The impact of abandoning social responsibility certifications: evidence from the decertification of SA8000 standard

Abstract

Purpose

In recent years, many companies have decided to decertify from their previously adopted corporate social responsibility (CSR) standards. The aim of this paper is to explore the phenomenon by focusing on the most important auditable CSR standard: Social Accountability 8000 (SA8000).

Design/methodology/approach

First, an event study is performed on a dataset composed of 136 SA8000 decertified public listed companies to analyse the possible relationship between certification, decertification and firms’ operating performance. Second, the authors shed light on the differences between 94 SA8000 (still) certified and the abovementioned 136 decertified firms. Finally, 10 interviews are conducted with decertified firms in the dataset to deepen the outcomes of the previous analyses.

Findings

The results show that, despite an initial positive effect in terms of sales and profitability, decertified companies experienced a reduction in productivity and profitability in the years following the certification, while positive outcomes emerged after the decertification. The study also highlights that certified and decertified firms differ in terms of home country, industry and labour intensity.

Originality/value

The paper contributes to the literature by opening the debate on an important but unexplored research area: the decertification from the most popular CSR standard, i.e. SA8000, and its relationship with firms' performance. In doing this, it also highlights the main differences between decertified and certified companies.

Citation

Podrecca, M., Orzes, G., Sartor, M. and Nassimbeni, G. (2021), "The impact of abandoning social responsibility certifications: evidence from the decertification of SA8000 standard", International Journal of Operations & Production Management, Vol. 41 No. 13, pp. 100-126. https://doi.org/10.1108/IJOPM-10-2020-0698

1. Introduction

Several initiatives, standards and codes of conduct have been developed to help organizations to integrate the corporate social responsibility (CSR) agenda with their business objectives and competencies. Among them, CSR standards stand out since they represent the most effective solution to operate in a socially sustainable way and to signal companies' efforts towards rules of conduct, ethics and moral integrity (de Andrade and Bizzo, 2019). While these standards are argued to contribute to enhance stakeholder relationships, increase sales and productivity performance and improve profitability (Sartor et al., 2016), recent data provided by some of the most important CSR regulatory bodies (e.g. Social Accountability International – SAI, 2020) show something unexpected: several organizations are revising their strategies by decertifying (withdrawing) from previously adopted CSR standards.

If compared to certification, the decertification process may have different characteristics and drivers. The effects of this decision at the company level (on business processes, on the relationships with the stakeholders and on operating performance) and macro-level (on social/ethical issues) could be significant. Moreover, the adherence to the requirements of CSR standards asks for major investments, and decertification could indicate a lack of benefits, thus posing several questions on their effectiveness for some firms. Except for the analysis of some characteristics of companies that left the CSR standard United Nations Global Compact – UNGC (Knudsen, 2011), scholars have not investigated CSR decertifications till now. Under this premise, our paper aims to explore the CSR decertification topic, focusing on the Social Accountability 8000 (SA8000). This choice has a twofold motivation. On the one hand, SA8000 is the most important and widely adopted CSR standard (Murmura and Bravi, 2020). On the other hand, it is recording an increasing number of firms that every year decide to decertify, moving from 280 out of about 3,000 total participants in 2010 (9% of the total participants) to 650 out of about 5,000 total participants in 2019 (13%) (SAI, 2020). This number of yearly decertified firms has currently become comparable to that of newly certified ones. Accordingly, the research questions that inspired our contribution are: What is the relationship between SA8000 certification, decertification and operating performance? What are the differences between SA8000 decertified and (still) certified companies?

To answer these questions, a mixed (quantitative-qualitative) approach was adopted. First, we analysed the possible relationship between certification, decertification and firms' operating performance through an event study performed on a dataset of 136 SA8000 decertified public listed companies. Second, we shed light on the differences between 94 (still) certified and the abovementioned 136 decertified public listed firms. Finally, we conducted 10 interviews with decertified firms in our dataset to understand the decision to leave SA8000. Results show that, despite some initial positive outcomes in terms of sales and profitability, decertified companies experienced a reduction in productivity and profitability in the years following the certification, while positive outcomes emerged after the decertification. By building on institutional theory, the study also shows that certified and decertified firms differ in terms of home country, industry and labour intensity.

Our contribution has several implications. From a theoretical point of view, we shed light on an unexplored research area: the decertification from the most popular CSR standard, i.e. SA8000, and its relationship with firms' performance. In doing this, we also highlight the usefulness of institutional theory in explaining the contextual differences between certified and decertified companies. From a managerial point of view, we show that the benefits associated with SA8000 certification may decrease over time. In this perspective, managers are required to constantly (re)think about how to keep the certification convenient and useful. When this is not possible, the decision to decertify may be a rational choice in some cases.

The remainder of the paper is organized as follows. Section 2 reviews the relevant literature; research hypotheses are here introduced. Section 3 presents the adopted methodology. Section 4 and Section 5 illustrate and discuss the findings. Finally, contributions and limitations are depicted in Section 6.

2. Literature review and research hypotheses

2.1 Literature review

Two research streams need to be introduced to understand the dynamics of SA8000 decertification. In the first Section (2.1.1), we focus on the literature analysing the effects associated with SA8000 certification. In the second Section (2.1.2), we focus on studies that shed light on the topic of decertification from (other) international standards and initiatives and in particular on the motivations for decertification and on the possible association between decertification and firms' performance.

2.1.1 Effects associated with SA8000 certification

Literature has investigated possible positive and negative effects associated with SA8000 adoption using different approaches: case studies, surveys and secondary data analyses (for a review on the topic see Sartor et al., 2016). As far as the positive effects are concerned (Table 1), scholars underline a potential relationship between some of the practices prescribed by SA8000 and the absenteeism/turnover rates of the adopting firms since workers may feel more protected and involved in the company's goals (e.g. Stigzelius and Mark-Herbert, 2009). Similarly, other studies (e.g. Fuentes-García et al., 2008) argue that SA8000 requirements can foster the firm's ability to attract a skilled workforce. SA8000 adoption could also contribute to the improvement of product quality, technical innovation (Beschorner and Müller, 2007) and labour productivity (e.g. Orzes et al., 2017). Other possible effects discussed in the literature involve increased confidence in the buyer–supplier relationship, reduced reputational risks and enhanced ability to attract new customers (e.g. Murmura et al., 2017). In this perspective, some authors observe a potential association with sales performance in terms of demand growth and revenue increase (e.g. Miles and Munilla, 2004): customers may be open to paying an extra fee (premium price) to buy from SA8000 certified firms (De Magistris et al., 2015). Moving the focus to the whole supply chain, the need to align suppliers on the same ethical obligations could strengthen communication and collaboration with business partners (e.g. Tencati and Zsolnai, 2009), with possible positive outcomes for supply chain performance (e.g. Ciliberti et al., 2009).

As for the negative effects, the most debated issues are related to costs. The number of obligations that companies have to respect (e.g. paid rest periods, extra compensation for the overtime work) is argued to contribute to increase the production costs (e.g. Koster et al., 2019). Moreover, some of the requirements imposed by SA8000 might involve limitations to the sourcing base (some suppliers may not be interested in meeting SA8000 obligations and therefore cannot be selected) and to supply chain flexibility (e.g. Merli et al., 2015). A possible association with increased coordination costs is also highlighted (e.g. Rohitratana, 2002). Finally, significant resources are reported as necessary to inform customers about the social commitment of the company (Salomone, 2008).

A key point emerging from our review is that several positive aspects (e.g. increased confidence in the buyer–supplier relationship, improved communication and collaboration with business partners) are not easy to quantify and their causal link with the certification is rather difficult to be explored, while the majority of the negative aspects are related to cost issues. At the same time, data provided by SAI show that, although the number of new companies that obtain SA8000 certification every year is high, equally high is the number of companies that decertify. This decertification trend poses several questions on the actual usefulness of SA8000 for some firms and on the long-term duration of its positive effects. Hence, a prominent need exists to conduct investigations on SA8000 decertification topic and to shed light on the differences between SA8000 certified and decertified firms.

2.1.2 Decertification from (other) international standards and initiatives

Decertification choices have been investigated mainly with reference to the ISO 9001 quality management standard. Within this research stream, we can identify studies focused on the motivations leading firms to decertify and studies shedding light on the possible association between decertification and firms' performance. As far as motivations are concerned, extant research has shown that firms leave ISO 9001 for two main reasons. On the one hand, ISO 9001 is sometimes considered too costly and time-demanding when compared to the operational outcomes (e.g. Lo and Chang, 2007; Zimon and Dellana, 2019). On the other, firms decertify because they perceive a reduction of market interest in the certification (Chiarini, 2019). Finally, Cândido and Ferreira (2021) and Ferreira and Cândido (2021) argue that ISO 9001 decertification decisions are related to a combination of both these aspects.

As for the decertification-performance association, Alič (2014) finds that 45% of surveyed companies closed during the four-year period after the decertification and argues that decertification could be perceived as a possible sign of declining financial performance. Similarly, Sansalvador and Brotons (2015) adopt the “discounted free cash flow methodology” to compare the net present value of ISO 9001 certified and non-certified companies, highlighting that the former exhibit higher values than the latter. In contrast, Cândido et al. (2016, 2021) find no significant relationship between decertification and financial performance.

The decertification from CSR standards is a phenomenon rarely considered in the literature. The only exception is the contribution of Knudsen (2011) about the UNGC; the author shows that the decertification [1] likelihood is related to some internal and external factors: size (small firms are more likely to decertify), home country (organizations coming from Africa, East Asia and Eastern Europe are more likely to decertify) and industry (firms belonging to oil and gas industry are less likely to decertify). Despite a common focus on CSR, UNGC and SA8000 differs in many aspects, UNGC is not an auditable standard (it lacks third-party inspection); it has a wider scope (covering not only labour rights); and it is based on less strict pillars (no quantitative obligations are imposed). We, therefore, argue that Knudsen's results (2011) cannot be generalized to SA8000.

Moreover, reviewed studies exhibit a paucity of theoretical underpinnings; none of them adopts a theory-driven approach to both formulate the research hypotheses and explain the outcomes. The only partial exception is Cândido et al. (2016, 2021), who however introduce a theoretical perspective (resource-based view) only to discuss the findings. Accordingly, following calls for more theory-grounded research on CSR (Zorzini et al., 2015) and SA8000 (Sartor et al., 2016), in the next section we will resort to a theory-driven approach to shed light on the differences between certified and decertified firms.

2.2 Research framework

Our research framework is structured in two parts. The first concerns the relationship between certification, decertification and firms' performance. The second deals with the differences between decertified and certified companies.

2.2.1 Certification, decertification and operating performance

Starting from Drucker's (1984) seminal contribution, where the author highlighted that companies should translate their social responsibilities into business opportunities, the effects of CSR practices on firms' financial/operational performance have gained noticeable consideration (see for example Carroll and Shabana, 2010; Bartolacci et al., 2020). A close examination of these contributions shows that, over the years, the initial voluntary and philanthropic approach to CSR has evolved into a more complex framework. CSR initiatives have currently become core aspects of firms' strategies (Boston Consulting Group, 2017) and are integrated with firms' competencies in order to improve profitability performance and support financial goals (Engert et al., 2016). This happens because the majority of a firm's stakeholders are interested in the financial outcomes of CSR involvement (Carroll and Shabana, 2010): managers are required to ensure that CSR strategies are consistent with the firm's overall strategy and produce economic benefits; shareholders are concerned about possible threats of CSR to the firm's economic priorities; socially-oriented consumers and social activists are well aware that companies will maintain their CSR commitment only if it is financially sustainable in the long term.

According to Kurucz et al. (2008), the CSR–profitability link can be described in terms of four strategies: reducing costs and risks (i.e. firms resort to CSR practices to improve their cost structure and stock value by reducing expenses associated with failures in addressing external pressures), gaining competitive advantage (i.e. firms adopt CSR practices to differentiate from the rivals and improve their marketing and selling capabilities), developing reputation (i.e. firms pursue CSR practices to enhance their reputation and attract consumers and investors), and seeking win-win outcomes through synergistic value creation (i.e. firms resort to CSR practices to improve their competitiveness by reconciling the different requests of their stakeholders). Without entering into the details of each motivation, it clearly emerges that there is a “business case” for CSR: firms seek an economic/financial argument that provides a rational justification for CSR activities (Carroll and Shabana, 2010). In line with the above reasoning, the final goal of CSR engagement can be summarized as “doing well by doing good” (Vogel, 2005). In other words, CSR activities – as for any management practice – are aimed at improving a firm's profitability and long-term value creation capabilities (Boston Consulting Group, 2020).

Enacted in 1997, SA8000 has currently become the most prominent standard in the area of CSR (Murmura and Bravi, 2020). It requires a redefinition of business processes in order to ensure that certified organizations address issues such as forced and compulsory labour, working hours, remuneration, disciplinary procedures, employee discrimination, health and safety in the workplace, freedom of association and the right to bargain collectively. Moreover, SA8000 is the CSR standard with the lowest level of decoupling (i.e. the mismatch between the certification dictates and what is actually implemented) (Behnam and MacLean, 2011).

In light of its CSR-related nature, we would therefore argue that SA8000 adoption represents “a rational economic decision” (Miles and Munilla, 2004, p. 6) “associated with expectations of business opportunities” (Stigzelius and Mark-Herbert, 2009, p. 48), which would be sustainable only if it translates into economic outcomes able to reward a firm's CSR engagement. Consistent with the above reasoning, we decided to focus on the business outcomes of SA8000 implementation and to investigate the relationship between certification, decertification and firms' performance in terms of operating/financial metrics (i.e. sales, labour productivity and profitability – Orzes et al., 2017; De Jong et al., 2014).

Literature shows that a certification could be abandoned if perceived as too costly when compared to the outcomes (Zimon and Dellana, 2019). As described in Section 2.1.1, the potential effects of SA8000 adoption on firms' performance are highly disputed; while some authors point out possible positive relationships, many others neglect these benefits or even highlight negative aspects. As far as commercial outcomes are concerned, some studies observe that the signalling effect provided by SA8000 is likely to increase sales performance (Sartor et al., 2016). This can be explained both by the opportunity to attract new customers (Battaglia et al., 2014) and the chance to charge a premium price by targeting “ethical consumers” (De Magistris et al., 2015). The contributions of Fuentes-García (2008) and Salomone (2008), however, question these findings claiming that SA8000 is not widely known and that significant resources are needed to inform customers about companies' social commitment.

Another debated topic is the relationship between SA8000 and the labour productivity of the adopting firms. Again, controversial findings are reported in the literature. Some studies highlight that some SA8000 practices might help to improve the working environment and the job satisfaction of the employees with potential benefits for staff motivation and productivity (Rohitratana, 2002). On the other hand, other scholars point out that the number of obligations that certified companies must respect in terms of wages, overtime work and social monitoring may contribute to increase production costs (Stigzelius and Mark-Herbert, 2009).

Finally, while positive relationships with personnel productivity, supply chain performance and sales should result in clear economic outcomes, both the contributions of Orzes et al. (2017) and Koster et al. (2019) do not report significant associations between SA8000 adoption and firms' profitability.

Summing up, the absence of relevant benefits could become an important driver for decertification as firms usually expect to be “rewarded” for gaining a certification (Fuentes et al., 2003). Hence, our first hypothesis is proposed as follows:

H1.

SA8000 decertified companies did not experience operating performance improvements after the adoption of the certification.

Specifically,

H1A.

SA8000 decertified companies did not experience sales performance improvements after the adoption of the certification.

H1B.

SA8000 decertified companies did not experience labour productivity improvements after the adoption of the certification.

H1C.

SA8000 decertified companies did not experience profitability improvements after the adoption of the certification.

Extant research on international standards (e.g. ISO 9001) highlights that certification benefits can gradually decrease as time goes by (Casadesús and Karapetrovic, 2005). As far as SA8000 is concerned, no study analyses the long-term effects of the adoption. Previous contributions have indeed only focused on short-term outcomes (only up to two years after obtaining the certification) (Orzes et al., 2017). However, positive effects associated with SA8000 could be time-sensitive too, as non-certified companies usually take countermeasures to reduce the gap with certified ones (Llach et al., 2015), thus lowering the differential advantages/benefits. In this respect, Orzes et al. (2017) show that the initial positive results achieved in terms of sales performance shrink even two years after gaining the certification. Koster et al. (2019) go even further, arguing that over time the choice to charge a premium price for SA8000 certified processes could lead consumers to choose cheaper products.

If the benefits associated with SA8000 adoption dissipate over the years, while the negative effects (e.g. increased labour and coordination costs, and the costs of surveillance audits) are likely to remain unchanged, decertification could be a rational choice to avoid (undue) costs and limitations. After the decertification, we might therefore expect an improvement of firms' performance. Our second hypothesis is, therefore, proposed as follows:

H2.

SA8000 decertified companies experienced operating performance improvements after the decertification.

Specifically,

H2A.

SA8000 decertified companies experienced sales performance improvements after the decertification.

H2B.

SA8000 decertified companies experienced labour productivity improvements after the decertification.

H2C.

SA8000 decertified companies experienced profitability improvements after the decertification.

2.2.2 Differences between decertified and certified firms

In light of the acknowledged context-dependency of CSR practices (e.g. Silvestre et al., 2020), several scholars use institutional theory to analyse the dynamics of CSR standards (e.g. Koster et al., 2019). This theoretical lens is based on the central tenet that organizations in the same competitive environment are subject to the same coercive (determined by other organizations firms depend on and by cultural expectations from society), normative (determined by the practices considered appropriate within specific contexts) and/or mimetic (determined by firms' orientation to imitate successful organizations) institutional pressures. This leads to “copycat” (imitation) practices and results in a subsequent homogeneity (isomorphism) across firms (DiMaggio and Powell, 1983). By applying institutional theory to SA8000 decertification, we would argue that the choice to decertify is also associated with some context-specific variables and therefore that certified and decertified firms compete in different contexts. According to CSR literature, some of the most important context-specific variables are country (e.g. Llach et al., 2015), industry (e.g. Chen et al., 2017) and labour intensity (e.g. Podrecca et al., 2021). We will develop below three hypotheses on these three dimensions.

As a result of the general increase in social awareness, many consumers are becoming reluctant to buy products coming from developing countries since some of them are characterized by low attention to workers' rights, mainly because of a lack of effective regulatory systems and/or controls (Parmigiani and Rivera-Santos, 2015). In these contexts, ethical standards like SA8000 allow firms to prove their social commitment (Gilbert and Rasche, 2007). Hence, companies in developing countries may perceive strong coercive and normative pressures to adopt and maintain CSR standards in order to gain legitimacy in the eyes of consumers (Huq et al., 2014).

Scholars have also associated CSR standards with the economic trade openness of the firm's home country; companies from open economies have commercial relations with a large number of nations; therefore, they often deal with foreign firms characterized by a high level of CSR (Wolf, 2014). Thus, to maintain these relationships, such companies usually need to comply with partners' coercive pressures towards CSR (Guler et al., 2002).

In sum, according to the underpinnings of institutional theory, national contexts could exert different institutional isomorphic pressures on the adoption and the maintenance of SA8000. Hence, the following hypothesis is posed:

H3.

SA8000 decertified and certified firms exhibit differences in home country characteristics (in particular the level of development and the economic openness).

Previous studies (e.g. Chen et al., 2017) have shown that the adoption and the effects of CSR practices are affected by the munificence (availability of resources to support firms' growth and reward their efforts), the dynamism (variability) and the competition (number and strength of competitors) of the industry in which organizations compete; the effects of these three dimensions are linked to the institutional pressures firms operating in distinct industries are subject to (Berrone et al., 2013).

In particular, high munificent industries are usually characterized by funding and tax/insurance premium reductions for sustainability and CSR programs, thus resulting in greater normative pressures for managers to engage in socially responsible practices (Lee et al., 2016).

As for the industry dynamism, rapidly changing customers' preferences, unpredictable technology development and inconsistent regulations (Prajogo and Oke, 2016) exert evolving coercive, normative and mimetic pressures resulting in frequent strategic plan redefinition, which may also concern CSR practices.

Finally, industry competition implies a more complicated situation. On the one hand, firms operating in highly competitive industries, characterized by market shares almost evenly distributed among a large number of competitors, perceive the need to resort to differentiation strategies such as the ones provided by CSR standards (Leong and Yang, 2020). On the other, in highly competitive industries with a large number of competitors, mimetic pressures are particularly strong; firms' strategies will be aggressively matched by rivals, reducing the possibility to generate abnormal rents from the diversification (Jansen et al., 2006).

Consistent with the dictates of institutional theory, we may expect that firms operating in different industrial contexts are subjected to different institutional pressures and therefore take different decisions as regards SA8000 decertification. Hence, the following hypothesis is stated:

H4.

SA8000 decertified and certified firms exhibit differences in industry characteristics (in particular munificence, dynamism and competition).

Over the last decades, many labour-intensive multinational firms (e.g. Nike, Adidas) have been affected by social scandals related to poor working conditions and child labour exploitation (Sartor et al., 2016). As a result, these companies are in the spotlight (Podrecca et al., 2021) and cannot conduct their business “under a cloak of secrecy” anymore (Post, 2013, p. 6). They are therefore asked to pay particular care to the labour conditions of their employees and to demonstrate these efforts to several influential stakeholders. Reading these requests through the lenses of institutional theory, we would argue that labour-intensive firms perceive strong coercive and normative pressures to adopt and maintain CSR standards, like SA8000, in order to show the attention paid to ensure decent work practices. Accordingly, the following hypothesis is posed:

H5.

SA8000 decertified and certified firms exhibit differences in labour intensity levels.

3. Methodology

3.1 Dataset

To analyse the hypotheses developed in the previous section, we decided to focus on the population of SA8000 certified and decertified firms listed on stock exchange markets. This approach is consistent with previous studies on quality management/sustainability standards (e.g. Corbett et al., 2005; Lo et al., 2014; De Jong et al., 2014) and allows us to measure performance through objective data rather than with self-reported (perceptive) measures. To build our dataset, two data sources were used:

1. The official list of SA8000 decertified and (still) certified firms available on the SA8000 website.

2. The Thomson Reuters Eikon database, which includes financial information for more than 30,000 public firms located in 88 countries.

The matching of SA8000 certified and decertified companies with Eikon data required a significant effort. Since company names often have a different format/wording in the two data sources, the direct word-for-word matching was not suitable. For this reason, a multi-step approach was followed. Two researchers independently searched in the Eikon database all the companies listed on the SA8000 website. In case of no matching, different parts of the name or formats were also searched. For each company, the address of the headquarter and the industry was carefully considered in order to ensure that the right firm was identified. As a second step, a further check on the firms' official websites, on Bureau van Dijk Orbis database and using the Google search engine was performed to verify (1) that the right firm was identified; (2) that the decertification was not related to events like plant/firm closure, mergers and acquisitions and (3) that firms' performance could not be influenced by other events within the same timeframe, following the suggestions of McWilliams and Siegel (1997) to look for “confounding events”. In this way, we identified a dataset of 94 still certified (Group 1) and 136 decertified firms (Group 2). For the purposes of our contribution, we needed complete balance sheet data covering the certification year as well as some years before and after this event. Accordingly, we focused on firms that adopted SA8000 up to 2013 (around 3,800 organizations) and that subsequently decided either to maintain their adherence to the standard (Group 1) or to decertify (Group 2).

Table 2 presents the distribution of the dataset by industry and country. Companies belong to several sectors, with a prevalence of apparel products, stone clay and glass products, and primary metal industries for SA8000 certified firms (Group 1); textile mill products, industrial machinery and equipment, and chemical and allied products for SA8000 decertified firms (Group 2). As for the distribution by country, SA8000 certified companies (Group 1) are mostly located in India, followed by China, and Vietnam; while SA8000 decertified firms (Group 2) are in India, China, and Brazil.

Our dataset captures the entire population of public listed SA8000 certified and decertified companies. There are, however, some differences between our dataset and the wider population of SA8000 organizations in terms of firms' size: small and medium enterprises are more frequent in the population of SA8000 firms (SAI, 2020) than in our dataset. This difference can be explained considering our focus on public listed companies. Nevertheless, several other studies on SA8000 and other certifications (e.g. ISO 9000, ISO 14001 and OHSAS, 18001) have focused on companies listed on stock markets for three main reasons: (1) they are more likely to adopt internationally recognized certifications, (2) they exhibit reliable cross-country accounting data, and (3) they have specific strategies and resources that require a separate analysis (e.g. De Jong et al., 2014; Lo et al., 2014).

3.2 Certification, decertification and operating performance

To analyse the hypotheses on the relationship between SA8000 certification, decertification and operating performance, we employed the long-term event-study methodology (Barber and Lyon, 1996). This approach was already used in the past for international standards and initiatives, like OHSAS 18001 (Lo et al., 2014), ISO 14001 (De Jong et al., 2014), and SA8000 (Orzes et al., 2017).

We adopted this methodology twice on a dataset of 136 companies that have decertified from SA8000 (Group 2): the first time to analyse the relationship between SA8000 certification event and firms' operating performance (H1) and the second to analyse the relationship between SA8000 decertification event and firms' operating performance (H2).

The certification process usually requires six months (La Rosa and Lo Franco, 2005). The same amount of time is also required, due to semi-annual audits, for the decertification (SAI, 2020). The event period was therefore defined as the year in which a company decided to adopt (first event-study) or to leave (second event-study) SA8000 (year t). The year preceding the event period (t − 1) was considered the base year and used to determine the control firm sample while year t−2 was taken into account to avoid endogeneity issues. In line with previous studies (Orzes et al., 2017; Lo et al., 2014), the following operating performance measures were adopted:

1. Labour productivity: the ratio of operating income to the number of employees;

2. Sales performance: the relative sales growth, defined by (SALES_t−SALES_{t −1})/SALES_t − 1 and

3. Profitability: the return on assets (ROA).

For each decertified firm, two portfolios (the first for the certification event study and the second for the decertification event study) of control firms that have never adopted SA8000 were created based on a combination of three criteria (Barber and Lyon, 1996):

1. The industry: matched firms have the same two-digit first SIC code of the dataset firm

2. The size: matched firms have 50–200% of the dataset firm's total assets in year t − 1 and

3. The measure: 90–110% of the dataset firm's considered performance (labour productivity, sales growth or ROA) in year t − 1.

We then estimated the abnormal performance change of the sampled decertified firms in relation to the control firms as follows:

AP(t+b)=PS(t+b)−EP(t+b)

EP(t+b)=PS(t+a)+(PC(t+b)−PC(t+a))

where AP is the abnormal performance, EP is the expected performance, PS is the actual performance of the sampled firms, PC is the median performance of the control firms, t is the year of certification/decertification, a is the starting year of comparison, and b is the ending year of comparison.

To evaluate the paired data of the event-study, Barber and Lyon (1996) suggest that non-parametric tests are more robust than parametric t-tests. Specifically, when the distribution is symmetric, the Wilcoxon signed-rank (WSR) test will suffice. However, if the distribution is highly skewed, the sign test is suggested to be more appropriate. Hence, we analysed whether the abnormal performance differed significantly from zero through the WSR test and the sign test.

3.3 Differences between SA8000 decertified and certified firms

To analyse the hypotheses on the differences between SA8000 certified and decertified firms, we ran some statistical tests on the dataset obtained by combining the 94 certified (Group 1) with the 136 decertified firms (Group 2). As far as the country is concerned, our research hypothesis (H3) postulated that certified and decertified firms differ in terms of home country development level and economic openness. Hence, we considered the following variables (Orzes et al., 2017; Knudsen, 2011):

1. Development: human development index of the certification year.

2. Economic openness: the ratio of international trade (the sum of imports and exports) to the gross domestic product (GDP) of the certification year.

As far as the industry is concerned, we hypothesized (H4) that decertified and certified firms exhibit differences as regards munificence, dynamism and competition. Hence, we included (Jacobs et al., 2015; Lo et al., 2013):

1. Munificence: the slope coefficient obtained by regressing sales over the five-year period before the certification year and dividing by the mean sales in the same period.

2. Dynamism: the standard error of the slope coefficient obtained by regressing sales over the five-year period before the certification year and dividing by the mean sales in the same period.

3. Competition: defined as 1- Herfindahl Index of the certification year.

Finally, we hypothesized (H5) that decertified and certified firms exhibit differences as regards labour intensity levels that we measured as follows (Lo et al., 2014):

1. Labour intensity: the ratio of the number of employees to the total assets of the firm.

We also included six control variables to account for some firms' characteristics (Orzes et al., 2017; Lo et al., 2013, 2014):

1. Firm size: the logarithm of the firm's total assets in the certification year.

2. Financial performance: the ROA in the certification year.

3. ISO 9001: a dummy variable indicating whether the firm has the ISO 9001.

4. ISO 14001: a dummy variable indicating whether the firm has the ISO 14001.

5. OHSAS 18001/ISO 45001: a dummy variable indicating whether the firm has the OHSAS 18001 or the ISO 45001.

6. UNGC: a dummy variable indicating whether the firm has the UNGC.

Given the non-normal nature of the data, to shed light on the differences between the 94 certified (Group 1) and 136 decertified (Group 2) firms, we followed Siegel and Castellan's (1998) suggestion and adopted non-parametric tests. Specifically, the Mann–Whitney U test for quantitative data, and the Fisher and Chi-squared tests for binary (dummy) data were used.

3.4 Qualitative insights

After the evaluation of the quantitative findings, we conducted some interviews with companies in our dataset of decertified firms to better understand the decertification decision. We contacted (by email) all companies included in our dataset asking their availability for an interview. Ten of them agreed. This sample of ten companies covers all the main continents of SA8000 certified (and decertified) companies and includes firms of different sizes (Table 3).

Each interview was conducted with the manager responsible for SA8000 and lasted on average 60 min. Four companies also gave us the availability of a second respondent knowledgeable about certification (and decertification) choices. A semi-structured approach was adopted to allow a more open dialogue characterized by knowledge sharing, mutual reflection, and wider coverage of the topic under investigation (Kvale and Brinkmann, 2009; Yin, 2017). The interview protocol, developed on the basis of the existing literature on decertification and SA8000 (Yin, 2017), consisted of five parts:

1. Company and interviewee profile. To start each interview, we asked the interviewee to introduce him/herself and his/her role in the company and to describe the firm, its product/service offer and the industry.

2. Objectives and effects of the initial SA8000 adoption. This part shed light on the driver(s) leading the firm to adopt SA8000, the effects associated with the adoption and the stability of such effects over time, competitor reactions, and the changes in the firm's practices due to the certification.

3. Reasons for the abandonment of SA8000. We asked firms to describe the reason(s) leading to decertify from SA8000. In particular, we investigated which specific problems they encountered in dealing with SA8000 and whether they believe that in some contexts SA8000 is less useful (or more difficult to manage) than in others.

4. Implications of the abandonment of SA8000. This part was aimed at understanding the effects of the decision to abandon SA8000 and whether SA8000 practices were maintained or not after the decertification.

5. Quantitative findings discussion. To conclude each interview, we showed the interviewees the quantitative results of our study in order to collect their feedbacks.

The interviews were conducted mainly in English. To improve the accuracy of the findings, we asked the respondents their availability to have the interview recorded; six out of ten agreed. To ensure inter-coder reliability (Kvale and Brinkmann, 2009), two researchers individually collected field notes and conducted the data analysis. At the end of all the interviews, the researchers shared their notes and transcripts and discussed their findings with the rest of the team. During this phase, an external researcher taking the role of “resident devil's advocate” was also involved.

4. Results

4.1 Certification, decertification and operating performance

Tables 4 and 5 summarize the outcomes concerning the event study performed on Group 2 to shed light on the relationship between certification, decertification and operating performance. As for the certification event (Table 4), we find significant positive abnormal sales (from t − 1 to t, t + 1, t + 2) and profitability (from t − 1 to t) performance in the short term. Subsequently, in the medium/long-term, negative abnormal labour productivity (from t − 1 to t + 3, t + 4, t + 5) and profitability (from t − 1 to t + 5) emerge. On the other hand, in the decertification event study (Table 5), firms exhibit significant positive abnormal labour productivity (from t − 1 to t + 1, t + 2) and profitability (from t − 1 to t + 1). For what concerns our research hypotheses:

1. Decertified companies initially benefited from their decision to adopt SA8000 in terms of both sales and profitability (H1A and H1C not supported), while they exhibit no short-term effects as regards labour productivity (H1B supported);

2. As time goes by, the initial positive outcomes dissipate and productivity-related barriers emerge. After the decertification, no significant results in terms of sales performance have been observed (H2A not supported), while positive effects have emerged as regards productivity and profitability (H2B, H2C supported).

We also checked for the absence of firms' financial issues before the decertification decision. On the one hand, the outcomes of the second event study show no significant negative abnormal profitability performance before the decertification event (from t − 2 to t − 1, see Table 5). On the other, we also tested the ROA of decertified firms before the decertification (t − 1 and t − 2) and found that it was positive and significantly higher than zero.

4.2 Differences between SA8000 decertified and certified firms

Table 6 shows the differences between SA8000 certified (Group 1) and decertified firms (Group 2). We found significant differences for the variables Development, Dynamism, Competition and Labour Intensity but not for Economic Openness and Munificence. In particular,

1. Decertified firms (Group 2) are located in more developed countries than certified firms (Group 1) (median value of the variable Development 0.69 against 0.58; p < 0.01);

2. Decertified firms (Group 2) compete in more dynamic industries than certified firms (Group 1) (median value of the variable Dynamism 3.70 against 2.24; p < 0.01);

3. Decertified firms (Group 2) operate in more competitive industries than certified firms (Group 1) (median value of the variable Competition 0.78 against 0.61; p < 0.001).

4. Decertified firms (Group 2) are less labour-intensive than certified firms (Group 1) (median value of the value Labour intensity 0.08 against 0.11; p < 0.05).

None of the control variables exhibit significant differences. In light of the outcomes presented so far, H3 and H4 are only partially supported, while H5 is fully supported.

Finally, as the analysed dataset only consists of public listed companies, we tried to understand the generalizability of our results to the wider population of all SA8000 decertified and certified companies. This was possible only for the country-related variables (Development and Economic openness) for which we could retrieve reliable data from the list of companies available on the SA8000 website. The analyses performed confirmed the results presented above: decertified firms are located in more developed countries than certified firms (p < 0.05). This gives us confidence about the robustness of our results and suggests that they could be generalized to the wider population of SA8000 decertified and certified firms.

4.3 Robustness check

As explained in Section 2.2.1, the strict dictates of SA8000 (i.e. third-party surveillance audits conducted by accredited bodies, sanction mechanisms and continuous improvement approach) led scholars to classify it as the CSR standard with the lowest level of decoupling (e.g. Behnam and MacLean, 2011). Nonetheless, we also verified this assertion by analysing the relationship between SA8000 certification, decertification and the CSR performance of the firms in our dataset. This check is based on the assumption that the effective adoption of the practices of SA8000 should be positively associated with the CSR performance of the adopting firms, while if firms decertify and SA8000 practices are discontinued, CSR performance is expected to deteriorate.

We adopted the “workforce score” from the Refinitiv (formerly Thomson Reuters) Environmental, Social, Governance (ESG) database. This variable measures the company's efforts towards job satisfaction, a healthy and safe workplace, maintaining diversity and equal opportunities, and development opportunities for its workforce and is therefore directly related to the areas covered by SA8000. We found data for about 40% of the companies in our dataset.

Two event studies were performed following the procedures outlined in Section 3.2. In the first event study, we analysed the association between the SA8000 certification event and the workforce score of both certified (Group 1) and decertified firms (Group 2). In the second event study, we analysed the association with the decertification event (Group 2). Results – reported in Table 7 – show that after the adoption of SA8000, companies exhibit significant positive values of the workforce score, while after the decertification, negative values emerge. These findings seem to confirm our assumptions and allow us to be reasonably confident that the outcomes of our study are driven by the adoption of SA8000 practices. To obtain further confirmation, these aspects have also been investigated during the qualitative interviews.

5. Discussion

In this section, we will draw on both the quantitative findings of our analyses and the qualitative evidence from the interviews to explain the possible relationship between certification, decertification and operating performance, and the contextual differences between certified and decertified firms.

5.1 Certification, decertification and operating performance

Our study shows that SA8000 decertified companies (Group 2) reported significant positive effects in terms of sales and profitability (immediately) after the certification. However, these outcomes dissipated over time, and, in the medium/long-term, negative findings related to labour productivity and profitability were recorded. Furthermore, after the decertification significant positive results in terms of labour productivity and profitability emerged, while no significant outcomes were found for sales performance.

As for the sales effects (H1A), the literature suggests that SA8000 adoption can provide reputational advantages to certified firms by signalling their CSR commitment to the customers (Murmura and Bravi, 2020), with possible commercial benefits (Werre, 2003). However, our findings show that, for some firms (Group 2), this effect is time-sensitive. Based on the qualitative insights from the interviews, the motivation is twofold. On the one hand, according to case companies C, D, F, and H, the positive commercial outcomes associated with the adoption of SA8000 led competitors to take countermeasures in order to maintain their competitive position. Company F, for instance, stated: “when we decided to join SA8000 we were the only firm in our market to have a certified tool to signal our CSR efforts, but in the following years several competitors adopted similar strategies weakening the differentiation power provided by SA8000”. On the other hand, according to the cases E, G, and I, companies experienced problems to get a premium price for their CSR efforts. For instance, Company E stated: “In the initial period of SA8000 adoption, we were able to charge a premium price on our customers and to acquire new clients attracted by our CSR-focus. However, after some years our business partners stopped considering SA8000 a winning factor and were no longer interested in recognizing our efforts with a premium price”.

As far as labour productivity outcomes are concerned (H1B), decertified firms did not exhibit any short-term significant effect at the time of their initial certification. From the qualitative insights, the impression emerges that decertified firms adopted SA8000 in a reactive rather than proactive way, i.e. their main goal was the compliance with the requirements of the standard rather than the continuous improvement of the internal practices. The majority of the interviewed companies reported that during the surveillance audits, there were often some non-compliances. Third-party auditors therefore made increasingly insistent requests with consequent productivity disadvantages in the medium/long term. Company B and H stated that one of the main difficulties they met was the need to conduct inspections on the suppliers. “Over time, this issue has become so critical that we needed some dedicated experts; this increased the costs to maintain the certification” (Company B). In two cases (Company E and Company H), SA8000 brought problems with workers. Company E − headquartered in a developing area – highlighted, for instance, that, despite the adoption of SA8000 brought better working conditions, the employees were not interested in this aspect; their main goal was to work as much as possible to maximize their household income. They, therefore, went on several strikes against the overtime restrictions imposed by SA8000.

Finally, we found a positive short-term association between SA8000 certification and profitability (H1C) (from t − 1 to t), which however became negative in the medium/long period (t − 1 to t + 5). As also confirmed by the interviews, this result can be explained considering the abovementioned results on sales (short-term positive outcomes, no medium/long-term outcomes) and productivity (no short-term outcomes, negative medium/long-term outcomes). Moreover, many interviewees highlighted that the commercial advantages provided by SA8000 have only partially translated into economic benefits due to the high cost of certification maintenance. For instance, Company C argued: “in the initial years of SA8000 adoption we improved our revenues. However, managing the standard took a lot of time and required some dedicated resources, and this increased the costs too”. This could explain why, despite the abnormal sales performance was significant for several years, profitability outcomes were limited.

With regard to the decertification, we found positive outcomes in terms of labour productivity and profitability, while no significant results emerged for sales. The absence of an association between decertification and sales performance (H2A) provides further confirmation to our previous reasoning; the differentiation power of SA8000 decreased over time to the extent that the standard was no longer having any (positive or negative) commercial effect. In this perspective, Company A's statement was particularly interesting: “although SA8000 was considered a mandatory requirement to create some business partnerships, after our decision to decertify no customer interrupted its relationship with us”. As for the positive findings related to labour productivity (H2B) and profitability (H2C), it emerged that, after the decertification, all sampled firms modified their processes dismissing some (costly) practices imposed by SA8000. This suggests that the choice to decertify could pay off in the short-term as it allows the company to avoid the limitations (and the costs) of SA8000. Thanks to the cost savings associated with the decertification decision, the majority of the case companies also reported “budget surpluses” that were then re-invested in the internal activities of the firm. Motivated by this assertion, we also performed some tests (WSR and sign test) on the total assets of the decertified firms (Group 2) and found a statistically significant increase from the first to the second year after decertification. This could explain why, in the post-decertification period, the profitability was significant only during the first year.

Summing up, all the interviewees explicitly stated that their main motivation to leave SA8000 was related to the worsening of the cost-benefit ratio over the years: “if we had to take stock of our experience with SA8000, we would say that initially we achieved some economic benefits, but over time the effort and the cost required to maintain the adherence with its requirements become greater than the economic return and for that reason we decided to leave it” (Company B).

To conclude, it is worth highlighting that our results are particularly interesting in light of the wider debate around the effects of CSR standards on firms' financial performance. In particular, they lead to the conclusion that the outcomes of CSR standards adoption may change over time and that companies would maintain their certified status (and the associated practices) only until the financial/operational benefits exceed or at least compensate the imposed restrictions (and the resulting costs). While future studies are needed to extend these findings also to non-public listed companies, in the medium/long-term, and to CSR standards other than SA8000, our paper highlights the pressing need to analyse more in detail the CSR decertification strategic choice.

5.2 Differences between SA8000 decertified and certified firms

Our study further highlights that SA8000 certified (Group 1) and decertified (Group 2) firms exhibit differences in terms of country, industry and labour intensity. From an institutional theory perspective, these findings show that the isomorphic pressures resulting from the different institutional environments in which firms are embedded affect their choices and lead companies with analogous backgrounds to resort to similar strategies as regards SA8000 decertification. In other words, the context plays a role in the decision to decertify from SA8000 and more in general from CSR standards.

As far as country characteristics are concerned (H3), we found that decertified (Group 2) firms are in general located in more developed countries than still certified (Group 1) firms. While three case companies located in developing countries perceived the need to implement SA8000 to legitimize their activities (at least initially), companies in developed countries reported statements such as “differently from ISO 9001 and ISO 14001, SA8000 is not strategic for our business” (Company B). In particular, Company B adopted SA8000 to show their external stakeholders the supplementary attention paid to social responsibility aspects. This difference seems related to the regulatory background, as governments in developed countries usually pay higher attention to workers' rights and labour practices (Sartor et al., 2016). Reading this result through the lenses of institutional theory, our findings show that companies in developed countries already face the need to comply with strict legislative systems and therefore are not (or less) required to demonstrate their social commitment (i.e. CSR standards are generally implemented for differentiation purposes). On the contrary, in developing countries labour and social laws are (or are perceived to be) weaker, and this forces firms to conform with strong coercive and normative customers' pressures towards CSR standards adoption.

As for the differences related to the industrial context (H4), we found that decertified companies (Group 2) operate in more competitive and dynamic industries than still certified organizations (Group 1). According to institutional theory, in highly competitive markets, where firms fiercely fight for customers, the mimetic pressures are particularly strong and lead to imitation practices. This happens because in these contexts, mimicking rivals' actions is considered a safe choice to preserve the competitive position of the organization (DiMaggio and Powell, 1983). As a result, firms tend to become isomorphic, and diversification strategies are usually aggressively matched by competitors. This reduces the possibility to keep the extra gains generated by these tactics in the medium/long run (Jansen et al., 2006). By contrast, in monopoly or oligopoly industries, mimetic pressures are generally low, and firms have strong power to charge a price premium for their products and to increase their profits by adopting differentiation approaches, like SA8000 adherence. Further support for our theory-based reasoning comes from the interviews: organizations belonging to more competitive industries lost their initial competitive advantage within a few years since the competitors soon resorted to similar CSR practices. Moreover, case companies highlighted that in competitive industries, consumer choices are more cost-driven, and premium prices for CSR commitment are less viable. As for the dynamism of the industry, our findings are aligned with the predictions based on institutional theory. In particular, while firms operating in less dynamic industries are subject to relatively stable pressures and are characterized by steady processes, firms competing in more dynamic contexts often face the need to modify their activities to keep up with constantly evolving coercive, mimetic and normative isomorphic pressures. One of the main drawbacks found during the interviews was that in dynamic environments, SA8000 suffered from “a lack of fit to the business needs” (Company F) as its obligations usually resulted in very rigid processes, which reduced the firms' ability to quickly adapt to market requirements. In this regard, Company H also added: “we often experienced the need to modify some activities to be in line with customers' demands. In many cases, however, the rules imposed by SA8000 prevented these changes”.

Finally, we found that decertified firms (Group 2) are less labour-intensive (H5) than certified firms (Group 1). This result can be explained in a twofold way. First, in contexts potentially more exposed to social issues (as it is for labour-intensive organizations), companies have to deal with a high level of scrutiny: these firms are faced with relevant stakeholders' pressures asking them to declare their social commitment and offer guarantees about their practices (Podrecca et al., 2021). CSR standards are benchmarks in CSR levels set by adherent organizations (Husted et al., 2016); they therefore allow labour-intensive companies to answer these pressures and gain the required legitimacy to operate in the market. Second, highly labour-intensive activities are often based on manual production routines. As a result, it is easier to adapt to the dictates of a standard when firms only need to modify some simple operational practices, rather than when production methods include significant technological complexity and require specific equipment (Lo et al., 2013).

6. Conclusions

6.1 Contributions to theory

Our paper contributes to operations management and CSR literature in at least five significant ways.

First, we are the first to shed light on the potential performance effects of SA8000 adoption on decertified companies. Findings show for these firms a statistically significant positive association in the short-term between SA8000 adoption and operating performance, followed by a decrease in the following years. These results enrich the debate on SA8000 and show that, for some firms, SA8000 medium/long-term benefits are not enough to overcome the significant resource investment required and the constraints imposed by the standard. Furthermore, by highlighting that the outcomes of CSR standards adoption may change over time, they also contribute to the wider literature on the effects of CSR practices on firms' performance. This should suggest scholars to adopt a more time-dependent perspective and to consider also the medium/long-term outcomes.

Second, we are also the first to analyse the effects of SA8000 decertification on operating performance, showing positive results in terms of labour productivity and profitability. In this way, we highlight that CSR standards are not beneficial for all the companies that decide to get certified, and we provide a potential explanation for the high number of yearly SA8000 decertified firms. Moreover, while extant literature focuses mainly on the certification choice, our study emphasizes the need to consider also the decertification from other CSR and sustainability standards (e.g. ISO 14001, Global Reporting Initiative) and to explore both drivers and performance-related outcomes of this decision.

Third, while only a few scholars have resorted to grand theories to discuss the findings of their contributions on the decertification topic (Cândido et al., 2016, 2021), we are the first to adopt a theory-driven approach to shed light on the differences between certified and decertified organizations. Specifically, building on institutional theory, we highlight that the institutional environment (country, industry and firm-specific characteristics) and the resulting isomorphic pressures faced by companies might play a role in the decision to decertify from international standards and in particular from CSR-related ones. This finding provides two important contributions. On the one hand, it shows the usefulness of institutional theory in shedding light on the differences between certified and decertified firms. On the other, it calls OM and CSR scholars for more theory-based research on firms decertification choices.

Fourth, according to Sartor et al. (2016), extant research is characterized by a paucity of contributions shedding light on the role of contextual factors in affecting the adoption and the outcomes of SA8000. By showing that decertified firms mainly operate in highly competitive and dynamic industries, our contribution extends the (scant) knowledge on the topic suggesting scholars develop a deeper consideration of the industrial context in which SA8000 is implemented and highlighting the need to conduct comparative studies among contexts with different characteristics.

Finally, by resorting to a cross-country/industry dataset based on objective data and performing rigorous matching and analyses deepened by qualitative evidence, our study provides a solid approach to clarify the relationship between certification, decertification and firms' performance. This represents a significant advancement in a research field that mainly builds on single country case-based research and survey studies and which is often characterized by conflicting evidence.

6.2 Contributions to practice

Thousands of firms are regularly called to decide whether or not to renew their commitment to SA8000. This choice is strategic for both companies and society, considering the importance of the topics addressed and the potential effects associated with the adherence to this certification scheme.

Since our study shows that SA8000 benefits might decrease over time, managers should constantly adopt approaches and strategies to ensure that the certification remains convenient year after year. In terms of sales-related aspects, for instance, firms could regularly communicate to the customers the added value provided by SA8000 and the relevance of having socially acceptable practices in the workplace. This could guarantee commercial benefits in the long run, thereby allowing companies to cope with the costs associated with the ongoing management of SA8000. As for the productivity-related findings, a proactive approach (which may also include periodic internal surveys/interviews with the staff) may shed light on problems and limitations arising from some practices imposed by SA8000. This may help firms to improve their activities, reverse negative outcomes and thus avoid decertification. Should SA8000 still prove problematic despite these actions, managers might consider decertifying from SA8000 and possibly moving to other CSR initiatives (e.g. industry-specific standards or company codes of conduct) more suited to the characteristics of their firms and of the markets where they compete.

Moreover, the paper highlights some contextual differences between decertified and certified firms which may warn managers interested in joining SA8000. Since the list of SA8000 certified and decertified organizations is publicly available, managers could apply some benchmarking analyses in order to observe the profile of companies keeping and leaving the standard. This could allow them to understand whether similar firms are preserving their adherence to SA8000 or whether many of them have already resorted to different strategies.

Our study also has implications for regulatory bodies (e.g. SAI and ISO). The fact that there are country, industry and firm-related differences between SA8000 decertified and certified firms might, for instance, advise SAI to investigate the specific factors that make SA8000 ineffective in such contexts. In this way, the certification body could support organizations to correctly orientate their activities and help them to anticipate the problems that may lead to decertification.

Finally, as “citizens of the World”, we believe that shedding light on the phenomenon of CSR decertification will lead all the stakeholders to consider the problem. Disengagement from social responsibility is not good news: our main aim is to have a future in which responsibility is deeply rooted in the consciousness and practices of individuals and companies. The first step in this direction is precisely to understand when and in which contexts the economic interests of companies and the concerns for social conditions are diverging.

6.3 Limitations and future research

The results of our study should be viewed in light of some limitations.

First, we only included public-listed companies. Such choice granted reliable financial data, but, at the same time, it also imposed some restrictions in terms of the composition of the analysed dataset. While previous contributions (e.g. Orzes et al., 2017) already faced this shortcoming and we partially tried to overcome this limitation by testing country-related variables on the entire list of SA8000 certified and decertified organizations, future research could check the generalizability of our results by focusing on all SA8000 firms.

Second, we did not consider the drivers that lead firms to adopt SA8000 and how these might differ between certified and decertified firms. Further research might exploit the impact of several antecedents on the decertification outcome.

Third, we matched SA8000 certifications, which involve individual sites, with corporate financial performance. As outlined by Corbett et al. (2005), this choice makes the tests more conservative and therefore should not be considered a major issue. Moreover, the interviewed firms argued that, although the certification was issued only on one or few plants, SA8000 practices were generally adopted in the whole production network. As a future step, if reliable disaggregated performance data on individual sites become available, it would be worthwhile to replicate our analyses at the individual site level.

Fourth, while we carried out a first attempt also to also collect qualitative evidence to deepen the outcomes of our quantitative analyses and highlight the main drivers of SA8000 decertification, future research should conduct more structured case studies to further investigate the decertification process.

Finally, as this study is focused only on SA8000, another direction for further research is linked to a comparative analysis of the decertification from other CSR initiatives and standards.

Table 1

Overview of positive and negative effects associated with SA8000

Positive effects	Exemplary references	Negative effects	Exemplary references
Improvement of workingconditions	Castka and Balzarova (2008)	High costs	Fuentes-Garcia et al. (2008)
	Tencati and Zsolnai (2009)Murmura and Bravi (2020)		Stigzelius and Mark-Herbert (2009)Koster et al. (2019)
Enthusiasm among employees	Stigzelius and Mark-Herbert (2009)Merli et al. (2015)Murmura and Bravi (2020)	Labour constraints and production cost increase	Miles and Munilla (2004)de Andrade and Bizzo (2019)Koster et al. (2019)
Absenteeism and staff turnover reduction	Henkle (2005)Stigzelius and Mark-Herbert (2009)Murmura et al. (2017)	Constraints in supplier selection	Klassen and Vereecke (2012) Merli et al. (2015) El Abboubi et al. (2021)
Increased attractiveness of skilled workforce	Miles and Munilla (2004)Fuentes-Garcia et al. (2008)Merli et al. (2015)	Decreased supply chain flexibility	Leipziger (2009)Leipziger (2010)Merli et al. (2015)
Product quality and technical innovation improvement	Beschorner and Muller (2007) Koplin et al. (2007) Gilbert and Rasche (2007)	Poor knowledge of the standard by the customers	Salomone (2008)Fuentes-Garcia et al. (2009)Koster et al. (2019)
Increased labour productivity and business process efficiency	Rohitratana (2002)	Complex data management	Leipziger (2001)
	Battaglia et al. (2014)		Leipziger (2009)
	Orzes et al. (2017)		Leipziger (2010)
Improvement of corporate image and reputational risks reduction	Rohitratana (2002)	Increased delivery times	Leipziger (2009)
	Battaglia et al. (2014)		Leipziger (2010)
	Murmura et al. (2017)
Brand loyalty and demand growth	Werre (2003)
	De Magistris et al. (2015)
	Orzes et al. (2017)
Increased communication and collaboration with business partners and improved supply chain performance	Tencati and Zsolnai (2009)
	Ciliberti et al. (2009)
	Chiarini and Vagnoni (2017)

Table 2

Dataset breakdown by industry and country

SA8000 certified firms (Group 1)				SA8000 decertified firms (Group 2)
Industry	SIC code	Number	Percentage	Industry	SIC code	Number	Percentage
Apparel products	23	11	11.70	Textile mill products	22	22	16.18
Stone, clay, and glass products	32	9	9.57	Industrial machinery and equipment	35	20	14.71
Primary metal industries	33	9	9.57	Chemical and allied products	28	18	13.24
Fabricated metal products	34	8	8.51	Electronic and other electric equipment	36	10	7.35
Industrial machinery and equipment	35	5	5.32	Wholesale trade – Durable goods	50	5	3.68
Other		52	55.32	Other		61	44.85
		94				136

SA8000 certified firms (Group 1)			SA8000 decertified firms (Group 2)
Country	Number	Percentage	Country	Number	Percentage
India	42	44.68	India	36	26.47
China	19	20.21	China	16	11.76
Vietnam	13	13.82	Brazil	16	11.76
Pakistan	8	8.51	Italy	15	11.03
Italy	3	3.19	Taiwan	12	8.82
Other	9	9.57	Other	41	30.14
	94			136

Table 3

Company and interviewee profiles

	Company A	Company B	Company C	Company D	Company E	Company F	Company G	Company H	Company I	Company J
Region	Europe	Europe	Europe	Europe	Asia	Asia	Asia	South America	South America	South America
Sector	Manufacturing	Manufacturing	Service	Manufacturing	Manufacturing	Manufacturing	Manufacturing	Manufacturing	Manufacturing	Utilities
Employees	1,000–5,000	<1,000	1,000–5,000	1,000–5,000	>10,000	1,000–5,000	1,000–5,000	>10,000	>10,000	<1,000
Interviewee(s) role	Quality manager	Human resources manager	Quality manager	Human resources manager + social compliance manager	General manager operations	Vice president human resources + social compliance manager	Vice manager human resources + quality manager	Operations director	Quality manager + sustainability manager	Quality manager

Table 4

Abnormal performance of the certification event (Group 2)

Certification event study
Period	AP median	AP mean	Normality	Skewness	p-value (WSR)	p-value (sign test)
Labour productivity (operating income/number of employees)
t − 2 to t − 1	386.79	−1204.31	No	X	0.265	0.237
t − 1 to t	517.53	−1743.32	No	X	0.525	0.268
t − 1 to t + 1	766.40	−2663.08	No	X	0.956	0.434
t − 1 to t + 2	−546.54	−4519.96	No		0.326	0.434
t − 1 to t + 3	−1836.57	−5234.83	No		0.008**	0.002**
t − 1 to t + 4	−2488.96	−7640.71	No	X	0.000***	0.001**
t − 1 to t + 5	−3571.07	−9298.90	No	X	0.000***	0.001**
Sales performance (yearly percentage change in industrial sales)
t − 2 to t − 1	1.03%	−1.49%	No	X	0.922	0.396
t − 1 to t	5.23%	8.32%	No	X	0.001**	0.003**
t − 1 to t + 1	9.01%	18.38%	No		0.000***	0.000***
t − 1 to t + 2	6.21%	14.87%	No	X	0.001**	0.002**
t − 1 to t + 3	4.70%	7.81%	No	X	0.067	0.341
t − 1 to t + 4	1.54%	1.12%	No		0.482	0.864
t − 1 to t + 5	2.00%	3.90%	No		0.460	0.797
Profitability (return on assets)
t − 2 to t − 1	−0.58%	−0.88%	No	X	0.155	0.253
t − 1 to t	1.35%	0.97%	No	X	0.028*	0.039*
t − 1 to t + 1	1.07%	1.05%	No		0.096	0.253
t − 1 to t + 2	0.98%	0.75%	No	X	0.159	0.277
t − 1 to t + 3	0.51%	−0.76%	No	X	0.813	0.608
t − 1 to t + 4	−0.83%	−1.94%	No		0.207	0.514
t − 1 to t + 5	−1.51%	−2.46%	No		0.004**	0.017*

Note(s): *, ** and *** significant at 5, 1 and 0.1% levels, respectively (Benjamini-Hochberg's (1995) false discovery rate correction)

Table 5

Abnormal performance of the decertification event (Group 2)

Decertification event study
Period	AP median	AP mean	Normality	Skewness	p-value (WSR)	p-value (sign test)
Labour productivity (operating income/number of employees)
t − 2 to t − 1	−115.00	2678.30	No	X	0.875	1.000
t − 1 to t	622.67	3264.57	No	X	0.379	0.641
t − 1 to t + 1	955.62	5297.52	No		0.046*	0.271
t − 1 to t + 2	1745.17	8665.25	No	X	0.007**	0.000***
Sales performance (yearly percentage change in industrial sales)
t − 2 to t − 1	−0.53%	−0.48%	No		0.582	0.932
t − 1 to t	1.62%	1.17%	No		0.671	0.646
t − 1 to t + 1	2.09%	1.43%	No		0.249	0.265
t − 1 to t + 2	−0.69%	0.65%	No	X	0.897	0.797
Profitability (return on assets)
t − 2 to t − 1	−0.24%	0.40%	No	X	0.655	0.668
t − 1 to t	0.52%	0.21%	No		0.267	0.461
t − 1 to t + 1	1.17%	1.02%	No		0.028*	0.002**
t − 1 to t + 2	0.89%	0.46%	No	X	0.236	0.106

Note(s): *, ** and *** significant at 5, 1 and 0.1% levels, respectively (Benjamini-Hochberg's (1995) false discovery rate correction)

Table 6

Differences between SA8000 decertified and certified firms

Variable	Median decertified (Group 2)	Median certified (Group 1)	p-value (Mann Whitney)	p-value (Fisher)	p-value (Chisq)
Development	0.69	0.58	0.010**
Economic openness	3.90	3.87	0.334
Munificence	41.90	35.95	0.130
Dynamism	3.70	2.24	0.001**
Competition	0.78	0.61	0.000***
Labour intensity	0.08	0.11	0.047*
Firm size	5.41	5.23	0.093
Financial performance	8.78	7.70	0.761
ISO 9001				0.763	0.800
ISO 14001				0.774	0.889
OHSAS 18001/ISO 45001				0.788	0.865
UNGC				0.106	0.121

Note(s): *, ** and *** significant at 5, 1 and 0.1% levels, respectively

See Chapter 3.3 for a detailed description of the operationalization of each variable

Table 7

Robustness checks

Certification event study
Period	AP median	AP mean	Normality	Skewness	p-value (WSR)	p-value (sign test)
Workforce score (decertified firms – Group 2)
t − 2 to t − 1	−0.48	−0.92	No		0.426	0.524
t − 1 to t	1.34	2.14	No		0.151	0.233
t − 1 to t + 1	3.40	5.03	No	X	0.033*	0.028*
t − 1 to t + 2	5.25	6.87	No	X	0.012*	0.017*
t − 1 to t + 3	6.11	7.95	No		0.003**	0.010*
t − 1 to t + 4	6.62	8.43	No		0.002**	0.003**
t − 1 to t + 5	7.04	8.80	No		0.002**	0.002**
Workforce score (certified firms – Group 1)
t − 2 to t − 1	−0.21	−1.45	No	X	0.541	0.647
t − 1 to t	2.47	3.52	No	X	0.089	0.158
t − 1 to t + 1	4.36	6.24	No		0.016*	0.023*
t − 1 to t + 2	5.68	7.33	No	X	0.007**	0.009**
t − 1 to t + 3	7.20	8.49	No		0.001**	0.003**
t − 1 to t + 4	8.85	9.90	No		0.000***	0.001**
t − 1 to t + 5	10.18	11.32	No	X	0.000***	0.000***

Decertification event study
Period	AP median	AP mean	Normality	Skewness	p-value (WSR)	p-value (sign test)
Workforce score (decertified firms – Group 2)
t − 2 to t − 1	1.24	1.53	No		0.188	0.265
t − 1 to t	0.19	−0.36	No		0.714	0.798
t − 1 to t + 1	−2.67	−3.71	No		0.062	0.079
t − 1 to t + 2	−5.45	−6.68	No		0.008**	0.011*

Note(s): *, ** and *** Significant at 5, 1 and 0.1% levels, respectively (Benjamini-Hochberg's (1995) false discovery rate correction)

Note

1. The term “decertification” is generally used to refer to standards with “third-party” audit (e.g. ISO 9001), while the right word for initiatives without formal auditing procedures (e.g. UNGC) should be “delisting”. However, to be consistent within the article, and not introduce too many different words, we decided to use the term “decertification” also for the UNGC.