Part III - Evolution and Resilience in Sustainability and Food Safety Regimes

7 Human Rights Due Diligence and Evolution of Voluntary Sustainability Standards

Enrico Partiti

7.1 Introduction

Private regulators of social and environmental sustainability such as voluntary sustainability standards (VSS) have proliferated. Private schemes such as the Forest Stewardship Council (FSC), FairTrade, and the Marine Stewardship Council (MSC) define sustainability-related product features and production processes by means of voluntary standards. VSS discipline aspects of production including human rights, labor rights, and environmental impacts ranging from pollution prevention to impact on forests and biodiversity. Like other private governance structures,¹ VSS are characterized by contingency and context-dependency that makes them receptive to critical events² including regulatory developments, even prospective ones. Legislative developments at the national level such as the convergence around criteria of timber legality in EU, US, Australian, and South Korean legislation contributed to align VSS requirements. In addition, the goals of VSS were partially refined toward assessing compliance with national provisions and demonstrating due diligence of legality of timber origin as required by those instruments.³

Not only events and rules at the national level are capable to affect this form of private authority. Transnational private regulation is a vehicle to “harden” voluntary obligations and make them applicable to individuals.⁴ Transnational private regulators are therefore also affected by relevant international soft law instruments in their field of operation.⁵ For standards such as VSS, the emergence of human rights responsibilities of corporations represents a major, albeit understudied, development. The 2011 adoption of the United Nations Guiding Principles on Business and Human Rights (UNGP) affirmed a corporate responsibility to respect human rights throughout business activities, parallel to a State duty to protect human rights and a right for victims to obtain remedies.⁶ As an integral part of the responsibility to respect, firms must perform human rights due diligence (HRDD) to identify, assess, avoid, mitigate, remedy, and report about human rights impacts in their value chains, which include both social and environmental aspects.⁷ Some counties passed legislation making HRDD mandatory. EU rules currently require HRDD in the supply chains of minerals associated with armed conflict in Central Africa, and a proposal for a general HRDD Directive is expected soon.

VSS, like other transnational private regulators, are characterized by considerable flexibility and organizational resourcefulness,⁸ insofar as they are capable to rapidly adapting governance structures, procedures, and content of their standards to better fit their contextual environment. Organizational resourcefulness confers resilience to private regulators, as they can reorganize in the face of change affecting the pursuit of their objectives and withstand discontinuity while adapting to new environments.⁹ Given VSS’ receptivity to soft law and, especially, to (perspective) national legislation, an alignment of standards in line with the responsibilities, processes, and constructs of HRDD is expected to be visible. With respect to norms of responsible business conduct, voluntary standards defining responsible production and sourcing must be aligned with HRDD and its requirements if they are to support certified firms at different levels in the value chain toward compliance with their HRDD responsibilities and emerging legal obligations. However HRDD also directly affects VSS, as they are private organizations with their own responsibility to conduct HRDD. Responsibility, or even legal liability under future legislation, could result from VSS association to human rights impacts caused by certified entities or members or to that which they contributed. VSS must also implement grievance mechanisms in line with the UNGP and perform their own HRDD toward firms with which they have a business relation such as certified and noncertified members.

Studying the evolution of VSS in light of HRDD allows us to better understand the influence of requirements established by public authority on private standards. Is public authority capable to influence transnational private regulation or is it bound to fail? To what extent does the resilience stemming from the capacity of VSS to adapt to change (in regulatory frameworks) allow them to retain their regulatory prerogatives or bring private regulators to (partially) reorient their goals?¹⁰ This chapter focuses on private standards connected to deforestation, conversion of ecosystems, and human rights concerns that certify forest products and agricultural commodities. While it does not focus on one case study, it adopts a comparative perspective to analyze the effects of HRDD on some of the most relevant multi-stakeholder and industry-driven initiatives in this domain. The analysis takes place on the basis of requirements contained in production standards, codes of conduct for members, other documents and policies, and NGO reports. These sources are complemented by fifteen semi-structured interviews centered on the impact of HRDD on standards held with NGO representatives and certification managers from ISEAL and from six schemes active in the domain of timber, palm oil, soy, sugarcane, cocoa, and coffee certification.

This chapter is structured as follows. Section 7.2 situates emergence and evolution of VSS in connection to regulatory crises and a recently changed climate toward certification that also contributed to the demand for mandatory legislation. Section 7.3 explains how HRDD can be seen as an organizational crisis, which could be both an opportunity for VSS to consolidate their regulatory prerogatives and a potential threat in light of the establishment of other risk management tools and initiatives. Section 7.4 illustrates relevant aspects of HRDD for VSS. Section 7.5 discusses how VSS are aligning their requirements and policies to the value chain dimension of HRDD and its engagement dimension. Section 7.6 concludes by reflecting on the refinement of VSS relation of complementary with public rules generated by HRDD and the capacity of public authority to align transitional private regulators to public rules. It also reflects on the resilience of VSS and their capacity to expand their activities to novel domains intersecting with HRDD.

7.2 Regulatory Failures and Transnational Private Regulation

Regulatory crises are a critical moment for private regulators.¹¹ These crises are events of varying scale and scope resulting from the unintended or unforeseen consequences of the design or operation of a regulatory system and its interactions with other systems.¹² A regulatory crisis may pressure the industry to self-regulate to protect reputation and avoid liability but also to preempt more demanding regimes. In the domain of sustainability, private regimes appeared as a response to a regulatory crisis exposing shortcomings in the regulation of global production. Certifications for forestry products, coffee, and other agricultural commodities were established after the collapse of commodity prices and the worsening of deforestation caused by agricultural production.¹³ Labor schemes, such as Social Accountability International and the Fair Labour Association, emerged in the aftermath of extensive campaigns in the mid-nineties exposing sweatshop conditions and incidents in the garment industry.¹⁴ The first wave of biofuel certification (before EU regulatory intervention) was linked to the 2007/2008 food crisis to avoid biofuel production displacing food crops.¹⁵ Many of these VSS appeared in an environment characterized by a lack of binding international frameworks that resulted in the emergence of heterogeneous standards.

Where a crisis is connected to a regulatory failure, it may reverberate on public authority as well, which may expressly support private regimes as a solution. Remarkably, voluntary private regulation in the area of sustainability was often suggested as a possible solution also by the very NGOs that brought up attention to the crisis in question.¹⁶ However, in recent years, the wide acceptance of HRDD and the demands for making it mandatory were accompanied by a growing dissatisfaction from certain civil society organizations about the effectiveness and impact of corporate social responsibility¹⁷ and voluntary initiatives including private standards and certifications. NGOs campaigned for the introduction of mandatory legislation aimed at value chain transparency and mandatory HRDD noting how voluntary private standards failed and that they should only play a very limited function in future instruments.¹⁸ NGOs are also experiencing “certification fatigue” in participating in VSS and offering monitoring functions to ensure that firms comply with the standards – a role that they consider as very resource-intensive and better performed by public authority.¹⁹ In recent years, prominent civil society organizations left the VSS that they contributed to establish. Among several instances, the most visible is arguably that of Greenpeace International leaving the FSC, of which it was a founding member, in 2018 due to the controversies around the “FSC Mix” certificate.²⁰

Some environmental NGOs are also growing frustrated at what they consider as an obstructive attitude of business toward attempts of reform in VSS about transparency, auditing, and stringency of the requirements.²¹ NGOs filed complaints against Bonsucro and RSPO for breaches of international standards for responsible business conduct.²² A recently published report by the Institute for Multi-Stakeholder Initiatives Integrity went as far as concluding that VSS “have peaked” and that they will be replaced by alternative, rights-centered, models of private governance similar to the Bangladesh Accord.²³ While quantifying effectiveness and impact of VSS remains complex and debated,²⁴ certification managers respond to this alleged lack of impact of VSS by noting how certification was never intended to be a “silver bullet” capable to tackle deeply rooted structural problems that can only be solved with the involvement of all public and private actors and mandatory rules. Part of the disappointment among certain civil society organizations would stem from having put too high expectations on certification,²⁵ which should be seen as a complement rather than a replacement of public governance.²⁶ This is also the position of VSS in public consultations and lobbying activities.²⁷

Regulatory crises not only brought private actors together in the establishment of voluntary sustainability regimes but also mobilized civil society and governmental support for private solutions instead of more profound public intervention. This establishes competitive dynamics under which private regimes hinder or delay the emergence of more profound and mandatory public rules.²⁸ Competition arises where private and public regimes fight for legitimacy, uptake, support, the authority to set rules and key terms thereof, or the acceptance of a regulatory regime over the other.²⁹ Competition could result in substitution where public rules are challenged by, or replaced with, private regimes that are less stringent than public regimes or limit their effectiveness, pursue business interests to a larger extent than public goals, or that are ineffective and “symbolic.” The fact that public authorities, at least in the EU, are committed to introduce or have already introduced mandatory legislation on the social and environmental impact of global production therefore stands in contrast to initiatives hitherto enacted on both sides of the Atlantic and grounded on voluntarism and multi-stakeholderism. It arguably testifies to a possible co-optation outcome, where public regulation takes over private regimes, either by turning elements of private regulation into a (mandatory) public regime or by narrowing down the regulatory space for private governance.³⁰

7.3 HRDD as Threat and Opportunity for VSS

From the internal perspective of transnational private regulators, the perception of critical factors or a change in (regulatory) context as threats to the status quo is linked to the notion of organizational crises. An organizational crisis represents a threat for an organization that prevents it from attaining its goals or reduces its ability to do so. Organizations seek to resolve such crises also because they are an opportunity to achieve their goals even further – and beyond the issue in question.³¹ Organizational crises catalyze opportunities to cooperate in new or existing institutions and experiment with alternatives that would not otherwise be considered, resulting in rethinking, reorganization, and new institutional settings.³² The introduction of HRDD, especially in mandatory legislation, from the perspective of VSS can be seen as an organizational crisis.

The goals of VSS do not just include the regulation of sustainability. VSS also pursue institutional goals such as increasing market uptake and gaining legitimacy from their association to legislation.³³ HRDD and mandatory HRDD constitutes an opportunity for schemes to extend their uptake among firms and consolidate their regulatory prerogatives, possibly even in new regulatory domains and through new regulatory tools. As VSS contribute to social and environmental risk management, HRDD could incentivize their use as part of companies’ responsibilities and obligations. The UNGP raised awareness and demand for supply chain transparency that VSS are well placed to provide, by giving firms a tool showing that they “do not harm” and to monitor progress and improvements.³⁴ HRDD also requires firms to engage with their value chains, as further illustrated in Section 5.2, thereby generating a demand for guidance and new institutional forms to that purpose.

In parallel, however, HRDD could push firms to design their own internal due diligence systems for sourcing and tackling social and environmental risks, which could be less stringent and less transparent than private certification, nor based on a multi-stakeholder approach and without third-party assurance mechanisms.³⁵ Private business programs in the context of sustainability supply chain management³⁶ proliferated in sectors covered by VSS. There is evidence that they displaced certification especially in the cocoa space.³⁷ These initiatives cover a company’s entire sourcing and could create fragmentation and ultimately additional burdens for compliance by upstream producers. Scheme managers are concerned with this increased competition by firms’ proprietary systems:

We need to be very clear what is the difference with other [firm-level] schemes. Legal deforestation is not the same as zero deforestation. Third-party certification with accreditation is not the same as one simple, single audit firm certifying every scheme.

In connection to deforestation, alternative forms of private governance have indeed emerged that are not necessarily alternatives to VSS but that could reduce their role in regulating sustainability in value chains. Multiparty pledges such as the Soy Moratorium reduce the relevance of voluntary certification initiatives at least for deforestation-related concerns as they include public enforcement and strong enforcement mechanisms to avoid that noncompliant products are traded. Auditing is also supplanted as a monitoring mechanism by the possibility to use remote sensing and publicly available satellite imageries,³⁸ which are, however, also integrated in VSS under the awareness of the limits of audit systems.³⁹ A respondent from a nongovernmental organization summarized the implications of HRDD as follows:

If [VSS] look at the mandatory human rights due diligence requirements that are increasing particularly in Europe, and realize that they have to lift their game and this is what they’re going to need to do to essentially provide that service for companies so that they can make their human rights due diligence requirements, then that’s an opportunity. If they do it, that’s an opportunity but on the other hand, if they don’t rise to that challenge then companies will decide they’re not an effective tool for human rights due diligence and find other ways to do it …. it will only be a legitimate process for human rights due diligence if the [VSS] and its own process of certification, etcetera, is robust.

As organizations, VSS themselves also bear the responsibility in the UNGP not just to respect human rights but also to avoid associations to human rights violations to which they are directly linked through their commercial relations. Recent dispute resolution before the national contact points (NCP) for the OECD Guidelines confirmed that this can be the case, thereby opening the door to other complaints. In two cases against RPSO and Bonsucro in Switzerland and the United Kingdom, both NCPs confirmed previous practice to expand what they considered as a “multinational corporation” under the OECD Guidelines for Multinational Corporations⁴⁰ – an instrument that expressly operationalizes the UNGP. This notion was interpreted to include other transnational private actors such as NGOs and sport bodies such as FIFA. The Swiss NCP’s involvement in the RSPO case was rather narrow in light of jurisdictional limitations.⁴¹ In the Bonsucro case, however, the UK NCP held that it could be possible for a multi-stakeholder initiative to breach provisions of the OECD Guidelines such as the presence of a human rights policy and the continuous performance of HRDD including the exercise of leverage and mitigation of adverse human rights impact.⁴² Membership was explicitly considered as a business relation directly linking human rights harms committed by a (prospective) member to a VSS.⁴³ The factual assessment of these claims is currently pending after failure of the parties to reach a mutually agreed solution.

This process was described as:

an important wake-up call [for VSS] in the sense that, “Look, we have to be more reactive to this type of thing and we need to have a system where really what we’re asking of our members is broader than just our standard and that certification part of it. It’s that broader alignment with human rights over to the UNGP.”⁴⁴

7.4 Relevant Aspects of HRDD for VSS

The type of due diligence legislation currently discussed in the EU,⁴⁵ with specific rules for agricultural commodities and ecosystem conversion,⁴⁶ will likely require companies to undertake due diligence for all human rights and environmental impacts in the entire value chain. This would reflect the UNGP and sector-specific OECD Guidance documents asking firms to account for the entire adverse social and environmental impact they caused, to which they contributed, or are directly linked through their business relations.⁴⁷ Within mandatory HRDD and with respect to the business responsibility to exercise HRDD the function of VSS must be explained. Firms demand supply chain risk-management tools, to manage risks, ensure conformity, and enhance productivity, reputation, and profitability⁴⁸ and also to ensure respect of legal requisites. The requirements of a scheme must therefore be aligned to HRDD as provided in the UNGP, OECD Guidelines, and the specifications of future regulatory instruments for a VSS to be of assistance in firms’ responsibilities. Where issues and risks covered by a scheme align with those faced by a firm, standards are suitable for integration in that firm’s HRDD processes as a non-dispositive evidence of low risk,⁴⁹ as also done under the EU Timber Regulation.⁵⁰ As a consequence of a possible narrower scope of VSS, compliance with a scheme would not grant a presumption of conformity with legislation but would serve as a rebuttable presumption of “low risk.” This approach has been problematic for VSS in the timber legality space, as certified firms were disappointed that the cost of certification did not lead to opening up market access and ensuring legal compliance in the EU.⁵¹

The limitations of VSS must be clear. Firstly, HRDD responsibilities include all possible human rights affected by business operations,⁵² but VSS may have a narrower human rights scope. For example, FSC does not generally refer to all human rights in its Principles and Criteria but, as its focus lies on forest operations, it covers human rights affected by forest management operations such as workers’, customary, community, and Indigenous Peoples’ rights.⁵³ The effectiveness of VSS as risk mitigation tools depends also on the extent to which a given social and environmental concern or harm can be detected. This links to the vexed question of whether certification is an effective mechanism to verify and ensure that the scheme’s criteria are implemented properly.

Secondly, HRDD responsibilities apply throughout the entire value chain. From the perspective of a downstream firm marketing in a jurisdiction with (future) HRDD legislation, HRDD must identify, mitigate, and remedy possible risks and harms all the way upstream. Furthermore, a downstream firm could be implicated in adverse impact through a producer from which it sources both certified and noncertified material and in which human rights violations occur in the context of noncertified volumes. Different would be a scenario where harm occurs in a production unit whose products are not traded or marketed by downstream firms. While some NGOs are keen to expand the possible responsibilities of downstream firms,⁵⁴ there would not be a “direct link” with adverse impact through business relations. A similar situation would arise where harm is generated by a subsidiary or associated entity of a firm with which the downstream firm does business but with which there is no direct relation. However, from the perspective of the human rights responsibility of that upstream entity, there would be association to human right harm. This situation is particularly challenging for VSS’ own HRDD, as they would be certifying entities causing or contributing to human rights violations and also breaching their standards in noncertified operations.

However, most requirements of schemes apply to the level of harvest, plantation, and unit of production. Even where the entire value chain must be certified under forms of chain of custody certification, intermediary entities such as mills, plants, and processing facilities are rarely requested to comply with requirements concerning environmental impacts and human rights. A 2019 comparative study of FSC and PEFC’s principles and criteria and chain of custody requirements⁵⁵ concluded that the forest management standards are aligned to the ILO Fundamental Conventions, the UNGP, and the OECD Guidelines, and these are assessed by auditors. However, the chain of custody requirements include only compliance with ILO Conventions and are limited to an indirect reference to FSC Policy for Association and a self-declaration of compliance by firms.⁵⁶ Scheme holders consider that the purpose of chain of custody standards is to assure credibility of claims, and therefore human rights issues are not necessary, also because they consider risk to lie at the farm level.⁵⁷ Furthermore, requirements for nonproducing members have traditionally been lower than those for producers in order to attract downstream firms to participate.⁵⁸ However, as Section 7.5 illustrates, these features of VSS are changing.

7.5 Impact of HRDD on VSS

Some VSS are attempting to build in respect for human rights, including an obligation to perform HRDD, within membership requirements applicable also to downstream actors and retailers, that is, the noncertified members. The emergence of HRDD responsibilities for upstream entities is also visible in the expanded criteria for certification to noncertified volumes and entities. Some schemes are introducing the principle that producers whose only part of their operations is certified cannot breach key requirements in the noncertified areas or production units. For intermediary supply chain actors such as mills, this approach results in extending criteria to all sourced volumes, thereby transmitting upstream a request for certification-compliant production. Additional human rights criteria are also appearing in chain of custody certification. This expands the substantive obligations so that the scheme covers broader supply chain segments for the downstream firms and therefore better aligns with HRDD requirements. Section 7.5.1 discusses these developments concerning VSS regulatory activities. Section 7.5.2 focuses instead on the recent expansion of novel forms of nonregulatory activities centred on value chain collaboration and engagement.

7.5.1 Expanding Requirements

FSC established in 2011 a Policy for Association that attempts to extend FSC standards beyond certified operations. Certificate holders, certification bodies, partners, or members associated with FSC can be disassociated if responsible – as a company or because of activities of subsidiary companies or subcontractor – for violations of key criteria including illegal logging or trade, destruction of high conservation value forests, significant conversion of forests to planation, GMO use, violations of traditional rights, human rights, and breaches of ILO Core Conventions.⁵⁹ The Policy of Association is enforced by FSC and is part of the due diligence performed by the organization. FSC is attempting to move beyond a self-declaration for prospective members toward actively performing due diligence about whether firms are involved in unacceptable activities under the Policy for Association.⁶⁰

Between 2015 and 2017, FSC attempted to reform the Policy for Association by formalizing how it performed due diligence. The proposed policy provided that prospective members will be subject to additional screening in high-risk cases. Data is collected via a self-assessment complemented by stakeholder input.⁶¹ If there is evidence of violations, the matter would be dealt with under the procedure to process complaints against breaches of the Policy for Association, with a Complaint Panel that will make recommendations about association. Disassociation should take place only in the presence of repeated violations, as the organization prefers to address violations through cooperation given that dissociation would not produce positive outcomes for forests.⁶² Revision and expansion of the Policy for Association has been complex, and an attempt to strengthen its enforcement by FSC, as well as clarifying when a member may breach the key criteria,⁶³ produced no result and had to be put on hold.⁶⁴ Another attempt for revision started in 2020. Through its Shared Responsibility policy, RPSO included the requirement that all members such as NGOs, banks and investors, retailers, manufacturers, processors, and traders must respect human rights, especially free prior and informed consent, in their entire operations and have grievance mechanisms in place.⁶⁵ In this way, RSPO supports respect for human rights and the performance of HRDD within its membership requirements for downstream firms.

Similarly, Bonsucro scaled-up its membership requirements through a Code of Conduct. In March 2020, Bonsucro aligned it to the UNGP and the OECD/FAO Guidance for Responsible Agriculture Value Chains by recognizing human rights responsibilities of members in relation to their suppliers. The Code therefore requires certified and noncertified members (i.e., respectively, mills and all other supply chain actors) to commit to continuous improvement, respect human rights and protect natural ecosystems, embed this commitment in operations, and communicate progress.⁶⁶ By incorporating the UNGP’s concept of “direct link” to human rights harm that determines the boundaries of companies’ responsibility to respect, the Code applies also to products and services linked to sugarcane production, processing, and sourcing.⁶⁷ The self-assessment performed by members requires them to improve compliance of their production, processing, and sourcing. As under the UNGP,⁶⁸ the expected commitment of members varies according to the risks at hand and the nature and size of operations.⁶⁹ Members are also expected to provide remedies to adverse impact, including via operational-level grievance mechanisms and remediation in line with UNGP Principle 31.⁷⁰

Bonsucro ensures compliance with its Code of Conduct not by including its requirements in audits but via reporting appraised by the organization. Action plans may be requested in case of noncompliance, with the possibility to refer to Bonsucro’s Grievance Mechanism.⁷¹ Bonsucro also acknowledges their own responsibility toward members through risk assessment.⁷² An enhanced due diligence of members was introduced, assessing their social and environmental risk. The process entails online searches, consultation of court records in the country of operation and with other organizations that may possess information about relevant social and environmental impacts of the perspective member, and comments by interested parties on the basis of which the level of risk and expected actions are determined.⁷³ The process may lead to additional requirements imposed on the (candidate) member. In a recent case, Bonsucro engaged in discussion with local stakeholders and used its leverage to require a prospective member to establish corrective plans including disengagement with suppliers breaching human rights.⁷⁴ Where allegations were raised about the involvement of another candidate member’s with forced evictions of indigenous communities, Bonsucro engaged with different stakeholders and ascertained that, while the candidate was not directly involved, some of its suppliers might have been responsible. Bonsucro thus requested in the action plan the implementation of risk management systems, a requirement of continuous dialogue, and a disengagement strategy.⁷⁵ This is in line with the UNGP requirement that leverage should be exercised as much as possible, and disengagement should only take place where leverage failed to achieve results.

Concerning the expansion of certification requirements to noncertified volumes and organizations to account for the human rights responsibility of the firms at hand, some VSS are expanding the human rights requirements applicable in their chain of custody certification. Some schemes “don’t want to create a Chain of Custody Standard that is covering human rights issues. The purpose of that standard is to assure credibility of claims.”⁷⁶ Other VSS are instead broadening the applicable requirements under chain of custody standards. FSC recently incorporated core labor standards into auditing requirements in chain of custody.⁷⁷ As they become part of the audit criteria, this update strengthened enforcement of human rights provisions that would otherwise only be covered under FSC Policy for Association and its self-assessment. In addition, the expansion of human rights in chain of custody allows downstream entities sourcing FSC-certified products to receive assurance of low risk of at least certain human rights violations in the entire supply chain. In a similar manner, various additional social and environmental requirements have been introduced in Rainforest Alliance’s 2020 version of its supply chain standard.⁷⁸

Bonsucro is introducing additional rules for certification of mills and processors concerning noncertified volumes. In the current standard, the supply area included in the unit of certification comprises the farms supplying cane in conformity with Bonsucro requirements. Where this is less than 100 percent of the supply, a respective percentage of production is considered as certified.⁷⁹ In fact, mills on average select an area to certify that represents only 23 percent of the mill supply, and the production standard applies only to that area. Bonsucro is revising its standards to introduce a system where human rights requirements apply to the entire mill supply, including areas neither controlled by the mill nor certified but that are managed by smallholders whose certification is complex and where environmental and social risks lie.⁸⁰ These requirements include enacting “sustainability policies” to respect human rights, mapping vulnerable stakeholders, and assessing risks.⁸¹ While auditing is limited to assess whether sustainability policies and other requirements are int place, these new criteria – if implemented successfully by mills – are capable to expand the reach of human rights standards under the HRDD responsibility of the mill. The standard would therefore acknowledge that mills’ responsibility extends beyond certified volumes and includes all entities to which they are directly linked via their sourcing activities.

Also RSPO similarly introduced in the 2018 revision of its Principles and Criteria requirements that mandate the entire unit of certification, in all its business operations and transaction, to have a policy to respect human rights at all value chain levels.⁸² Other standards focusing on GMO such as ProTerra require mills and processors employing inputs from noncertified farms to design and implement supply chain control systems to ensure that core GMO and social and environmental indicators are met.⁸³ While these requirements avoid commingling of GMO and non-GMO materials, where human rights requirements are monitored, the standard also covers the entire human rights responsibility of the mills concerning the farms to which it is directly linked. Verification of supply at the farm level is undertaken over a five-year period through third-party audit.⁸⁴

In light of HRDD responsibilities, VSS are spurred to reflect over their own due diligence structures.⁸⁵ “Some of these elements were already considered good practices, but now from this lens of mandatory due diligence, these issues are going to become probably more important.”⁸⁶ In addition to a more thorough screening of prospective members discussed above, where schemes are not detecting noncompliance with certain criteria in spite of the presence of a high risk, they may be expected to carefully assess whether audits are working properly and whether they should not employ additional venues to have access to information, or establish complaints and grievance tools to accede to it.⁸⁷ In the 2020 standards revision, UTZ-RA strengthened mechanisms to collect geospatial data complemented by remote sensing and baseline mapping to supplement auditing in determining whether land conversion occurred. In this way, the scheme already knows which farms present a high risk of past deforestation and will inform auditors about possible concerns.⁸⁸

Generally, VSS also have to tackle the unintended adverse impact stemming from compliance with their standards. While certification may make visible existing conflicts,⁸⁹ in other cases, the standard may generate adverse impact or may have to balance between different types of harm. To lessen the negative environmental impact of burning sugar cane by farmers, standards may contemplate requiring increased mechanization. This may however impact on human rights of the workforce. Bonsucro includes this type of risk into its risk management systems. As a form of impact mitigation, the scheme and the certified entities offer retraining programs, and information is shared with other certified firms that aim to increase mechanization and could generate similar harms.⁹⁰

In the context of deforestation-related criteria and their possible human rights implications, a scheme is in the process of implementing enhanced definitions of forests and covered ecosystems for the entire production of farmers and mills located in the Brazilian Cerrado independently from whether certain areas will be certified or destined to certified processors and mills. Both definitions and elements of verification system will build on those provided by the Accountability Framework initiative (AFi),⁹¹ a global benchmark for deforestation claims across value chains established by NGOs and aligned to the UNGP. Benchmarking criteria therefore affect VSS substantive requirements. The effects of AFi on other definitional elements that scale-up standards’ formal requirements are also visible. The 2020 revision of Rainforest Alliance’s production standards has explicitly incorporated AFi’s approach centered on non-conversion of forests and natural ecosystems and has embedded relevant concepts and definitions.⁹² Other standard-setters are in the process of including aspects of AFi relevant to their schemes in their requirements, and others – such as the Responsible Leather Roundtable – already use AFi definitions.⁹³

7.5.2 Collaboration in Risk Mitigation and Remediations

Human rights abuses and social conflicts within value chains are likely to endure without collaboration and engagement among all stakeholders involved.⁹⁴ HRDD requires collaboration between downstream firms and upstream entities.⁹⁵ Collaboration is also essential with non-business stakeholders and human rights holders to ensure mitigation of impacts and remediation. Collaboration may require investment in value chain mapping and transparency and even supporting upstream producers. Firms should avoid risk-adverse behavior such as disengaging from noncompliant suppliers (which may in fact aggravate the situation for human rights) or stop sourcing from high-risk areas.⁹⁶ Finally, HRDD also requires that human rights violations, where they occur, are remedied and the status quo is restored, a requirement that is more easily fulfilled through collaboration.

Engagement can take various forms, such as committing to higher wages or purchase volumes, longer-term contractual relations, as well as investment by downstream firms to improve working, social, and environmental conditions upstream.⁹⁷ The provisions of economic incentives and tools to improve productivity intend to remove some of the economic drivers of social and environmental harm. At the same time, engagement ensures change on the ground and avoids that certification creates segregated markets where compliant products are sold in Western markets and noncompliant produce is sold elsewhere. However, this principle is complex to operationalize as companies lack the knowledge and incentives to actively engage. Engagement may be burdensome, requiring cost-sharing and direct financing. The allocation of costs and responsibilities remains unclear and contestable unless fairness considerations are incorporated to offset frequent downstream firms’ exploitation of their suppliers.⁹⁸

While offering risk mitigation and remediation through collaboration is the responsibility of firms, requiring engagement has proven difficult for standards, but this is an area where VSS are increasing their focus:

That’s where you need organizations to take companies by the hand, bring them together, and say, “Look, this is where your investment is going to go.” The great thing about a mandatory law would be that then they actually have to put in that investment, they can’t just walk away, but you’re still going to need that glue between companies to actually do something in a more collective sense. I think standard system to a certain degree, they can be that glue although, again, they are still maybe not active enough in that space to think about what this really comprehensive mitigation or remediation look like. Some of them have developed mechanisms and to a certain degree, just the noncompliances in the standard, you can build investments around it …. Anyway, the interesting role or where I definitely hope that standard systems will play a greater role is exactly in that wider remediation and mitigation space where it’s not just that they provide some information about which producers are compliant or noncompliant, but they can actually provide an entry point for companies that have to invest in mitigation and remediation because their inspections are linked to those problems.⁹⁹

Beside their traditional multi-stakeholder structures, VSS are introducing or strengthening collaborative features traceable to the UNGP requirement of engagement between upstream and downstream in mitigating and remedying impacts. This pathway confirms the intuition of those suggesting a reframing of VSS functions, one less concerned with authoritative rule-making and more centered on assistance in broader practices concerning sustainable supply chain management.¹⁰⁰ One area where collaboration was enhanced concerns the amount of certified products that entities downstream commit to purchase. Purchase commitments ensure a steady demand for certified products that guarantees price premiums for producers to undertake the necessary investments.¹⁰¹ Certification managers raised concerns about the “magnificent claims” made by downstream companies joining an initiative while only sourcing a limited amount of certified products and receiving a positive image return.¹⁰² While some schemes tackle this issue by strengthening their rules concerning claims about certified sourcing,¹⁰³ some organizations took structural steps to actively engage chain actors and require downstream entities to provide support to farmers in facilitating compliance with the standards. This allocates responsibilities for the costs of sustainability, mitigates the risk of social and environmental harm, and possibly scales up impact to areas and products also not sold in Western markets.

RSPO introduced in 2019 a “Shared Responsibility” policy applicable to all members but is particularly relevant for noncertified members. Similar to the FSC and Bonsucro cases discussed above, RSPO members comply with a Code of Conduct requiring them to implement requirements for their own entire organization that must align with the RSPO standards.¹⁰⁴ This broad requirement was expanded though the notion of Shared Responsibility defining the commitments for collective action, collaboration, and accountability needed to transform palm oil markets toward more responsible outcomes.¹⁰⁵ Members must comply with common principles and policies, support small farmers, raise awareness, and offer training as well as technical and personnel support to RSPO. The most salient aspect concerns the identification of volume targets for buyers. Manufactures and retailers commit to purchase an extra 15 percent of certified palm oil in the first year of implementation, while traders and processors have a 2 percent target. This commitment from downstream firms matches the commitment from farmers to comply with more stringent standards.¹⁰⁶ Members must report on their purchase commitments, which are independently verified and included in the audit under the Chain of Custody Standards. Systems for sanctions and incentives are currently being discussed, as well as the provision of financial contributions to support small-holders.¹⁰⁷

A similar approach was introduced by RA-UTZ in the 2020 standards revision to ensure that risks, costs, and benefits of sustainability transformations are evenly distributed between producers and buyers. The new standard introduced a “sustainability differential,” i.e. a price premium to certified producers to recognize farmers’ efforts and to support sustainable production. While a price premium is already present in schemes such as FairTrade, buyers of certified products under RA-UTZ certification are also required to make (and report about “sustainability investments” necessary to enable farmers in their value chains to comply with production requirements or the cost of audit and on the basis of investment plans designed by certificate holders themselves.¹⁰⁸ Both sustainability differential and sustainability investment are paid by the first buyer, included in the sale contract, and recorded in RA’s traceability system.¹⁰⁹ Bonsucro’s Implementation Guidelines of the 2020 Code of Conduct also require continuous commitments and improvements that can be demonstrated by sourcing increasing percentage of certified material or supporting suppliers toward certification.¹¹⁰

Remediation for social and environmental harms has also been introduced in recent iterations of RA and RSPO standards. Since 2014–2015, RSPO requires certified members that engaged in noncompliant land clearance for plantation or other facilities after the 2005 cutoff date to use remediation and compensations mechanisms. Members will have to designate protected areas to offset previous conversion.¹¹¹ If land-use change impacts on the human rights of affected communities, social remediation and compensation plans must be negotiated with right holders.¹¹² In the 2020 revision of RA standards, in line with HRDD, a separate protocol was introduced requiring remediation and offering guidance on how to effectively remediate human rights violations.¹¹³ A strong remediation guidance is also included in AFi,¹¹⁴ which was developed building on the UNGP requirements.¹¹⁵

Also VSS’ growing involvement with integrated jurisdiction and landscape management shows an extension beyond individual producers as a unit of analysis, in combination with strong elements of engagement with and among public and private actors. As the eradication of adverse social and environmental impacts requires addressing structural issues with the involvement of all relevant public and private actors,¹¹⁶ VSS are supporting efforts in specific jurisdictions to identify smallholder lands and establish district-level multi-stakeholder governance structures to monitor, report, and verify land-use change. These approaches are part of a trend linking private initiatives with REDD+.¹¹⁷ However, the focus of landscape and jurisdictional initiatives moves beyond certified producers and aims at structurally involving other supply chains actors including the financial industry.

While VSS have already been cooperating within landscape and jurisdictional initiatives for a few years,¹¹⁸ some are beginning to offer jurisdiction-based certifications. RSPO has finalized a second consultation on a “Jurisdictional Approach to Certification” that aims to establish, in partnership with public authorities, a stepwise process toward granting certification against RSPO standards to an entire jurisdictional organization.¹¹⁹ ISEAL recently released a code of good practices applying not to standard systems but to both landscape and jurisdictional initiatives that wish to make credible claims about their activities and to other initiatives developing frameworks for landscape and jurisdictional projects.¹²⁰ While recognizing the potential of these initiatives, ISEAL stresses that they remain complementary to current supply chain tools like standards systems, which are capable to verify and incentivize specific sustainability improvements at the farm level.¹²¹ Also here, the perceived risk for schemes is that firms may decide to source from certain landscapes or jurisdictions, giving up certifications.¹²²

7.6 Conclusion

As public authority intervenes in the regulation of sustainability and human rights across value chains, both through soft law and via mandatory rules, isomorphic pressures among private certifications generate convergence among VSS requirements and approaches, with VSS increasing also their nonregulatory activities. A visible trend among private schemes is the expanded application of key requirements to noncertified volumes and firms to account for the human rights responsibilities of entities at different levels of the value chain. As a form of transnational private governance applicable to the firms and producers that wish to comply with their standards, VSS are complementary to international and national provisions in the social and environmental domains.¹²³ This happens by design, so that VSS can be used by firms to demonstrate compliance and manage social and environmental risks. If VSS want to retain their complementarity, they must adapt to emerging public requirements. This allows schemes to better fit in firms’ HRDD systems as they cover risks for a broader number of value chain entities. VSS themselves are enacting enhanced due diligence and risk management procedures to account for their own HRDD responsibilities vis-à-vis possible human rights impacts by members and certified firms.

With emerging obligations of HRDD, schemes are no longer competing with public rules in a transnational space or deterring their emergence. As public regulators step up the regulation of responsible business conduct, it will be public requirements that determine what represents sustainable or responsible conduct across value chains. This process can be seen as a co-optation of VSS where VSS do not anymore independently define sustainable practices but operationalize detailed requirements of what constitutes HRDD that are emanations of public authority both at the international and at the national/regional level. In this context, the implementing functions of VSS in transposing legal obligations in the social and environmental domains are diminished in autonomy. Therefore, public intervention is capable to effectively align transitional private regulators to public rules. This could be seen as an instance where public authority has been capable, if partially, to get a handle on economic private activism.¹²⁴ However, HRDD spurs VSS to account for impacts of various entities associated to them, thus further expanding the application of their standards to more firms across value chains. The extension of this form of indirect public control is therefore counterbalanced by an increased relevance of VSS in the supply chain they govern.

The impact of HRDD on VSS is linked to its double nature of opportunity and threat for VSS, in line with the notion of organizational crisis discussed in Section 7.3. HRDD gave VSS the possibility to leverage their organizational resourcefulness to engage in new activities and establish new institutional features. By strengthening their efforts in the area of engagement and collaboration between firms at different levels in the value chain, VSS function is also expanding and partially realigning. By providing standards and associated services, VSS also increasingly engage in nonregulatory activities such as offering fora for engagement for risk mitigation, remediation, and sharing costs of social and environmental compliance required by HRDD. This does not fully shelter VSS from the possible threat stemming from other alternative tools for HRDD. However, it creates a novel goal to which VSS are arguably well placed to contribute. The capacity of VSS to expand their activities to new nonregulatory domains despite the influence of public authority on their regulatory function also testifies to their resilience.

8 The Politics of Collaborative Governance in Global Supply Chains Power and Pushback in the Bangladesh Accord

Juliane Reinecke and Jimmy Donaghey

8.1 Introduction: Collaborative Action and Transnational Governance

On April 23, 2013, the Rana Plaza building complex collapsed claiming the lives of over 1,100, mainly women, ready-made garment workers and injuring many more. This was just the latest in a series of fatal factory fires and collapses that killed hundreds of garment workers in Bangladesh, many despite factories being audited against international private accountability standards, including some of those in the Rana Plaza complex. The collapse also demonstrates the failure of prevailing business practice to deal with labor issues in the global supply chain based on individual social auditing by international brands of factory compliance against accountability standards and codes of conducts. To date, the main approaches to transnational labor governance have been either the development of business-driven codes of conduct¹ or agreements between individual multinationals and international trade unions in the form of global framework agreements.² However, as discussed elsewhere, neither solves the chimera that is transnational labor governance. In this chapter, we focus on a novel transnational governance initiative that was developed after the Rana Plaza crisis: the Bangladesh Accord for Building and Fire Safety.

In common with a central theme of this volume, a central feature of this chapter is the role that crises played in determining the dynamics of transnational private governance. Drawing on the work of Louis Althusser, we view the idea of crisis as “a critical moment” for institutions where “we’re uncertain and don’t really know whether the crisis [or “situation” to avoid being tautological] will culminate in death or rebirth.”³ While not seeing crises as necessarily leading to such stark binary outcomes, we share the idea that the defining feature of a crisis is a period of extreme uncertainty and this uncertainty may force actors to make decisions that mark a departure from previous practice in order to emerge from the uncertainty. In this context, the chapter focuses on two primary crises that were central to the dynamics of the Accord. First, the role that the very crisis nature of the Rana Plaza disaster played in developing the collective approach that drove the Accord. Second, once established, the Accord itself faced a crisis in the form of a conflict between the transnational level governance of the Accord and the national regulatory system of Bangladesh. While the Accord had many novel features,⁴ in this chapter, we focus on one particularly novel aspect of the Accord – the nature of the collective action created through the Accord – and highlight both the institutional design features that were at its foundations but also how the collective action was highly political in nature and why ultimately this contestability of its actions became a weakness in the end.

The supply chain model, where both buyer and supplier firms are involved in multiple contracts and implementing multiple private governance standards, has led to fragmentation of labor governance.⁵ Thus, as globalization has deepened and multinational corporations seek to reduce labor costs, private transnational labor governance has emerged as a key area of focus where MNCs seek regimes with low labor standards while simultaneously wanting to avoid association with the most egregious labor abuses.⁶ This private governance often emerges in response to pressures exerted by civil society actors, such as unions and NGOs, particularly when crises emerge.⁷ In this context, neither public governance nor individual actors have the capacity to provide meaningful governance. In this chapter, we examine an example of how collaborative governance, in the immediate aftermath of a crisis, helped to overcome governance gaps inherent in prevailing practice that emphasizes individual actions but also how the collective action literature needs to be more cognisant of the politics within collectives and those outside the collective.

A collective action dilemma describes a situation where the action of individuals leads to a lack of investment or resources being overexploited unless an external authority, typically the government, intervenes to regulate access.⁸ However, due to structural changes in the global economy, where states often compete against each other for investment based on how liberally regulated their economies are (or not), the possibility of government regulation is slim. Elinor Ostrom suggested an alternative to either regulation, polycentric systems of governance, in which formally independent centers of authority interact to make allocation, regulatory, and sanctioning decisions.⁹ In such a system, private actors may seek to avoid the “tragedy of the commons” if governance systems are designed to tie in actors to a collective approach. In this chapter, we take the Bangladesh Accord as an example of collaborative governance and explore the key features that prompted its efficacy as a governance mechanism. However, while certainly the collective action features of the Accord had a positive effect on workplace safety in the Bangladesh garment sector, our findings highlight that the political consequences of collective action by private actors require greater attention from transnational governance scholars.

The argument developed here is that by understanding worker safety as a collective action problem, a more robust approach to global labor governance may be achieved, depending on the prevailing circumstances. However, while such an approach may be effective in terms of improving some governance areas, we highlight that the private power generated by collaborative governance to overcome the collective action problem made it a highly politically contested approach. This calls for greater attention to the dynamics of contestation and how these can affect the operation of collective approaches. We illustrate our arguments by examining our empirical research on a key response to the Rana Plaza tragedy, the Bangladesh Accord for Building and Fire Safety (Accord),¹⁰ from which wider lessons for labor rights in global supply chains may be drawn. The research is based on a six-year longitudinal project (2013–2019) with a total of 140 interviews conducted in Bangladesh and buyer countries with supplier companies, buyer companies, trade unionists, NGOs, and other related parties.

8.2 Worker Safety as a Collective Action Problem in Apparel Supply Chains

The apparel sector has become the epitome of how globalization is driving down labor standards. Single brands have neither the willingness, incentive, nor leverage to make individual mechanisms work, as the failure of social auditing demonstrates. Wary of exposure to negative publicity by mainly Western NGOs, labor activists, and unions, brands have heavily invested in social auditing of factory compliance to corporate codes of conduct, which accounts for up to 80 percent of their ethical sourcing budget.¹¹ Yet, while social auditing has helped companies manage their supply chains, safeguard individual reputations, and claim social responsibility, it has largely failed to improve working conditions.¹² Shortcomings have been tragically demonstrated by the failure to prevent a series of fatal industrial accidents: Rana Plaza famously housed two factories, Phantom Apparels and New Wave Style, which were audited against the Business Social Compliance Initiative’s (BSCI) standard.¹³

Critics have long argued that social auditing is primarily designed to limit buyers’ legal liability and to manage reputational risk, rather than improving working conditions¹⁴. In theory, the threat of sanctions – withdrawal of orders – is meant to encourage suppliers to address noncompliances. In practice, social auditing creates a collective action dilemma of its own. Even in cases of major noncompliances, there is little follow-up by buyers and contracts are rarely terminated, not least because this would create disruption to the supply chain and/or increase production costs for buyer firms.¹⁵ This renders the threat of an individual buyer withdrawing orders ineffective. In turn, the knowledge that individual action is likely to make little difference reinforces a buyer’s incentives to keep “eyes wide shut,” even if suppliers are found noncompliant with a company’s code of conduct. Companies and their suppliers then both have an interest in hiding labor violations rather than reporting them. The result is corporate complicity in a system where multiple buyers re-monitor the noncompliances of their supplying factories, leading to duplication of audits and “audit fatigue,” yet without significant remediation taking place.

This failure has implications for the interests of all actors and at all levels in the supply chain in the apparel sector:

• At the brand level, even though safe and sustainable factories carry collective benefit for the entire industry, buyers face individual disadvantage when pursuing costly sustainable actions, as such costs may not be borne by competitors. Brands are competing against each other on a low-cost model and competitors may have an incentive to free-ride based on safety upgrades that another firm has made. Without assurances that all buyers invest in safe and sustainable factories, it is difficult to convince individual buyers to take action due to this risk of free-riding.

• At the supplier level, competition between factories both from within the country and from other countries is cutthroat due to low entry barriers. Suppliers compete with each other on a cost-basis and push downwards pressure on workers in terms of wages, safety, and many other labor issues regimes.

• At the worker level, as long as an unlimited supply of labor is ready to take up work in garment sector, workers lack market power to demand safer workplaces. Efforts at developing a collective voice in the form of trade unions face strong resistance from employers and often government officials alike. As a result, out of over 4,500 officially registered garment factories in Bangladesh, only about 10 percent have registered unions. But according to local estimates, many fewer are functioning properly due to the immature system of industrial relations, fragmentation of unions and lack of organizing capacities.

• At the supply chain level, highly complex global webs of purchasing relationships involve multiple buyers sourcing from multiple suppliers with parties spreading their relationships across geographical spaces to minimize risk. The combined effect has been increasing fragmentation, multiple tiers involving subcontracting, and an overall lack of transparency. Fragmentation has produced a situation where, in the absence of state oversight, brands have not taken responsibility for labor standards within the factories that produce the goods which they sell.

• At the host government level, competition among sourcing destinations for low cost production drives down standards and wages. Bangladesh for many years has been the world’s second-largest exporter of ready-made garments, though has now slipped to third, and it faces increasing competition from neighboring Myanmar, Cambodia, Laos, and increasingly East Africa. The Bangladeshi government has generally resisted effective regulation out of fear that compliance will increase production costs and decrease the competitiveness of this sector that makes up 80 percent of exports in a highly competitive and mobile global market.

These contrasting interests made the issue of worker safety much more complex and certainly makes the question of what is the problem to be solved far from straightforward. In the Accord, for international brands and the trade unions, the shared problem they were seeking to solve through a collective approach was how they could establish and enforce a governance system to make Bangladeshi factory owners improve the safety in their factories. In this way, what they were seeking to establish was a new system that required the greatest change from an actor other than those actually involved in the design of the governance system. Thus, the collective action was one of trying to force other actors who each had their independent agency to change their behavior and invest resources, rather than the Accord actors themselves shifting their approach. Therefore, and unsurprisingly, the collective action became one of intense internal and external politics in terms of the creation and implementation of the Accord. While the internal politics played a role in both making the Accord more encompassing and probably more effective,¹⁶ as will be developed below, the external politics, particularly in terms of those who were the subjects of the collective action, ultimately led to the ending of the Accord.

8.3 Labor Governance in the Bangladesh RMG Sector

Since the 1980s, Bangladesh grew to become the second largest exporter of garments after China, growing from growing from $12,000 in exports in 1978 to annual exports exceeding $21.5 billion and 13 percent of GDP at the time Rana Plaza occurred. Despite Rana Plaza, the sector still grew to $34 billion by 2018¹⁷ and has been credited with creating employment for around 4.1 million people directly, about 65 percent of whom are women, and 5 million workers indirectly.¹⁸ At the time Rana Plaza occurred, workers only earned a minimum wage of $38, which increased to US$68/month in 2013 and US$95 in 2019. Despite these low wages, employment has contributed to the country’s rapid economic development and poverty reduction, halving halved the percentage of people living under the $1.90 poverty line since 1991.

The Bangladesh RMG sector epitomizes the global supply chain model, both in terms of the economic development and job opportunities it brought to the country, as well as in terms of the lack of regulation and exploitation of labor. It is a prime example of how the hypercompetitive market dynamic of globalized, industrial capitalism has overwhelmed the traditional bulwarks against capitalist exploitation – protective labor market institutions of the state and trade unions – and shifted the human cost of cheap fashion to the most vulnerable element of the chain – the garment worker in developing countries. Bangladesh is also an example of the “competition state,”¹⁹ where states compete for inward investment through cheap labor and lack of public regulation to keep prices low. As a result, public regulation has largely failed to protect workers’ rights in the Bangladeshi ready-made garment industry as the Bangladesh Labour Law and Building Code have remained largely unenforced. At the time Rana Plaza occurred, the government’s labor inspectorate had fewer than 100 inspectors for more than 24,000 factories across all industrial sectors, 3 million shops and 2 major ports, including the over 4,300 export-oriented garment factories.

In terms of labor relations, Mark Anner²⁰ describes Bangladesh as “despotic market labor control” where workers lack market power alongside ineffective state protection. Since Rana Plaza, the government has publicly criticized efforts to increase unionization in the ready-made garment sector. A hostile context for trade unionism, low density, lack of unity with thirty-four union federations in the garment sector alone, an immature system of industrial relations, and political corruption point to the limitations of traditional labor regulation in the sector. Following a change in the labor law post–Rana Plaza, the International Labour Organization²¹ reported a rise in factory-level union registrations to 437 by March 2015 out of at least 4,500 officially registered garment factories. Yet, according to the AFL-CIO Solidarity Center in 2019, only 200 were still active with many fewer functioning properly due to both employer resistance and lack of union organizing capacities and, with international pressure subsiding, the government rejected 73 percent of applications for new union registrations in 2015.

Weak labor power and hypercompetitiveness have not only depressed wages but investment in factory safety. With uncontrolled growth of the RMG sector, residential and commercial buildings in densely populated urban areas were repurposed for industrial use, often by adding additional stories without permission as in the case of Rana Plaza. As the industry grew, so did the number of industrial disasters: before Rana Plaza, more than 600 Bangladeshi garment workers had died due to unsafe buildings since 2006, often despite factories having been certified by reference to CSR auditing standards. Sixty-four were killed in the Spectrum Sweater factory disaster in 2005, 21 killed in the Garib & Garib sweater fire in 2010, 117 killed by the Tazreen factory fire in 2012, and eventually over 1,100 died and a further 2,000 got severely injured in the Rana Plaza collapse in 2013.

8.4 The Bangladesh Accord for Building and Fire Safety

The Accord was signed in the immediate aftermath of the Rana Plaza building collapse in 2013 to develop a more robust approach to worker safety in the Bangladesh ready-made garment industry. The Accord grew out of an earlier attempt to create a “Memorandum of Understanding” following the Tazreen fire that killed 112 workers but never came into force as only two brands, PvH and Tchibo, had signed up.²² The idea behind the Memorandum of Understanding was that brands would take a collective approach to developing greater worker safety in an initiative jointly managed with labor actors in the form of NGOs and unions. It was only, however, after the crisis of Rana Plaza that a number of labor actors including the Global Union Federations, IndustriALL Global Union, and UNI Global Union, as well as the labor rights NGOs Clean Clothes Campaign and Workers Rights Consortium were able to pressurize a critical mass of brands into signing up to this five-year, legally binding agreement between brands and unions that was unprecedented in global supply chain governance.

The Accord, at its peak in 2017, had been signed by 215 signatory companies, global and local unions, with labor rights NGOs as witness signatories. Its main governing body is the Accord Steering Committee, which is composed of three brand representatives and three union representatives with the ILO acting as its independent chair. Signatory companies include global brands such as H&M, Inditex, C&A, Primark, and Hugo Boss; large Western retailers such as Aldi, Carrefour, or Tesco; as well as a number of smaller apparel brands. The Accord was never intended on being a permanent approach to the governance of worker safety: it grew as a crisis response to Rana Plaza but was always viewed as a mechanism that could be leveraged to develop nationally based institutions. The original Accord had a five-year duration and expired in July 2018, to be replaced by the “Transitional Accord” to carry on the work with a proposed term of two to three years, depending on the progress made in terms of establishing a Bangladeshi-based alternative in which brands and unions had confidence. This, however, as will be outlined below was effectively brought to a premature end by court action in Bangladesh.

The Accord set out as its goal “to enable a working environment in which no worker needs to fear fires, building collapses, or other accidents.” Program activities included:

• Rigorous program of fire, electrical, and structural safety inspections carried out by trained Accord engineers

• Monitoring remediation progress and facilitating brand support for remediation

• Online publication of all Corrective Action Plans (CAPs)

• Mobilizing collective brand leverage through escalation and termination of business relationships with nonparticipating supplier factories

• Training program for joint labor-management safety committees

• Accord Safety and Health Complaints Mechanism to resolve safety complaints

By signing the Accord, company signatories made legally binding commitments that were enforceable in the national courts in the country within which they were registered. Thus, while the initiative was essentially a private governance mechanism, national regulatory mechanisms were to be the ultimate enforcer of the agreement. The legal commitments included financial responsibility for funding the Accord’s activities above. In terms of the operations of the Accord, the Accord employed its own engineers and other members of its secretariat who acted to implement a unified approach to safety across all factories who supplied its members. This was specifically designed to reduce the multiplicity of different codes of conduct where minor differences between multiple codes and multiple inspections by each brand meant that suppliers lacked a clear approach in terms of their buyers requirements.

The success of the Accord is most clearly demonstrated by improvements to safety, reducing workplace accidents. The average remediation rate had reached over 84 percent by the time the first Accord expired in April 2018, rising to 93 percent by December 2020 (see Figure 8.1).

Figure 8.1. Remediation progress

Source: Accord, 2020²⁴

The Accord found more than 80,000 safety issues in its first round of inspections of 1,100 factories. Critical issues were found in almost each factory often despite having previously passed multiple social audits. As of 2017, 74 percent of identified safety issues have been reported or verified as fixed,²³ such as fire proofing the electrical wiring, installation of fire doors and fire systems, as well as redistributing weight loads and strengthening the factory building’s columns.

The Accord illustrated how a collaborative approach by brands and unions had, in the context of a crisis, the potential to generate leverage to change the system through collective action. What features made this approach effective? Eight dimensions helped to establish the institutional conditions that made the Accord a successful example of collaborative labour governance. These will be divided into institutional design for collective action and operating principles for collective action.

While the term “institutional design” is often used in a loose way, in the Accord, there was a very deliberate “design” adopted as a mechanism of developing the Accord as a governance approach. These four design features of the Accord both enabled the development of a collective approach and gave other actors assurance that their interests would be represented through the governance structures. The four design features as follows:

Transnational Co-determination

A central feature of the Accord was that it was a jointly governed initiative by representatives of business and labor: worker representatives, and not just representatives of capital, were included in the design and oversight of the transnational labor governance regimes. Rather than other initiatives where business organizations co-opted NGOs and the like onto their bodies in advisory-type capacities, in the Accord, equal status of business and worker interests within the institution underscored its purpose. Two Global Union Federations and six local Bangladeshi unions are full signatories with 50 percent voting rights on the Accord Steering Committee. In addition, four NGOs (Clean Clothes Campaign, Workers Rights Consortium, International Labor Rights Forum, Maquila Solidarity Network) are “witness signatories” who enjoy observer status only. The International Labour Organization (ILO) acts as neutral chair.

Rather than promoting common business interests in protecting reputation²⁵, inclusion of recognized labor representatives meant representation of interests of the agreement’s intended beneficiaries: garment workers. Through such a collective approach, different parties were able to cooperate and build a consensus through which problems were solved, even if their individual interests were in competition or the definition of the problem differed. In this way, the Accord was built around the idea that while interests may be mutual, they are not necessarily shared. Mutuality is built upon the recognition that, at times, the interests of the parties involved, typically workers and managers/owners, may not be shared but that a common solution can create mutual benefit despite divergent interests.²⁶ An obvious actor missing from the Accord as an employment relations agreement are the Bangladeshi employers. For this reason, the Bangladesh Garment Manufacturers and Exporters Association (BGMEA) and Government of Bangladesh fiercely contested the quasi-regulatory authority of the Accord, leading to a High Court case which will be discussed later in the chapter.

Industry-wide, Pre-competitive Collaboration

Tragedies such as Rana Plaza have demonstrated that an industry’s reputation is a shared resource, subject to reputational spillovers.²⁷ In collaborative approaches, encompassing interest groups thus need to collaborate to achieve collective action and sanction free-riding. In this aspect, under the Accord, brands effectively were accepting the need for pre-competitive collaboration as a way of removing from competition issues of collective concern, such as labor rights. Collective action then spreads the cost of economic adjustment, increases sanctioning capability, and reduces the incentives for free-riding.

Legally Enforceable Commitment

By signing the Accord, company signatories make legally binding commitments that were enforceable in the national courts in the country within which they are registered. Due to its legally binding nature, the Accord departs from voluntary CSR standards. The Accord created an enforceable contractual relationship in the home country of the buyer brands. The point about the establishment of a legally enforceable contract was a very significant new departure in global supply chain labor governance. Through the Accord, brands transferred oversight of their supply chain to a body that had a right to initiate legal action against the brands where they did not meet their commitments. While heavily resisted in the past, a legally binding agreement was achieved due to the pressure placed on brands by the harnessing of the complementary capacities of labor rights NGOs and trade unions. This legally binding nature was tested when two cases were filed in October 2016 at the Permanent Court of Arbitration in The Hague by IndustriALL Global Union and UNI Global Union to hold two unnamed signatory companies to account for failing to meet Accord terms ²⁸ – in particular, to require their suppliers to complete their remediations and to agree on commercial terms that are financially feasible for their suppliers to cover the remediation costs. The first case settled in December 2018. The second case settled in January 2018, involving agreed payments of $2.3 million to cover remediation in more than 150 garment factories in Bangladesh ($2 million) and pay into IndustriALL and UNI’s joint Supply Chain Worker Support Fund ($0.3 million).²⁹

Developing Worker Voice

Central to the Accord, and pushed by the labor caucus, was the idea that having a transnational apparatus was not enough: the governance of factory-level safety needed to have worker participation at all levels. Worker voice was thus seen as central to the Accord as it recognizes the potentially competing interests of management and workers over core organizational issues³⁰. As such, the Accord oversaw the development of a more comprehensive structure of worker voice in the area of workplace safety, including joint worker-management safety committees in all factories and a robust complaints mechanism. By December 2020, the Accord Safety and Health Complaints Mechanism had resolved over 693 complaints raised by workers and ensured the creation and training of over 1,260 Joint labor-management Safety Committees. The training foresees a central role for workers and worker representatives, including direct trade union participation in factory training and factory inspections.

Alongside these design features of the Accord, four operating principles became essential to the establishment of the collective approach.

Leverage through Collective Action

The Accord creates collective leverage through the combined power of its corporate signatories. Unlike codes of conduct or International Framework Agreements, which are generally agreed between one MNC and a GUF, the Accord covers multiple brands (in excess of 200 in the first Accord). Collective action by a large proportion of buyers provided far greater leverage for effective sanctioning than any buyer would have individually. As expressed by a buying brand: “If you don’t remediate you lose your orders from 215 brands. That’s leverage, that’s how you get things done in Bangladesh.” Under Accord rules, signatory firms agreed to terminate contracts with factories that failed to make safety improvements. Facing the loss of orders from not just one but a large group of buyers committed the factory to invest in remediation. Effective sanctioning led to the most unsafe factories being temporarily or permanently shut, with remediation efforts monitored in almost all other factories, potentially saving the lives of thousands. In 17 factories safety concerns were so severe that the Accord recommended immediate evacuation, and immediate remedial actions were necessary in another 110 factories. While not all factories were covered by the Accord, approximately half of all workers in the sector and most of those in directly exporting firms are covered by the Accord. The collective brand approach of the Accord also brought benefits to supplier firms in Bangladesh by having a unified set of standards, rather than suppliers attempting to satisfy a multiplicity of codes of conduct for different buyers.

Accountability through Collective Oversight

Collective oversight over inspections meant it was possible to overcome some of the limitations of previous auditing approaches, such as lack of transparency. By including unions, the ILO, and NGOs in oversight, as well as placing factory reports and Steering Committee minutes in the public domain, buyers were incentivized to act on noncompliance. Parties without profit rationales, such as unions and NGOs, can expose firms who seek to circumvent the collective approach. Transparency also reenforced collective leverage, because even brands not covered under the Accord were less likely to source from factories that have found to be unsafe.

Pooling of Resources

The Accord brought brands together to pool resources and share costs, information, responsibility, and risk. Pooling of resources increased marginal per capita return, which incentivizes participants because they know that their individual contribution made a bigger difference.³¹ With industry-wide contributions, a collective safety mechanism funded higher-quality inspections with engineering teams specializing in fire, electrical, and structural safety. Companies assume responsibility for funding the activities of the steering committee, safety inspectors, and training coordinators based on their annual volumes of garment purchases from Bangladesh on a sliding, pro-rata scale up to $500,000 per annum. With industry-wide contributions, a collective safety mechanism was argued to fund a program of high-quality inspections and remediation monitoring with engineering teams specializing in fire, electrical, and structural safety. To illustrate the scale of the task, by the time the first Accord expired in April 2018, Accord engineers had carried out a total of 25,656 follow-up inspections in a total of 2,055 factories. This yielded 134,489 findings, with safety hazards present in each and every factory, such as lack of fire escapes. Respondents repeatedly stated that another large-scale disaster would have been imminent without immediate intervention. Pooling of resources was argued to overcome the deficiencies of single-brand approaches such as lack of expertise, under-funding of specialized inspections, and protocols for follow-up action and remediation. Cost sharing was also viewed as making governance more accessible especially for smaller buyers with limited resources and further reduces incentives for free-riding.

Highly Focused Approach

One of the criticisms of the Accord was also one of its strengths. The Accord had a narrow focus on building, electrical, and fire safety. In order to build an agreed approach encompassing so many brands, as well as union and NGO actors, meant such a narrow focus was necessary as maintaining agreement is easier. This also meant that a highly specialized approach was enabled to be taken by the engineers employed by the Accord in terms of implementing an agreed common standard across all factories who wished to supply for any Accord brands. While preventing fatalities within the industry, one criticism though was that it has done little to increase poverty wages or extend worker rights beyond safety. In contrast, wide-ranging approaches such as the UN Global Compact on the other end of the spectrum have been criticized for achieving few of their many wide-ranging objectives. Similarly, global framework agreement negotiation by GFU with MNCs cover a wider variety of industrial relations issues but are often less able to deliver in terms of the expertise required and monitoring involved. Focusing on a clear and tangible problem enabled the Accord to concentrate actions and resources on delivering more effective problem solving.

8.5 Contesting the Regulatory Power of Collaborative Governance: The Bangladesh High Court Case

A theme of the chapter to this point has been the central role that crisis played in initiating collective action and in developing the Accord. Without doubt, the Accord’s focus on creating structures and processes for collective action was a key feature that explains its relative success compared to other private governance initiatives. While the work on collective action by the likes of Ostrom is attractive in explaining the emergence of collaborative governance initiatives, two particular problems are worth highlighting in relation to our case. First, the literature on collective action is underpinned by a rational-economic mindset. In contrast, we highlight that a more political approach to developing collective action is necessary. Take, for example, Ostrom,³² where she argues that “collective-action problems occur when it takes the inputs and efforts of multiple individuals in order to achieve joint outcomes – and it is difficult to exclude beneficiaries of these actions from benefiting even if they do not contribute.” Here Ostrom is working from the perspective that the problem is actually relatively easy to identify but that the difficulty arises over who pays the price for such action. While problems such as free-riding have long been acknowledged as potential problems in collective approaches,³³ such an approach ignores the political nature of organizational life where, not just the solution or cost of that solution are contested but the very nature of the problem to be solved is one that is contested.

Secondly, by definition, collective action is about bringing parties together and including them in a decision-making process. However, who is included or excluded is often a cause of political tensions within any cooperative arrangement.³⁴ A central feature of this literature is that collective action is viewed as parties coming together in order to solve shared or mutual problems. However, such a picture paints an idealized picture where all parties are equally invested in solving the problem. As we will develop, while there was significant unity to form a collective initiative, one of the reasons for forming the collective initiative was to take action against parties who were perceived as being the root cause of the problem for which a solution was to be developed: Bangladeshi employers, factory owners and government cast as “negligent” in providing workplace safety. The Accord thus also created a separate but related crisis for the industry and government in Bangladesh. The latter perceived the Accord to be a threat to their competitive advantage – cost leadership – in the international garments sector. This was a substantially different crisis due to the nature of the interests involved. Our argument is that the role of interests and how they relate to “problems” that require solving is central to understanding the dynamics of collective action in transnational labor governance. In the context of the Accord, we will highlight that in terms of its creation, implementation, and de facto termination, the nature of collective action in the Accord was highly politicized in nature.

Within the literature on collaborative governance, there is an underlying assumption that parties cooperate to solve a shared problem and generate “win-win” solutions. While the “solutions” to such problems may be contested, the idea is that those participating do so with a clear understanding of what the actual problem is. However, this approach implies that there is one objective and identifiable factor that can be identified as the problem that requires fixing. As outlined above, for both the MNCs and the labor actors in the form of unions and NGOs, the problem was the state of repair of the Bangladeshi factories and the main obstacle identified as causing that problem were Bangladeshi factory owners. This, however, is an oversimplification and this came to be highlighted through the Accord. In our interviews with factory owners and managers in Bangladesh, a consistent picture emerged that they felt extreme pressure from the MNC buyers to drive down their prices. As we have argued in our research elsewhere,³⁵ problems that multinational corporations seek factory owners to remedy often arise out of the pressures associated with the sourcing model that the multinationals impose on their suppliers. Thus, factory owners in Bangladesh did not share the same understanding of the problem.

Initially, the opposition of factory owners and the government did not prevent the Accord from operating. With both the MNCs and unions viewing the Bangladeshi employers as the problem, the employers were excluded from the governance arrangements of the Accord. As such, they were clearly placed in the role of subjects of the initiative. This proved to be very controversial for both the employers and the government of Bangladesh. In our research, the metaphor of “being a guest in someone’s house, and then telling the host to change your house” was frequently raised by employers and managers in Bangladesh. Thus, since the start of the Accord, there was much opposition from industry and government actors in Bangladesh about not cooperating with the Accord. However, having the collective buying power of up to 220 brands from Europe, North America, and Southeast Asia meant that the vast majority of factory owners allowed the Accord to inspect their premises. Without doubt, though, there was a feeling among Accord actors that some factory owners in Bangladesh were “dragging their heels” in terms of implementing the remediation to “ride out” the period until 2018 when the initial Accord was due to expire.

A second feature of the Accord was that the fire, electrical, and particularly structural aspects of complying with the Accord meant that a significant cost had to be paid to raise the standards of factories in Bangladesh: under the approach of the Accord, while brands had a duty to help suppliers through loans or advance payment for orders, ultimately it was the suppliers who were expected to pay the cost. While the brands did face a cost in terms of being Accord members, which was to pay for the operational costs of the Accord, this was capped at an annual maximum of US$500,000. In contrast, costs associated with remediation were to be covered by factory owners that could run substantially higher. The IFC in 2016³⁶ estimated that total remediation costs for factories covered by the Accord would amount to ~US$403 million, or between US$120,000 and US$320,000 for the majority of factories (80 percent). As Accord signatories, brands were obliged to “ensure” that it was financially feasible for their supplier factories to cover these costs. But brands insisted that “ensuring” by no means meant covering the costs. Thus, the collective action in this case was not simply a positive sum game but rather was a zero-sum game with the actor responsible for paying the cost being the subject of governance rather than being a participant in it.

This exclusion and cost model caused considerable resentment within Bangladesh. The government, the industry bodies in the form of the BGMEA and Bangladesh Knitwear Manufacturers and Exporters Association, and individual employers engaged in a series of pushbacks against the Accord. From the outset, the government opposed the Accord and portrayed it, and the Rana Plaza disaster itself, as part of a conspiracy to undermine the lucrative RMG sector in Bangladesh. Both the prime minister and the minister for commerce were vocal in their opposition, but more significantly, stating that the Accord would not continue to operate past its initial five-year term ending in 2018. While there was much rhetoric from the political class in Bangladesh opposing the Accord, state actors did not take direct action to stop it from operating, presumably due to the dependence on the sector for exports.

However, ultimately it was the actions of the state, through the Bangladeshi court system that effectively ended the role of the Accord in Bangladesh. In this case, a factory owner was deemed to be noncompliant with the Alliance, an alternative system to the Accord but one that crucially shared findings and had a system of mutual recognition of noncompliant factories with the Accord. Thus, the factory owner was excluded from supplying either Accord or Alliance brands. After undertaking some remediation works, the factory owner applied to the Accord to be deemed compliant but the Accord deemed that the factory still had not met the required standards. In response, the factory owner applied to the national body for non-exporting factories to be deemed compliant. This was forthcoming and the factory owner went to court claiming that as his factories were compliant with the national body, his factories should be deemed compliant with the requirements of the Accord. In response to the case, the Supreme Court of Bangladesh took the unusual step of opening up a suo moto case on the entire legality of the operation of the Accord in Bangladesh.

In this case, the Supreme Court of Bangladesh, when ruling on the operation of the Accord, argued that in 2013, when Bangladesh was facing an emergency situation, the Accord taking on some of the responsibility of the Bangladeshi state was acceptable. But by 2019, when the Transitional Accord came into effect, this emergency state was no longer in existence. The Supreme Court’s decision ruled that the authority of the Accord to end contracts was based on the Accord being focused on implementing the National Action Plan and placed the Accord on the same footing as the less stringent National Tripartite Plan. In addition, the Court ruled that where a factor was regarded as being safe by the National Tripartite Plan, under Bangladeshi law, the Accord could not deem it as being unsafe. The Court then went further and argued that given the national emergency of 2013 had now passed, the Accord was given 281 days to end its operations within Bangladesh.³⁷ This judgment prompted a series of negotiations opening up between the Accord, its constituents, the BGMEA, and the government of Bangladesh. This ultimately brought an end to the operations of the Accord in Bangladesh in April 2020 and established a new governance body, ostensibly run in its operations by the BGMEA but with BGMEA, brand, and union representatives making up its governing body.

Much of the literature on collective action and also the literature on joint problem solving is one where parties seek to take joint action on a commonly held problem. What happens though where the collective action is effectively to get another party or parties to make changes that they see as contrary to their interests? This became the key political question in the case of the Accord. A key feature of the design of the Accord was that it was essentially designed to force Bangladeshi employers to take the substantive action to improve labor conditions in their factories. A second source of pushback came from the Bangladeshi state for whom the Accord raised issues on two levels. First, the Bangladeshi government viewed the Accord as impinging on is right as a sovereign government to govern within its jurisdiction. Secondly, the Bangladeshi government, as being heavily reliant on the garments sector for exports, had a key interest in maintaining the competitiveness of the industry in Bangladesh. Ultimately, the exclusion of these groups and how they framed the problems and responsibilities for remedying safety issues in the Bangladesh RMG sector initiated considerable resistance from those outside but subjected to and affected by the operation of the collective action of the Accord.

8.6 Conclusion

The Accord was an unprecedented example of how collaborative governance can generate collective action in global supply chains and solve a collective action problem – despite a lack of reinforcing institutional support, even resistance, from the host government. Without doubt, the crisis following the Rana Plaza disaster enabled the development of a collective approach that has halted the deadly incidents in the factories. The nature of the crisis played a central role in its success by bringing together a plurality of interests and having inbuilt mechanisms for ensuring accountability through participation of unions and NGOs. Viewing governance as a collective action problem helps improve our understanding of the complex institutional arrangements that can contribute to collaborative governance within complex global supply chains. A number of key lessons can be drawn. First, the collaborative approach enables a system that provides collective leverage between participants but also against those who may seek to free-ride. Secondly, collective oversight from an independent body provides a mechanism through which diverse actors can be guided down a common path. Third, the inclusion of multiple interests enables the identification of mutual solutions to interests affecting particular parties. A final and more general lesson to be drawn is that, in the current neo-liberal environment, individual action and intense competition is often elevated above the benefits of cooperative and collaborative behavior. Taking individual action as illustrated by the social auditing model adopted by many supply chains creates an inferior regime in terms of health and safety governance. The Accord demonstrates that collective responses have a central role to play in developing meaningful and sustainable governance mechanisms that can deliver common solutions despite competing interests.

However, two more cautionary tales emerge in terms of collective action and the nature of private governance. The first is that private governance, regardless of the parties who participate, operates in the shadow of public governance. This relationship between the public and private sphere is one that has attracted increased attention in recent years³⁸ but ultimately private initiatives can be overridden by public institutions. In the case of the Accord, the Bangladeshi state, through its High Court, took a decision to reassert national sovereignty over a transnational governance initiative: the Bangladeshi state certainly faced a legitimacy crisis in terms of its ability to defend its main export industry. Obviously, all those brands could have chosen to exit and withdraw their sourcing from Bangladesh in protest against the ruling, thus threatening nearly the entire export base of the country. But ultimately brands’ commercial interests triumphed. However, this does not mean that the Bangladeshi state was able to claw back on economic private activism entirely. A plethora of private social auditing regimes are still in place, requiring buyers to comply with the codes of conduct of their suppliers. In this sense, the very feature that endowed the Accord with its regulatory power – stringent enforcement through the collective leverage of signatory brands – was also the one that rendered it vulnerable to state backlash.

Secondly, collective action in this case was inherently political in nature and steeped in the extreme power asymmetries of the global supply chain. While a problem was identified that few could have argued was not a problem (the deaths of workers in Bangladeshi garment factories), the root cause of it was highly contested (negligent employers and factory owners or a “broken” supply chain where brands’ squeezing of suppliers creates downward pressure). As such, the “solution” of devising a private governance system that excluded the Bangladeshi employers while placing on them the main burden of paying the cost of the Accord and associated remediation, created an oppositional force. Employers had an inherent interest to bring the initiative to an end. Thus the creation of collective inclusion also implicitly meant the creation of exclusion and thus “problem solving” actually became a highly contested territory.

Finally, the case of the Bangladesh Accord contains important insights about the episodic nature of crises in transnational governance. While the crisis of the immediate aftermath of the Rana Plaza disaster for brands sourcing from Bangladesh played a key role in shaping the collective action by brands, unions, and NGOs, by creating a hitherto unmatched level of inspection and enforcement, a crisis was created for the Bangladeshi government and the factory owners. In response, they took action through public institutions that pushed back against the private governance. At time of writing, the dynamics are still playing out with efforts being made to recast the Accord with a more international focus, demonstrating the ebbs and flows of the interface of national regulation and transnational governance, as well as the dynamic effects which the episodic nature of crisis responses may have.

9 The Evolution of the Global Food Safety Initiative The Dynamics of the Legitimacy of a Transnational Private Rule-Maker

Tetty Havinga and Paul Verbruggen

9.1 Introduction

The Global Food Safety Initiative (GFSI) is an industry-driven meta-regulator that has proven to be influential in shaping both industry practices and public policy making in the field of food safety across the globe. GFSI makes a particularly promising case for investigating how a transnational private regulator has changed over time and shown resilience in overcoming deep-rooted legitimacy challenges (or “crises”) to its private authority. In 2000, a group of leading international supermarket chains founded the Global Taskforce on Food Safety, Quality and Security “to set voluntary standards for food products sourced by retailers around the world.”¹ Twenty years later, GFSI has become a leading global business organization in the food industry, promoting “continuous improvement in food safety management systems around the world”.² That goal is principally pursued by the assessment or, as GFSI calls it, “benchmarking” of private food safety certification schemes that regulate food production and processing to a common, global industry standard. In 2021, over 150,000 certificates from 12 GFSI-recognized food safety certification programs had been issued in 162 countries and numerous food safety experts participated in its committees and events.³

In our analysis of the evolution of GFSI, we will discuss its governance structure, its activities, and its framing. The lens through which we assess that evolution is the concept of legitimacy understood in terms of institutionalization theory. To survive, any organization needs social acceptability and credibility.⁴ In other words, they require legitimacy from those actors that have an interest in their activities. For private, non-state regulatory organizations such as GFSI, building and maintaining legitimacy is of vital importance.⁵ In the spirit of this book, one can say that this labor is an essential means to harness endogenous and exogenous forces that undermine the organization’s regulatory effectiveness and thus lead to its crisis, in order to adapt and demonstrate resilience.⁶ Whereas traditional state authority is often taken for granted, non-state regulators “must obtain legitimate rule-making authority … from salient constituencies in their organizational field.”⁷ Legitimacy provides justifications and a shared understanding of what is acceptable or appropriate. GFSI needs the cooperation of other parties in order to achieve its regulatory goals. More specifically, GFSI depends on the participation of retailers, certification program owners, food manufacturers, food producers, and others in the food supply chains, as well as on the support of authoritative organizations and persons, whether public or private. These actors constitute different legitimacy communities with their own values, priorities, and interests.⁸ Changes in these traits challenge the legitimacy of GFSI, put it at risk of organizational crisis, and thus require from it a response of resilience to find a new equilibrium allowing for the continuation of its activities.

We will argue that GFSI has evolved as a transnational private rule-maker through continued processes of pluralization of its constituents, increased transparency, ratchetting up of food standards’ quality, and globalization of its benchmarking activities. In these dynamics, we suggest, GFSI has sought to respond to criticisms and changing demands of its legitimacy communities. GFSI has evolved from a relatively limited retailer-led initiative into a leading actor in the field of global food safety. The GFSI benchmarking requirements are regarded as top of the bill and both national and international governmental organizations have accepted GFSI as a reliable industry partner in policy debates on food safety. In short, GFSI has expanded on many fronts as a transnational regulator, thus widening and deepening its legitimacy basis. Despite its growth and diversification in terms of (board) membership, GFSI has not changed its initial objectives: it still is an industry-led organization, of which the constituents are food safety experts of large food corporations. We will show that many of the changes the organization has gone through can be interpreted as a response to the opposition voiced by internal and external actors in the light of developments in the field of food safety and its certification.

Our account of the evolution of GFSI is mainly based on the content analysis of publicly available documents and on the research we carried out on GFSI before. We studied GFSI publications, newsletters, and press releases. More in particular, we analyzed how the governance structure of GFSI and the benchmarking requirements changed between 2000 and 2020. We also verified what activities GFSI has undertaken over the years and how it frames its mission. In addition, we use information from interviews with two member of the GFSI Global Regulatory Affairs Working Group and other interviews that we conducted in earlier research in the field of food safety governance.⁹

We start this chapter by discussing the concept of legitimacy as developed in the work of Richard Scott and of Julia Black.¹⁰ Subsequently, we will introduce GFSI as a transnational private rule-maker. This introduction lays the ground for the analysis of GFSI’s evolution, in which we focus on the processes of pluralization, transparency, globalization, and ratchetting up of standards’ quality. We will show how these processes have contributed to the gaining and maintaining of legitimacy, and we will analyze how GFSI has sought to manage the conflicting interests involved and be resilient to these dynamics.

9.2 Dynamics of Legitimacy in Transnational Private Rule-Making

Legitimacy, it has been recognized in academic literature on regulation and governance, constitutes a necessary attribute of any actor in the successful pursuit of regulatory goals.¹¹ Most scholars of regulatory governance would agree that legitimacy essentially turns on “the acceptability and credibility of the organization to those it seeks to govern.”¹² For lawyers, that may imply that a legal mandate is bestowed upon the regulator by a recognized state authority, such as a Parliament or a court, and that the actor operates within the bounds of that mandate.¹³ Political scientists may stress the democratic representativeness of beliefs, expectations, and interests of those affected by the regulatory activities or the effects and costs of these activities on the attainment of the regulatory goals.¹⁴

Private rule-making at the transnational level presents a challenging case for the construction of legitimacy. In that domain, traditional state-based mechanisms of democratic legitimacy are generally absent. Legitimacy is not self-evident and must be socially developed by the regulatory organization in relation with those that are sought to be governed.¹⁵ In the absence of such legitimacy, the risk of regulatory failure is apparent and can lead the regulator into organizational crisis. In the same vein, Bernstein and Cashore stress the “political legitimacy,” that is, the general support for a regime or governance institution, which transnational private governance systems require to be effective.¹⁶ In their view, it “requires institutionalized authority (whether concentrated or diffuse) with power resources to exercise rule as well as shared norms among the community.”¹⁷ In this sociological conception, legitimacy is an empirical observation that results from the acceptance of the organizations’ conduct by others. Legitimacy, then, is dynamic and may come and go with the change of legitimacy demands from those in and outside the regulatory regime, such as regulated entities, regulatory intermediaries, and beneficiaries.¹⁸

Richard Scott has provided a powerful and widely accepted theoretical framework to capture the dynamics of institutionalization and legitimation. In his framework, institutions are built on three pillars: regulative, normative, and cultural-cognitive systems.¹⁹ Related to these three pillars, he discusses three general mechanisms that may lead to institutionalization: (i) increasing returns stressing the role of incentives and interests as motivating force, (ii) increasing commitments stressing the role of identity and mutual social relations, and (iii) increasing objectification of shared beliefs embedded in routines and forms stressing the role of ideas (Table 9.1).²⁰ In line with this triptych, Black distinguishes between three sets of reasons for organizational legitimacy.²¹ “Legitimacy may be pragmatically based: the person or social group perceives that the organization will pursue their interests directly or indirectly. It can be morally based: the person or social group perceives the goals and/or procedures of the organization to be morally appropriate. Finally, legitimacy can be cognitively based: the organization is accepted as necessary or inevitable.” Black argues that the degree of resilience of organizational legitimacy of a transnational private regulator differs for the three types of legitimacy: pragmatic legitimacy is least resilient, cognitive legitimacy is most resilient, normative legitimacy is in between.²²

Table 9.1. Three pillars or reasons for institutionalized legitimacy

Pragmatic	Moral	Cognitive
In line with economic interests	Goals perceived as morally appropriate	Accepted as inevitable
Increasing returns	Increasing commitments and norm generation	Increasing objectivation of shared beliefs embedded in routines
Strategic / cost-benefit calculations	‘Thick’ institutionalization in procedures, routines, rituals, and symbols	Objectivation of shared believes in a community
Low resilience	Intermediate resilience	High resilience

As Scott stresses, “robust institutionalization is often the product of multiple mechanisms that interact with and reinforce each other.”²³ The three distinctive mechanisms or sets of reasons may be perceived as phases in an institutionalization process: first, cost benefit calculations are dominant. Subsequently, “thick” institutionalization may take place creating social entanglements by hardening rules and procedures and creating rituals and symbols. In a final stage of institutionalization, shared beliefs are objectified resulting in the assumption that things could not be otherwise. In line with this thinking, Bernstein and Cashore have shown for non-state market driven (NSMD) governance systems that political legitimacy is constructed in a three-phase process with different relationships between the actors and participation of different actors.²⁴ They distinguish between three phases in the process of gaining political legitimacy:

• the initiation phase, early support of a small number of firms based on strategic calculations;

• the phase of widespread support, gaining support from firms whose practices are further away from the NSDM requirements, norm generation begins; and

• the phase of political legitimacy, participation in shared community, strategic calculations occur within, not about, NSMD systems. The systems that were included in the study of Bernstein and Cashore had not reached this phase yet.

These three phases echo the triad of institutionalization mechanisms flagged by Scott and the reasons for organizational legitimacy discussed by Black. The early stage of NSMD governance systems is characterized by a logic of consequences and calculations based on self-interest; later on, a logic of appropriateness and normative motivations become more important. Bernstein and Cashore elaborate in detail the building of political legitimacy for what they call NSMD.²⁵ Key elements are the conflicting interests between firms and NGOs and that NSMD governance systems deal with global problems that firms have little incentive to address, such as fair trade or sustainable forestry. As these issues are not key to GFSI (NGOs are not involved in GFSI and food safety is a major concern for most companies in the food industry) this specific elaboration is not relevant for our study. What is relevant, however, is the question of whether we can recognize similar mechanisms of institutionalization of legitimacy in the evolution of GFSI.

Casey recently applied the thinking of private regulators as responding to legitimacy demands in an empirical study of the GlobalGAP, the world’s most widely adopted private food safety certification program, which has also been benchmarked by GFSI. Casey contends that for GlobalGAP “early institutionalization of structures, practices and processes have a significant influence on legitimacy by cementing the distribution of power within an organization and crystallizing a dominant organizational logic.”²⁶ We will assess whether and how the distribution of power and the dominant logic within GFSI has changed since its initial phase.

In our comparative analysis of the evolution of GFSI, we focus on changes in its governance structure, its activities, and its framing. We will see how these changes in the distribution of power, in the benchmarking requirements, in the kind of activities undertaken, and in the dominant narrative have contributed (or not) to the construction of legitimacy and to what extent they can be attributed to legitimacy demands.

9.3 The Rise of GFSI

GFSI was launched in 2000 against the background of the proliferation of multiple overlapping and competing private food safety standards initiated by retailers and food firms in the aftermath of the BSE crisis and other incidents.²⁷ In the 1990s, leading supermarket chains such as Tesco and Sainsbury in the United Kingdom and Albert Heijn in the Netherlands, started to develop their own comprehensive quality assurance schemes specifying detailed requirements for their suppliers. These retailers wanted to reduce risks and liability costs and inspire confidence for consumers.²⁸ Private retail-driven food safety standards have expanded ever since. Food retailers collaborated and created industry standards such as the British Retail Consortium (BRC) Global Standard (1998), EurepGAP (1997),²⁹ and the International Food Standard (2003).

GFSI was established to promote globally accepted food safety standards. The original taskforce consisted of thirteen European-based supermarket chains predominantly from the United Kingdom and France.³⁰ Two and a half years later, the taskforce had grown to fifty-two members, mostly from Europe.³¹ The retailers wanted one or a limited number of global food safety standards that they could ask their suppliers to meet. This should produce significant cost savings for retailers and suppliers by reducing the number of food safety audits. Rather than developing one single normative standard of its own, GFSI benchmarks rival standards following a set of requirements laid down in the GFSI Guidance Document in order to coordinate, converge, and ratchet up existing standards. The first version of this meta-regulatory standard was published in 2001. The GFSI Benchmarking Requirements are frequently updated, with Version 2020 being the eight edition. In December 2022, twelve certification program owners have earned GFSI recognition, and three government-owned voluntary certification standards were found “technically equivalent.”³² The GFSI Benchmarking program has become a powerful tool in the global food market, because many major supermarket chains and food manufacturers require their suppliers to be certified against a GFSI-recognized standard. Accordingly, we have suggested, GFSI should be seen as an industry-driven meta-regulator that has proven to be influential in shaping both industry practices and public policy making.³³

9.4 GFSI’s Organizational and Regulatory Evolution

The evolution of GFSI from its inception can be studied through four distinctive processes: (1) pluralization of its constituents, (2) increased openness and transparency of its governance and rulemaking activities, (3) expansion and ratchetting up of food standards’ quality, and (4) globalization of its benchmarking activities.

9.4.1 Pluralization of Its Constituents

GFSI started as a taskforce within CIES – The Food Business Forum, a membership organization of major food retailing companies and their suppliers.³⁴ The taskforce consisted exclusively of supermarket chains. Its membership expanded from thirteen European retailers in 2000 to fifty-two retailers in 2003. In 2004, a board was formed consisting of eight retail members. In 2005, GFSI transformed its organization structure from a membership organization to a not-for-profit foundation under Belgium law managed by the CIES,³⁵ which later merged into the Consumer Goods Forum (CGF).³⁶ It was CIES/CGF who took the final decisions on important governance issues for GFSI, including its structure, strategy, and membership.

The board of the separate legal entity GFSI was dominated by retailers in those first years. Gradually, this power distribution has changed. In October 2008, GFSI announced that it has created a new governance structure for the board. In addition to retailers, from that moment on food manufacturers and food service companies were also awarded membership on the noard. Two vice-chairs of the board are appointed from each of these two categories.³⁷

In 2020, the board consists of eleven retailers, ten manufacturers, and two food service providers.³⁸ So, retailers lost their majority. All retailer and manufacturer board members are from CGF member companies,³⁹ which are prioritized in the board and other GFSI committees.⁴⁰ In 2021, the board was replaced with a steering committee to align GFSI with the governance model of the CGF.⁴¹ In the steering committee, manufacturers have the majority.⁴² The implications of this “modernization” are unclear at the moment of writing.

In 2020, for the first time, the board elected its own chairs. They chose two co-chairs and two vice-chairs representing retail and manufacture in a parity-based system of governance. Previously, the chair was appointed by CGF. This seems to indicate that GFSI has gained some independence from the CGF. However, in the 2019 version of the governance model, the GFSI office managed by CGF was given more duties and powers.⁴³ And this is even more so in the 2021 version.⁴⁴

GFSI thus remains an industry-driven organization run by large international food corporations. According to its governance rules: “Steering Committee membership is not assigned to service providers … including associations, certification programme owners, certification or accreditation bodies, food safety related service providers.”⁴⁵ A balanced representation is one of the criteria for appointment in the steering committee: “Balanced geographical and industry sector representation i.e. manufacturing and retail, including the need to have significant representation from both large and small industry companies to ensure that decisions take into account the divers perspectives within the industry.”⁴⁶

Although GFSI was an organization exclusively for retailers, from the start other stakeholders in the global food supply chain were involved in its activities and the pursuance of its regulatory goals. A draft version of the first edition of the guidance document (2001) was circulated for external consultation and “external comments have been incorporated” in the second edition (2002). In 2002, GFSI announced the formation of a “Stakeholder group, open to representatives of all links in the food chain in order to participate in the endorsement [benchmarking] process.”⁴⁷ The Stakeholder Advisory Forum currently consists of retail, manufacturer, certification program owner members, audit and certification organizations, and an accreditation organization.⁴⁸ In September 2006, the “retailer only” taskforce was replaced by a technical committee including other stakeholders in the food chain as well as the certification industry.⁴⁹ This committee provides technical expertise and advice for the GFSI Board. For example, the committee formulates the benchmarking requirements, which are then determined by the Board. Technical working groups are composed of four manufacturers, four retailers, four certification/accreditation bodies, four service providers, and all recognized certification program owners.⁵⁰ The benchmarking committees were composed of experts from retailers, suppliers, or food service companies (balanced) and a representative from a national accreditation body.⁵¹ The 2020 benchmarking document states that the assessment of a certification program is performed by the GFSI technical manager and the benchmark leader.⁵² So again, more tasks for the GFSI management.

In addition, GFSI has strategically coordinated its regulatory activities with government representatives. One of the four top priorities of the organization in 2002 was “to encourage co-operation between the world-wide food sector and national and pan-national governments and authorities.” GFSI staff and participating firm representatives engaged in talks with national and international governments to explain GFSI’s mission and the value of third-party certification.⁵³ GFSI’s objective is to establish partnerships with governments. Over time, GFSI has entered into agreements, partnerships, and Memoranda of Understanding (MoUs) with various international and national government organizations, such as the International Accreditation Forum (2003), several Chinese agencies (2015), the United Nations Industrial Development Organization (UNIDO) (2016), and the Mexican National Service for Agroalimentary Public Health, Safety and Quality (2017).⁵⁴

From 2016 onwards, GFSI has hosted at its annual Global Food Safety Conference (GFSC) a meeting of governmental food officials and international governmental organizations, followed by a meeting of governments with the GFSI Board.⁵⁵ These meetings provide a global platform to discuss ongoing national food safety reforms and the role that third-party audits and certification can play. In these meetings, GFSI representatives explained the GFSI system and governmental representatives pointed out where areas need to be strengthened. National governments essentially voiced two concerns: (i) the quality and reliability of audits and auditors and (ii) the equivalence of the GFSI requirements with national food safety regulations. International governmental organizations such as the Food and Agricultural Organization (FAO) of the United Nations and the World Trade Organization (WTO) were very skeptical about private standards and perceived private food safety standards mainly as a trade barrier in emerging markets.⁵⁶ The organization of the government-to-business (G2B) meeting has been formalized with the creation of an organizing committee in September 2020. The committee has eighteen members (eleven representatives of national governmental organizations, six representatives of international public organizations, and one representative of the GFSI Board).⁵⁷ Members are expected to “commit to the GFSI outcome of building trust in GFSI” and to “engage their organisation.”⁵⁸

In 2014, GFSI introduced the option for government-owned voluntary certification programs to be benchmarked as a technical equivalent. Unlike GFSI recognition, technical equivalence does not include the assessment of the program’s governance and operational management. The reason for introducing this possibility seems to be the desire to accept Chinese certification programs that are managed by governmental agencies. Public voluntary certification programs that applied and achieved technical equivalence are China HACCP, Canadian Grain Commission HACCP, and USDA Agricultural Marketing Service. The fact that not only China but also Canada and the United States have had public standards benchmarked by GFSI is an indication of the dominance and the credibility and authority that GFSI has acquired globally.

GFSI Board members are CEOs from major food corporations – retailers, manufacturers, and food services who purchase many food products and raw materials. SMEs and farmers who produce and supply these products and other stakeholders are not represented in the board and in working groups. On its website, GFSI emphasizes the existence of a GFSI community: “GFSI exists thanks to a global community of passionate people who volunteer their time and expertise because they believe that everyone has a right to safe food. They share a common understanding that collaboration is the key to achieving what no one company or country can achieve alone …. Countless individuals from over 150 companies have contributed to over 25 Technical Working Groups, Local Groups, Task Forces, Committees and the Board of Directors.”⁵⁹

Finally, it should be noted that none of the recognized certification programs covers the retail of food, even though major corporate retailers initiated GFSI and still are a powerful constituent of GFSI. This illustrates that GFSI is meant to develop the private regulation of supply chains belonging to major food retail and food manufacturing corporations. Apparently, there was no need to regulate the retail sector itself.

9.4.2 Advances in Openness and Transparency

In the early years, the governance of GFSI was only loosely organized as a taskforce of retailers within CIES, the food business forum. Over the years, the openness and transparency of the organization has gradually increased. This applies to the governance of GFSI, its benchmarking procedure, and the transparency of the certification programs submitted to the benchmarking procedure.

The benchmarking requirements are publicly available on the GFSI website and outline the benchmarking procedure and requirements. The current GFSI governance model and rules of procedure specifies its objectives and the internal governance of GFSI: mandates, decision-making procedures, the frequency of meetings, eligibility criteria, election and responsibilities of the steering committee and sub-committees, the working groups, and local groups. A code of ethical conduct, obligatory statements of commitment, and a complaints procedure are also included.⁶⁰ The internal organization has now been highly formalized through a rulebook of eighty-seven pages.

Appointment of members of the GFSI board and committees was initially “by invitation only.” Currently, a call for candidates is distributed and interested parties can apply to be appointed. The eligibility criteria, the appointment process, and – in some cases – the distribution of members over categories are laid down in the GFSI governance model.⁶¹ The 2015 and 2019 version contained the provision that GFSI holds a register of Benchmarking Committee members on its website.⁶² However, this register could not be found on the website. According to the benchmarking document from 2020 a “list of all GFSI-approved Benchmark Leaders is available from GFSI upon request.”⁶³ So it seems that this information is not easily available (anymore). The benchmarking process is regulated in the Benchmarking Requirements document.⁶⁴ The benchmark leader and the GFSI technical manager perform the assessment of a certification program and give their recommendation whether to recognize the program. It is the board who approves the benchmark leader and takes the final decision on approval.

GFSI publishes on its website which standards are recognized and which are under review. The latter was not so obvious in 2005. Hugo Byrnes, director of food safety at CIES noted: “There are three other standards under review, which cannot be made public yet. If these are rejected by GFSI, it could cause embarrassment under their respective owners.”⁶⁵ Elsewhere, we can read which three standards were “under consideration.”⁶⁶ The current website provides information for each certification program in which of the seven steps of the benchmarking process it is.⁶⁷

On its website, GFSI has a library of free downloadable documents including training material of the Global Market Program, consultation documents and reports and case studies. GFSI is less transparent about financial issues. It is clear that the participation of food safety experts on the board, working groups, and local groups is paid for by the food companies where these experts are employed. There are complaints about the high fees for conferences and the reluctance of the management of GFSI to account for this.

The benchmarking requirements include ever more requirements related to the governance structure of standard owners. Certification programs need to be available for all food firms and all certification bodies should be able to certify against the program. Information that a standard owner should make publicly available include the normative document, the list of certification bodies that are accredited to certify for the standard, rules to prevent conflicts of interest, and a register of valid certificates.

9.4.3 Expansion and Ratchetting up of Food Standards’ Quality

From its inception, GFSI aimed at improving the quality and credibility of food safety standards. This includes the substantive norms in the benchmark requirements, transparency and integrity of food scheme owners and certification bodies, and the quality and reliability of audits.

Already, the first GFSI benchmarking requirements were based on international authoritative norms from the Codex Alimentarius Commission and the ISO/IEC Guide 65. Initially, the benchmark requirements only consisted of a list of items that an application should contain, the consequent steps in the procedure, and some time limits.⁶⁸ With every new edition, the benchmarking requirements have been tightened and new themes have been added. Usually, when a weakness or issue that needs attention is in the spotlight, GFSI decides to introduce a new technical working group to investigate the matter and to make recommendations as to how to tackle the issue. Such groups were introduced for third-party certification, auditor competences and auditor training, alignment with national authorities, capacity building in global markets, food defense, food safety culture, and managing risks in produce and leafy greens. The activity of these groups often results in new or adapted requirements in the GFSI guidance document. New topics that were introduced in the benchmarking requirements include, for example:

• – reducing the risks of food fraud (in response to the horsemeat scandal);

• – auditor competences and auditor training;

• – stricter requirements for the frequency and scope of audits;

• – unannounced audits were added, first optional, later mandatory;

• – extended requirements related to scheme owner management and governance; and

• – requirements related to food safety culture.

Subsequently, new scopes and sector-specific requirements followed. The first editions only had general requirements for all schemes. The 2020 edition consists of general requirements for all certification programs and specific requirements for twenty scopes covering the whole food supply chain from farming, processing, distribution, and the retail and supply industry, such as feed production, food safety services, and food packaging.

What remains constant, despite all the new themes, is the exclusive focus on food safety. Other themes such as animal welfare and environment or ethical sourcing are explicitly kept out.⁶⁹ Moreover, just like the very first edition, the 2020 edition contains three key elements: HACCP; a food safety management system; and Good Practices for Agriculture, manufacturing, and distribution (GAP).

One of the main concerns over the past twenty years has been the reliability of food safety audits and certificates. This is a crucial issue as the value of the recognized certification programs depends on thorough and reliable food safety audits. Doubts about food safety certificates are voiced over and over again by governments, media, and parts of the food industry. GFSI recognizes these weaknesses and tries to explain the system to everyone and to ratchet up the quality of food safety audits. The first step was the choice for third-party certification, adding an independent observer to the scene. Despite checks and balances in the system of third-party certification⁷⁰ concerns remain. In each updated edition, requirements for audits and auditors were tightened. There have been years of discussing and negotiating the introduction of unannounced audits and requirements related to auditor competence and training. Due to conflicting interests and resistance, several of these changes took a long time before a final decision was made and the adaptation could be implemented. Recently, GFSI concluded that benchmarking of food safety certification programs was not sufficient. “There has been an understandable concern about the efficacy of audits and more specifically the competence of some food safety auditors themselves.” GFSI introduced its Race to the Top (RTTT) plan. This plan “is intended to address the specific challenges GFSI has been facing in relation to trust and confidence in GFSI certification outcomes …. driving improvements in the food certification system is vital to achieve our mission.”⁷¹ In 2021, GFSI introduced its benchmarking and harmonization of professional recognition programs for food safety auditors.⁷²

Over time, more detailed requirements were imposed related to the scheme governance and management. For example, the 2020 edition introduces the requirement for CPOs to carry out annual performance reviews of certification bodies. The relation between scheme owner and GFSI is increasingly formalized. Cooperation with the IAF resulted in adaptation of the guidance document. Efsis, a standard that belonged to the first four accepted standards, subsequently declined a new application because the requirements that the program should not be limited to members or own customers and that a scheme owner could not be the organization performing the audits, could not be met.

In addition to benchmarking, the dissemination of knowledge and the exchange of experiences is an important characteristic of GFSI’s activities. GFSI provides a global platform for professionalization of food safety experts working at certified food companies, at certification bodies, at certification program owners, at consultancy firms, and at regulatory agencies and departments. To pursue this objective, GFSI organizes annual global food safety conferences, focus days in different continents and countries to promote GFSI, and the system of third-party certification.

The modifications and increase in requirements make it ever more difficult and more expensive for certification programs to meet all requirements. For this reason, one of the first recognized schemes has decided not to register anymore. The organization behind this scheme has split into two systems, FSSC 2000 as a GFSI-recognized system and Dutch HACCP as a local system for traditional production.

9.4.4 Globalization of Benchmarking Activities

GFSI started in Europe and it was dominated by Europe-based retailers for a long time, although its ambition always was to be a global initiative. Soon after its launch, some US and Canadian supermarkets joined the taskforce. The first four standards that were recognized by GFSI were European.⁷³ In January 2004, a US retail standard was approved, SQF 2000. The breakthrough to the United States food industry was in 2007, after Walmart made GFSI certification mandatory for its suppliers.

Not only the member companies and the recognized food safety certification programs but also the other activities employed by GFSI were during the first phase located in Europe only. One of five strategic priorities 2011–2015 was to “continue presence in Europe, build momentum in North America and develop a strategy for APAC.”⁷⁴ Of the twenty annual global food safety conferences (GFSC), fifteen took place in Europe. It was not until 2011 that the first GFSC outside Europe was held, in Orlando, Florida, in the United States. Four years later, in 2015, the first GFSC was held in Asia.

From 2011, GFSI organized focus days in sourcing countries in Latin America, Asia, and Africa. “GFSI Focus Days and Regional Events aim at raising awareness around the benefits of GFSI in regional markets. They were initiated to provide local stakeholders in the food industry with an opportunity to find out more about the Global Food Safety Initiative. They also serve as a unique opportunity for networking and knowledge exchange.”⁷⁵ In China and Japan, GFSI organized such a meeting every year. Local groups are regional networks of companies that want to promote GFSI and share knowledge to improve food safety. The first local group was located in Japan (2012). Now there are local groups in Mexico (2013), China (2013), US-Canada (2013), South Latin America (2015), Europe (2016), and Australia-New Zealand (2019).⁷⁶

GFSI also aimed at cooperation with governments and international organizations all over the world. Responding to critics that private food safety standards are a barrier for SMEs and food businesses on non-Western markets, GFSI launched its global markets program in 2011 (manufacturing) and 2012 (primary production). The GFSI Global Markets Programme is designed to assist companies who have underdeveloped food safety systems to learn and adhere to best food safety practices. It is a systematic continuous improvement process that companies can follow to establish, and achieve recognition for, effective food safety systems. The Programme is “a stepwise pathway towards GFSI-recognised certification for companies that lack or wish to improve their food safety systems.”⁷⁷

In 2012, the majority of certifications for most recognized schemes were in Europe, exceptions are SQF (N-America) and FSSC 22000 (more equal distribution).⁷⁸ GFSI-recognized certificates have been issued in 162 countries.⁷⁹

9.5 Constructing and Managing GFSI’s Legitimacy

9.5.1 Legitimacy Dynamics

To what extent can the processes that we have highlighted be understood from the perspective of legitimacy? Some of the processes are clearly a response to the criticisms and demands expressed by external or even internal actors. Prominent examples of this are the reliability of food safety certificates, the vulnerability for food fraud, and capacity building in emerging markets. From the very start, the reliability of certification schemes has been the subject of persistent criticism by governments, NGOs, and businesses in the supply chain. GFSI could not simply ignore these critiques, as they go to the core of the organization’s regulatory goal, that is, to promote improvement in food safety management systems around the world. To provide these critiques with an answer, bolster its legitimacy, and be resilient, GFSI has sought to make these systems more reliable by promoting third-party certification and by imposing increasingly stringent benchmarking requirements for audits, auditors, and the certification bodies that employ them. The collaboration between GFSI and IAF, and the requirement that certification bodies must be accredited by an accreditation body that is a member of the IAF, are expressly meant to improve credibility. The modernization program, Race to the Top (RTTT), presented in 2020, is an explicit response to systematic doubts about the functioning of the certification system. The program aims at a fundamental shift from compliance being enough to continuous improvement at all touchpoints.⁸⁰ Accordingly, it seeks to address the concern among governments, NGOs, and part of the food industry itself regarding poor auditor performance and a lack of a sufficiently strong food safety culture in the industry.

GFSI stresses in its communication the existence and the importance of a GFSI community: “GFSI has grown into a vast, global multi-stakeholder movement.”⁸¹ At GFSI conferences and meetings – as one of the authors has observed as an invited speaker and participant – there certainly is a kind of community of global food safety experts.⁸² However, within this community, GFSI has to deal with, at least partly, conflicting interests of stakeholder categories. Governments, transnational corporate retailers, transnational top food manufacturers, farmers, SMEs, certification bodies, certification program owners, and economic development organizations all have a certain interest in securing safe food, but they also have different interests. GFSI is finding the middle ground in all of this. The politics involved in global food safety standards can, for example, explain why it took years before some form of mandatory unannounced audits were included in the GFSI benchmark requirements. Also, how to deal with the issue of auditor training and competences is contentious. Some recognized schemes preferred their own existing training program and examinations. Certification bodies feared that auditors would need to pass various exams, resulting in higher costs while no added value is expected. Since 2019, auditors are required to take and pass the GFSI Knowledge Exam. The examinations are provided by the certification programs and the requirements include mutual recognition of exam results between CPOs to assure that an auditor only needs to pass the exam once.⁸³

To sum up, the continuous improvement of the quality of the food safety certification programs is associated with GFSI’s desire to build and maintain its legitimacy as a transnational private rule-maker. This organizational concern closely aligns with the original regulatory goal of GFSI, namely to enhance global food safety.

Advances in openness and transparency can also be interpreted as reactions to criticism and changing international norms. Transparent procedures and requirements to ensure integrity, independence, as well as checks and balances in the certification system are in line with international authoritative norms and concerns voiced by governments and IGOs.

However, the first and foremost driver for the globalization of certification and GFSI’s benchmarking activities are the commercial interests of the retailers and manufacturers represented in GFSI. Transnational retailers and manufacturers want to source products and ingredients from all over the world. For that reason, it is important that their suppliers, not only in Europe but also in Asia, Latin America, and Africa, can deliver products from certified firms complying with the GFSI requirements. The Global Markets Programme helps in building and extending this supplier network. At the same time, however, this program also is a response to the concerns that international governmental organizations such as the Codex Alimentarius Commission and the STDF have voiced around the barriers to trade that private food safety standards create for developing countries.⁸⁴

The pluralization of its constituents is more difficult to directly link to GFSI’s desire for legitimacy. From the perspective of legitimacy, it would seem wise to include all stakeholders in the decision-making process. However, in GFSI’s organizational structure, this is only partly what happened. In the decision-making bodies, the GFSI steering committee/board and the CGF, many categories of stakeholders are not represented. This includes certification program owners, certification bodies, auditors, farmers, SMEs, governments, IGOs, consultants, accreditation bodies, and consumers. Although some of these categories are represented in working groups providing advice to the board/steering committee, they lack formal voting rights. Nevertheless, this degree of pluralization did bolster GFSI’s credibility and regulatory clout. GFSI has created a global platform around food safety and succeeded in forming a kind of GFSI community working together to improve food safety and exchanging knowledge and experiences. GFSI also succeeded in establishing partnerships with governmental organizations. Collaboration and partnerships with (international) governmental organizations further add to the legitimacy of GFSI as governments are widely believed to serve the public interest, while industry is believed to put self-interest first.⁸⁵

9.5.2 Institutionalizing Legitimacy

To what extent do we observe the three phases in the process of institutionalizing legitimacy as discussed in Section 9.2 in the evolution of GFSI as a transnational private rule-maker? During the first years, it indeed seems that economic reasoning is the most important driver for GFSI’s regulatory activities. Other arguments have gradually gained weight. Nevertheless, economic interests remain the most fundamental considerations. The ratchetting up of the benchmarking requirements leads to higher compliance costs, usually for parties up of the supply chain. These involve the companies that want to obtain a certificate and the firms that conduct audits and carry out certification. Too high costs could drive these actors out of the GFSI system. GFSI therefore treads carefully and proceeds step-by-step in laying down more stringent benchmarking requirements for the certification programs. Also, other actors, including the transnational supermarket chains and food manufacturers that are the driving forces behind the private standards, are sometimes tempted to opt for low costs. Nevertheless, so far GFSI has managed to build a fair degree of legitimacy both in the world of the food industry and food safety experts.

The discussed changes are predominantly gradual. Clearly distinguishable phases as suggested by Bernstein and Cashore are hard to uncover. The same is true for a (causal) link between changes in GFSI’s regulatory standards and food safety crises. What is clear in the case of GFSI is that, as Casey has argued for GlobalGAP, the first phase of institutionalization lays down the aims, structure, and power distribution within the transnational private rule-making body. Despite its growth and the inclusion of other participants next to retailers, GFSI stays true to its initial mission: “to enhance food safety, ensure consumer protection, strengthen consumer confidence, benchmark requirements of food safety systems and improve cost efficiency throughout the supply chain.”⁸⁶ Similarly, GFSI remains an industry-led organization relying on the participation of food safety experts of large food corporations, some of which have been represented from the very start. Also, GFSI’s main regulatory activity still is the benchmarking of food safety certification schemes. The ambition “once certified, recognised everywhere” still stands.⁸⁷

At least three changes have brought about change in GFSI’s initial organizational structure and regulatory activities: the joining of food manufacturers in the board and retailers losing the majority in board and committees, the introduction of the Global Market Programme, and the introduction of the Technical Equivalence Procedure. The latter implies that government-owned food standards are subjected to assessment by an industry organization such as GFSI. For example, the USDA-harmonized GAP audit was augmented to meet GFSI equivalence standards.⁸⁸ This collaboration inevitably lends legitimacy to GFSI as a global private rule-maker.

To conclude, GFSI has been deeply concerned in constructing and maintaining its legitimacy as a transnational private rule-maker. It has succeeded in gaining acceptance and credibility among major parts of the global food industry and even in governmental circles by being responsive to needs and criticism of various stakeholders. As such, GFSI has shown strong potential for adaptation, nourishing its resilience and dominance. However, responsiveness cannot explain all organizational changes we have discussed. Economic self-interest of the leading constituents of GFSI has initiated some of these changes yet delayed or prevented others.