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 conductFootnote ¹ or agreements between individual multinationals and international trade unions in the form of global framework agreements.Footnote 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.”Footnote 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,Footnote 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.Footnote 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.Footnote ⁶ This private governance often emerges in response to pressures exerted by civil society actors, such as unions and NGOs, particularly when crises emerge.Footnote 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.Footnote 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.Footnote 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),Footnote 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.Footnote 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.Footnote 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.Footnote 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 conditionsFootnote ¹⁴. 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.Footnote 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,Footnote 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 2018Footnote ¹⁷ 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.Footnote 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,”Footnote 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 AnnerFootnote ²⁰ 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 OrganizationFootnote ²¹ 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.Footnote 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, 2020Footnote ²⁴

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,Footnote 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:

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,Footnote 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,Footnote 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.Footnote 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,Footnote 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 2016Footnote ³⁶ 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.Footnote 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 yearsFootnote ³⁸ 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.