Beyond audit: the regulatory legitimacy of multi-stakeholder initiatives in the apparel sector

ABSTRACT This article is about multi-stakeholder initiatives that seek to regulate the human rights impacts of global apparel supply chains (Apparel MSIs). MSIs aim to improve human rights for millions of apparel workers worldwide, but after two decades they show little evidence of such improvement. Civil society critics argue that MSIs are an ineffective, unreformable model of private regulation. This article contends that a particularly pressing issue facing Apparel MSIs is the manner in which they seek to control their targets of regulation and ensure regulatory compliance. Most Apparel MSIs rely predominantly on social audit to monitor and ensure compliance with human rights standards. Yet there is now a strong body of evidence showing that social audit is an inherently flawed tool that is not sufficiently powerful to implement meaningful and consistent change. Accordingly, this article argues that Apparel MSIs must urgently move beyond social audit as a regulatory technique. They must instead pay attention to power relations and complex political contexts, amplify the voice of workers and harness complementarity between public and private regulation. Apparel MSIs must transform to improve effectiveness, preserve legitimacy and realise their ultimate aim of improving human rights outcomes for apparel sector workers.


Introduction
Multi-stakeholder initiatives (MSIs) first emerged in the 1990s as a means of addressing the social and environmental impacts of transnational business.This article analyses the specific case of MSIs that address the adverse human rights impacts of global supply chains in the apparel sector (Apparel MSIs).Despite being in operation for more than two decades, Apparel MSIs have not demonstrated substantial or consistent human rights advances for global apparel supply chain workers.
This article argues that a particularly pressing issue facing Apparel MSIs is how they seek to control their regulation targets to ensure compliance.To date, most Apparel MSIs have relied predominantly on social audit to monitor and ensure compliance with human rights standards.Although conflicting views around the value of social audit remain, there is now a strong body of evidence showing that social audit is an inherently The emergence of MSIs in the apparel sector Apparel MSIs emerged in the mid-1990s in response to the human rights challenges posed by complex global supply chains.The global apparel sector, valued at an estimated US$1.5 trillion2 and employing more than 60 million workers worldwide3 relies heavily on a globalised system of production.Today, the typical brand or retailer (buyer) in the apparel sector no longer manufactures the goods that it sells.Instead, it outsources production to multitiered networks of contractors (suppliers), typically located in developing economies where production costs (in particular labour costs) are low and regulations weak. 4his outsourced business model has created significant opportunities for employment and aided the development of emerging economies.It has also given rise to profound human rights challenges, with systemic human rights breaches a daily reality for tens of millions of apparel supply chain workers across the globe. 5n response, major apparel buyers initially turned to self-regulation in the form of supplier codes of conduct. 6Growing dissatisfaction with these unilateral and non-binding codes in turn gave rise to a new experiment in private transnational regulation: the MSI. 7Defined at a high level, an MSI is a voluntary initiative involving a group of networked stakeholders working together to address the social and environmental impacts of transnational business. 8Participating stakeholders are typically business, trade unions and non-governmental organisations (NGOs).MSIs seek to respond to global governance gaps and the limits of self-regulation through a combination of tools including more standardised codes, independent monitoring, greater transparency and broader civil society participation. 9espite much initial promise, evidence suggests that Apparel MSIs have had only a limited impact on human rights protection. 10The seeming lack of progress has caused disillusionment among a growing cohort of civil society advocates who argue that all MSIs (including Apparel MSIs) are an entirely ineffective model of private regulation, that they are not fit-for-purpose, and are entirely unreformable.Instead, they point to a recent innovation, known as the worker-driven social responsibility model (worker-driven model) as a more promising and effective private regulatory alternative. 11

Apparel MSIs and regulatory legitimacy
This article analyses MSIs at the intersection of two disciplines: political science, and law and regulation. 12Both place primacy on the role of MSIs as regulators, each bringing a different focus.
Although MSIs are often described as a form of private transnational regulation, not all MSIs perform a regulatory function.This analysis focuses on those Apparel MSIs that regulate or perform a combination of learning and regulatory functions.These are: the Fair Labor Association (FLA), Social Accountability International (SAI), Worldwide Responsible Accredited Production (WRAP), the Ethical Trading Initiative (ETI) 13 and Fair Wear Foundation (Fair Wear).According to political science scholars of private transnational governance, for regulation beyond the state to be justified, the assumption of regulatory functions by private actors must be democratically legitimate.That is, there must be a 'socially shared belief' that the regulator has the capacity and the authority to impose rules on a community of citizens'. 14Unlike states, MSIs are not embedded within established democratic mechanisms, nor do they have access to traditional mechanisms of state power to ensure compliance. 15Thus-to the extent that they regulate-MSIs must establish legitimacy in order to justify, and recruit stakeholders to, their regulatory scheme. 16emocratic legitimacy is constituted by both input and output legitimacy.17Input legitimacy considers to what extent regulations are perceived as justified or credible, in light of their procedures and governance structures. 18Output legitimacy is outcome focused and is concerned with the extent to which regulation effectively regulates the issue being targeted. 19The present analysis focuses on a key component of output legitimacy, and arguably the most challenging component of legitimacy overall: the ability to ensure that rules are followed and applied in practice. 20presentative of the Worker-Driven Social Responsibility Network (26 September 2018); Interview with a representative of IndustriALL (10 October 2018).Examples of the worker-driven model include the Bangladesh Accord on Fire and Building Safety; Action Collaboration Transformation; and the Lesotho agreements to prevent genderbased violence and harassment.
This requires a consideration of the efficacy of an MSI's functional control mechanisms. 21

An MSI typology
An MSI typology proposed by a group of business and human rights experts provides a useful framework to analyse the functions of an MSI. 22The typology distinguishes between three distinct operational models: learning (Learning MSIs); certification (Certification MSIs); and governance (Governance MSIs).
Learning MSIs typically undertake or promote activities such as awareness raising, dialogue, best practice sharing, data sharing, training, networking, collaboration and discrete projects.They may also engage in some form of standard setting.They do not, however, perform any oversight function or impose consequences for non-performance.That is, they do not have control mechanisms and therefore do not perform a regulatory function. 23ertification MSIs impose concrete obligations, usually set out in a code of conduct.In contrast to Learning MSIs, Certification MSIs verify compliance with these obligations, usually through a combination of reviews, social audit and other forms of monitoring.Enforcement occurs through the certification process itself, as those deemed non-compliant are not granted certification.Certification MSIs certify sites of production and (sometimes) associated chains of custody.Thus, suppliers are the principal rule targets.
Governance MSIs are based on a membership structure; typically, buyers are at the apex of supply chains.Members are required to comply with concrete obligations, usually set out in a code of conduct, and face the prospect of suspension or termination in cases of non-compliance.Each Governance MSI verifies member compliance through its own unique combination of reviews, social audit, reporting and other monitoring requirements.
Table 1 summarises the key characteristics of each model.Of the Apparel MSIs considered in this analysis, the SAI and WRAP are both Certification MSIs: they set out specific codes of conduct with concrete obligations and oversee a system of certification in respect of individual supplier facilities.Therefore the primary control mechanism utilised by these Certification MSIs is social audit.
The FLA and Fair Wear are Governance MSIs.They adopt substantially similar obligations and oversight techniques.Member companies commit to adopting the respective codes of each MSI and monitoring code compliance among their first tier suppliers.Members are also required to implement comprehensive management systems at the headquarter level.The FLA and Fair Wear verify member compliance through four main control mechanisms: (1) social audits of a proportion of its members' suppliers; 21 Noting that control is a component of input legitimacy, but is inextricably linked with effectiveness: Nicolas Hachez, and Jan Wouters 'A glimpse at the democratic legitimacy of private standards' (2011) 14(3) Journal of International Economic Law 677-710. 22Dorothée Baumann-Pauly and others, 'Setting and enforcing industry-specific standards for human rights: the role of multi-stakeholder initiatives in regulating corporate conduct' in Dorothée Baumann-Pauly and Justine Nolan (eds), Business and human rights: From principles to practice (Routledge 2016). 23Bernstein and Cashore (n 15); ibid 108-116.
(2) annual reviews of members' management systems and controls; (3) third-party complaints mechanism for workers and their representatives; and (4) formal member review procedures, which can lead to termination of membership.The third and fourth forms of control, while important, act as safeguards, and apply only to a handful of cases.Thus the primary forms of control for these Governance MSIs are social audits and reviews of companies at the headquarter level.The FLA also provides an additional layer of accountability through its accreditation process.Both the FLA and Fair Wear place the primary onus on buyers, rather than suppliers, to implement their respective codes of conduct.
Like FLA and Fair Wear, the ETI requires its members to adopt and implement its ETI Base Code throughout their supply chains.In contrast, however, the ETI does not monitor its members' supplier base through social audit-instead focusing on learning and collaboration through projects, training and best practice sharing.The ETI also imposes some level of oversight: member companies are required to report annually on the steps they are taking to improve working conditions in their supply chains, with consequences for members who do not comply.A senior ETI representative stated that ETI both tries 'to hold members to account for change' and is also a 'learning and doing and sharing organisation.'24On this basis, ETI is classified as a combined Learning and Governance MSI.Despite reflecting different models, these MSIs share a critical underlying feature: each MSI (save the ETI 25 ) relies on social audit as a significant form of control.

The failure of social audit
Social audit has been subject to extensive scrutiny and criticism by NGOs, trade unions and scholars.Although there is some research supporting the effectiveness of social audit, 26 there is a growing consensus that social audit is a limited regulatory tool, capable of delivering only superficial and short-term outcomes at best, and wholly inadequate for detecting and addressing more serious and systemic human rights violations. 27The failure of social audit is most tragically exemplified by major disasters in Bangladesh such as the Rana Plaza collapse and the Tazreen and Ali Enterprises factory fires.Each of these facilities had passed multiple prior social audits. 28 significant body of academic research provides the empirical basis for the assertion that social audit is routinely failing workers.Studies over time have found a lack of sustained improvements across most categories of rights, with suppliers moving in and out of compliance.Combined, existing research suggests that at best, social audit has led to inconsistent improvements in basic rights such as health and safety, wages, working hours and leave entitlements, and had no impact on broader systemic rights such as freedom of association and non-discrimination. 29hese limited outcomes can be attributed to two separate but interconnected issues: inherent limitations in the manner that social audit is conducted; and broader systemic and structural impediments to changes in conditions for workers.

Inherent limitations
The inherent limitations of social audit have been extensively canvassed by academics and civil society.First, audits are usually undertaken as a short checklist exercise 25 Noting that while the ETI does not itself conduct audits, its members have typically relied on auditing to implement the ETI code throughout their supply chains: Ethical Trading Initiative (ETI), 'ETI perspective 2020 (generally over the course of a few days, at one to two-year intervals), which at best, provide only a snapshot in time. 30Second, workers are not meaningfully involved in the audit process.Management often pre-select coached workers to be interviewed, while on-site interviews inhibit frank and open responses. 31This limited inclusion affects the accuracy of information collected by auditors, and ultimately serves to disempower workers, the intended beneficiaries of the process. 32In addition, local NGOs and trade unions are often not consulted; yet these are the groups most likely to shed light on violations. 33Third, instances of supplier fraud and evasion-for example, falsifying workers' records and 'cleaning up' facilities prior to inspections -are commonplace. 34Fourth, there are concerns about independence and appropriate expertise.The fact that auditors are paid (either by buyers, supplier factories or MSIs), combined with their often limited experience in human rights and labour issues, raises questions about the quality and independence of social audits. 35Fifth, audits rarely look beyond tier one of the supply chain, while unauthorised subcontractors and temporary workers are usually not picked up by social audit at all. 36et these are some of the most vulnerable workers within the supply chain. 37inally, the audit process is not sufficiently transparent or accountable.At present, of the Apparel MSIs considered in this article, only the FLA publishes its audit reports.A lack of transparency renders stakeholders unable to assess the basis on which certification was granted. 38oader, systemic limitations Even if the inherent limitations described above were overcome, and social audit was undertaken in a perfect matter, existing evidence indicates that social audit is simply not a powerful enough tool to protect the human rights of workers. 39In the words of a senior representative of the UK Trades Union Congress: … it's an external fix that you're trying to lay on a fundamentally broken system and generally speaking it does not work because it doesn't lead to sustainable long-term changes in behaviour or activities or values … 40 Broader structural conditions must also be addressed. 41This article identifies two particularly significant structural issues that social audit completely fails to address: poor purchasing practices and local, social and political conditions.A majority of interviewees in this research identified either one, or both, as critical issues that buyers and Apparel MSIs must address.

Purchasing practices
Major brands and retailers operate in a highly competitive environment, where short-term profit expectations of shareholders-combined with consumer pressure in relation to price, variety and speed-propel the ongoing search for cheaper and faster production. 42ombined, these commercial and competitive pressures result in the imposition of unfair purchasing practices-including fluctuating order demands, unrealistic lead times, smaller and more frequent shipments and greater product variety-which place unrealistic production pressures on suppliers. 43These practices in turn have a significant flow on effect on the human rights of workers who are subjected to variable working hours, excessive and forced (and at times unpaid) overtime, limitations on leave and breaks and unsafe working conditions, in an attempt to meet the commercial expectations of buyers. 44urther, the relative power of buyers allows them to conduct aggressive price negotiations, leading to outcomes that do not reflect the true cost of production, labour and social compliance efforts. 45Again, it is workers who suffer, as aggressive pricing practices are passed on in the form of below living wages, unpaid overtime and benefits and unsafe working conditions. 46The global pandemic exacerbated these dynamics.A 2020 survey of 75 apparel suppliers from across the globe found that, on average, 65% of suppliers had received demands for price cuts from buyers, resulting in 56% of suppliers forced to accept orders below cost. 47A representative of the Institute of Workers and Trade Unions (Vietnam) highlighted these cost pressures: … according to our [local] survey we see that the purchasing price is lower for 5 years or 10 years recently.That means they pushed down the price and here the minimum wage increase and the union tried to push it up.So in the middle, the factory here, how can they survive? 48 independent expert similarly argued: '[i]n some ways it's become a joke because you've got a company pushing on one side with purchasing practices and then they're auditing on the other and those are in conflict'. 49Consequently, buyers must be compelled to look inwards and address the human rights impacts of their own conduct and business models.Eighteen out of 33 interviewees in this research highlighted purchasing practices as a critical issue that buyers must seriously address.For example, a senior representative of the Sommilito Garments Sramik Federation (a union based in Bangladesh, that is an FLA member) stated: So I don't know, I don't much believe in [social audit].Because if the company pays good money, good salary or the brand pays a fair price why you need the audit, you don't need this, because you are not paying, you are not serious about the workers' issues, that is why you are saying we are doing audit, we are doing that.You don't need this, just pay the workers and build their capacity for their rights and responsibilities. 50representative of the Institute for Workers and Trade Unions (Vietnam) argued: We need the pressure from the buyer, we need the buyer to be more responsible to pay more for the labour cost to calculate the price so that it includes the labour cost.You need that, and you need commitment from the buyer … 51 (emphasis added) This is a significant point.In addition to reflecting the commercial terms of engagement between buyers and suppliers, purchasing practices also encompass the broader nature of the relationship between a buyer and supplier.This includes the extent to which buyers are willing to establish more long-term, committed relationships with their suppliers. 52evels of buyer commitment are directly connected to their ability to influence supplier conduct, or in United Nations Guiding Principles on Business and Human Rights (UNGPs) 53 terms, to exercise leverage.A company is most likely to be able to exercise leverage when it constitutes a large proportion of a factory's output, places regular orders, and has a long-term relationship with the supplier. 54In the apparel sector, suppliers typically manufacture for a multitude of buyers, with even large brands sometimes only constituting a small proportion of a supplier's total production. 55Further, research shows that buying decisions remain predominantly motivated by commercial terms rather than social compliance. 56For the majority of buyers, social compliance remains an afterthought, something that is considered subsequent to entering into a contract with a supplier, rather than a critical component at the beginning, or when considering whether to maintain a commercial relationship. 57In circumstances where suppliers are generally not rewarded for improved compliance, it is difficult to see how buyers can expect to exercise leverage in this regard.
Locke and his colleagues suggest that the power of buyers to compel better human rights outcomes can be addressed by what they term 'the commitment model'.This is in effect a form of capacity building, whereby buyers assist suppliers to improve social compliance through 'joint problem solving, information sharing, and the diffusion of best practices'. 58Evidence shows that this kind of commitment model is not widespread.According to the International Labour Organization, while more than 90% of suppliers (across sectors) were expected to follow a code of conduct, 49% of those did not receive any assistance with compliance. 59Social audit, with its supplier focus, is clearly not capable of addressing these issues, nor was it ever designed to do so. 60Ultimately, buyers must also come to the table, and be willing to factor in compliance into their own practices and business models.

Local, social and political conditions
There is widespread consensus that social audit has had no impact on broader systemic rights such as freedom of association and non-discrimination.Such rights enable the negotiation and achievement of other rights and are critical to improving working conditions and advancing worker empowerment and social justice. 61Yet these are complex, 'less visible [and] more deeply embedded' rights, which are not readily identified, let alone remediated by an overly technical checklist audit approach. 62As noted by an ETI company member: I believe in the countries that we source from … there are systemic issues there, like gender discrimination and like lack of voice which will never ever be picked up in a social audit and I almost think we could get rid of social auditing and it wouldn't really change anything, we wouldn't be worse off, in fact we would be better off.We would have more time and more cash. 63 these complex rights, the right to freedom of association is arguably the most critical, due to its direct role in enabling the negotiation and achievement of other rights.Six interviewees explicitly highlighted freedom of association as a critical issue that buyers needed to seriously address.Fair Wear and the ETI have also each highlighted freedom of association as a critical challenge facing the apparel sector. 64ocial auditors are not adequately trained to identify complex rights such as freedom of association. 65Even in cases where social audit does identify violations, remediation tends to be in the form of improved management policies and training programs.This is a wholly inadequate response to violations of rights that are inextricably intertwined with underlying social and political conditions. 66As a senior representative of CiviDep India (an FLA NGO member) argued: Social audits might be effective in dealing with non-compliance to physically verifiable aspects of production such as fire safety.However, when it comes to industrial relations and the various vulnerabilities of workers, audits have failed. 67milarly, social auditors rarely pick up discrimination, 68 despite gender discrimination and gender-based violence and harassment being widespread in the apparel supply chain. 69 majority of interviewees acknowledged that MSIs should not be expected to solve these complex, country specific, political problems on their own and that the primary responsibility lay with host states.Yet it is also somewhat disingenuous of Apparel MSIs to claim to promote freedom of association in their codes of conduct and continue to rely on such an obviously inadequate tool as social audit.Apparel MSIs must develop new strategies to address these complex rights.

Certification and governance MSIs
It is at this point that the distinction between Certification and Governance MSIs becomes significant.Although both MSI models rely on social audit, the centrality of social audit to Certification MSIs limits their ability to address the challenges outlined above.

Certification MSIs
The two Certification Apparel MSIs-SAI and WRAP-seek to regulate the conduct of suppliers rather than buyers.Their primary form of control is certification (or non-certification), which takes place by way of social audit.Certification MSIs are not able to employ the threat of termination as they are not membership-based organisations.Ultimately, the very functional model adopted by Certification Apparel MSIs poses an obstacle to a meaningful evolution beyond audit.Their focus on individual facilities means that they are inherently supplier focused and too narrow in scope to address broader systemic issues such as purchasing practices and social and political context.Ongoing reliance on this increasingly discredited and narrow tool as the only form of control leads to the conclusion that Certification Apparel MSIs do not utilise an effective control mechanism and are therefore not performing satisfactorily in relation to rule implementation.This severely limits their ability to establish regulatory legitimacy.

Governance MSIs
By contrast, the three Governance Apparel MSIs-FLA, Fair Wear and ETI-have recognised, and sought to act upon, both the inherent and broader, systemic limitations of the traditional social audit model.

Inherent limitations
The FLA and Fair Wear have attempted to ameliorate at least some of social audit's inherent limitations.In 2012, the FLA introduced its 'Sustainable Compliance Methodology' (SCI), precisely because it came to the conclusion that auditing was not preventing recurring violations or contributing to sustainable improvements.Rather than focusing solely on specific, visible violations, the SCI seeks to focus on root causes of non-compliance. 70The FLA has not, however, evaluated the effectiveness of its SCI approach.
As part of its audit process, Fair Wear considers not only working conditions but also broader causes such as members' purchasing practices, their monitoring efforts, the factory's management systems and worker-management communication. 71It conducts audits via an in-country team of independent experts and places a heavy emphasis on interviews with workers and local stakeholders (including trade unions, NGOs, business associations and local authorities). 72It also conducts audits over a relatively long period of time, ranging from 6 to 13 days. 73Nonetheless, empirical evidence suggests that Fair Wear's improved audits have still resulted in only minor, non-statistically significant improvements, which are not consistent or sustainable over time. 74n contrast to Fair Wear and the FLA, the ETI does not utilise social audit as a form of control at all, having taken the public position that social audit 'has not delivered the expected credible assurance of standards that [brands] and their customers seek.' 75 Instead, the ETI promotes a UNGPs human rights due diligence approach, which companies may then supplement with high-quality audits.However, as detailed below, the ETI does not, at this stage, have in place sufficiently strong alternate control mechanisms to ensure that its members are implementing effective human rights due diligence programs.

Purchasing practices
Perhaps the most significant, systematic, way in which Governance Apparel MSIs have moved beyond the audit is through the use of headquarter level assessments as a control mechanism.These assessments focus on buyer conduct, including purchasing practices and levels of commitment to suppliers, which is consistent with broader public acknowledgment by each Governance Apparel MSI of the need for buyers to address purchasing practices. 76LA member companies commit to, among other things, the ten FLA Principles of Fair Labor and Responsible Sourcing (FLA Principles).Principle 2-Responsible Purchasing Practices requires members to 'align planning and purchasing practices with commitment to workplace standards'.77 That is, members are required to implement planning and purchasing policies and procedures that address appropriate financial terms, adequate lead times and planning processes.Member companies are also expected to provide suppliers with incentives to comply with FLA standards.The Principle also sets out some high-level Key Performance Indicators, which members self-report against annually.The FLA also assesses compliance with the Principles for those members seeking accreditation.A representative of Adidas, an FLA member, highlighted the rigour of the accreditation process: … I mean they go on the ground, they meet with the teams, they evaluate, they shadow audit.They effectively are seeking to accredit and say yes this is a substantial and robust program that upholds the charter obligations and then report out on it.So that to us is a value because there are actually very few organisations out there that I can think of that actually can look at a compliance program that companies are running and evaluate it with the knowledge that they have.I don't think others can do that so easily. 78ile the overall accreditation process appears to be reasonably rigorous, the review of accreditation reports undertaken for the purpose of this study reveals a relatively highlevel assessment of member's purchasing practices-with a focus on policies in place, rather than actual changes in purchasing practices-arguably reflecting the high-level nature of the Principle itself.
Fair Wear assesses the policies and systems of member companies through its 'Brand Performance Check', which incorporates purchasing practices.Fair Wear adopts a more rigorous approach in assessing member performance in this area, setting out clear and specific indicators for each issue.For example, buyer commitment to suppliers is assessed by indicators such as what proportion of a members' production volume comes from locations where the member company buys at least 10% of production capacity, and the percentage of production volume from business relationships that have existed for at least five years.Adequate pricing practices are determined by indicators such as the extent to which pricing policies allow for payment of a minimum wage and whether members assess and respond to the root causes of wages below a living wage. 79All member companies are scored and then given an overall ranking of 'leader', 'good', 'needs improvement'.Fair Wear publishes the full results of its Brand Performance Checks, allowing external stakeholders to assess the progress made by member brands.
The ETI also undertakes a headquarter level assessment of member companies through its annual reporting procedure.Member companies are expected to self-assess and report against an ETI Reporting Framework, the details of which are not public. 80A representative of the ETI stated that the reporting framework was 'structured along the lines of business and human rights due diligence' and required companies to identify, among other things, their relevant policies and governance processes, monitoring systems, salient human rights risks, responses to those risks, progress over time and future objectives. 81A senior representative of ETI further indicated that the ETI had recently begun to focus on the broader question of business models: So real impact is not a CSR thing and it's not even just the purchasing practices.If your business model depends on an unequal power relationship and an unfair share of value, you're going to see the sorts of outcomes for workers that we see, and sticking to the past approach of audit and compliance is not going to fix it.So we are trying to open up the debate more around business practices, not compliance, which is too often an add on.Real change will not be made by CSR, but it goes to the core of how you make business decisions in your supply chain. 82though this reflects a promising move beyond audit, ETI's ability to implement this approach through its reporting process is limited, due to ETI's limited assessment of the veracity of member reports. 83ETI member interviewees expressed different opinions on the reporting process.One smaller ETI company member found the ETI reporting process to be a helpful framework for preparing their own strategic plans and measuring progress. 84Another, larger, company member stated that 'the time it takes versus the benefit which is nil, it's not worth it'. 85Conversely, another company member argued that the ETI reporting process was 'mainly to inform debate with the unions and NGOs', concluding that the ETI 'isn't holding me to account.It's not an accountability organisation, it's a helping and engaging organisation'. 86The results of the independent review, combined with the divergent views of members, indicates that while the ETI does attempt to impose some level of control over its members through the reporting process, it does not currently compare in rigour or stringency to the headquarter level assessments undertaken by the FLA and Fair Wear.

Addressing complex rights
The Governance Apparel MSIs have typically sought to address complex rights through discrete training and projects.These kinds of activities can be classified as 'learning activities', which fit into the typical Learning MSI model.Yet such discrete activities will only ever make small inroads, as they are non-binding, and incapable of overcoming powerful political forces.As noted by a senior representative of the Bangladesh Institute of Labour Studies: We have a lot of capacity building program, training and everything … But what are they telling when [workers are] dismissed?You gain some people, you ask them that you should organise a trade union, this is their rights and the moment they start organising they are dismissed and then their photograph is circulated in the area so they are not getting any job from other factories.So what is the meaning of this training? 87e other tool increasingly utilised by Apparel MSIs to address complex rights is political engagement.Both Fair Wear and ETI have explicitly highlighted the need to pursue greater advocacy with governments to promote freedom of association 88 and the FLA has historically highlighted the importance of engaging governments to address 'systemic' challenges. 89Against this strategic backdrop, Apparel MSIs have made direct representations to supplier countries. 90hile project and advocacy work represents an attempt to address complex rights through techniques beyond audit, there is a real question as to whether non-binding approaches to these issues are sufficient.Without more, these efforts remain piecemeal and discrete, rather than comprehensive regulatory strategies.
It is important to acknowledge recent attempts by Fair Wear to take a more systemic approach to addressing two complex sets of rights: gender-based violence and harassment and freedom of association, both significant human rights challenges in the apparel supply chain. 91Fair Wear has now explicitly incorporated both rights into its 2022 Brand Performance Check and updated its associated guide. 92It requires buyers to incorporate a gender lens in their analysis of, and response to, human rights risks.In relation to freedom of association, it requires brands to commit to a sourcing strategy that privileges countries where workers can freely form or join a trade union and/or bargain collectively.It also requires buyers to engage with suppliers in respect of both gender and freedom of association prior to entering into purchasing orders.
While its impact remains to be seen, this recent development reflects a promising advance in addressing these issues in a more systematic, consistent, binding and measurable way.

Beyond audit
Original research conducted for the purpose of this study confirms unanimous acknowledgment of the limitations of social audit among a range of MSI stakeholders.Overall, all 33 interviewees agreed that social audit was an imperfect and, at times, problematic tool and one quarter of interviewees expressed the view that companies should cease social audit altogether.The remaining interviewees considered that while social audit had its limitations, it nonetheless remained a relevant tool in the overall social compliance toolbox.Those categorically against social audit viewed it as a cynical exercise intended to protect brand reputation.As stated by a representative of Anti-Slavery International (an ETI NGO member), it's about establishing plausible deniability for business … Auditing is a way of showing they are doing something but without changing any system. 93milarly, a representative of the Worker-driven Social Responsibility Network argued 'they've been able to use these social auditing programs as an excuse to really deflect pressure away and protect their reputations.' 94Other interviewees argued that although social audit was flawed, it formed a necessary basis for corporate relationships and decision-making, allowing companies to form a baseline assessment of supplier conditions and assess risk.
The preceding section argued that the very functional model adopted by Certification Apparel MSIs poses an obstacle to evolution beyond audit.In contrast, the Governance MSIs have each acknowledged the limitations of social audit and sought to move beyond it in at least some capacity.These efforts render Governance Apparel MSIs a significantly more legitimate form of regulation than Certification Apparel MSIs.They must now take further action to realise their regulatory potential. 95While the headquarter level assessments are a positive step beyond audit, these assessments should be strengthened and adapted to reflect a more committed, partnership-based approach with suppliers.This model, referred to in the literature as 'the commitment model', is a form of long-term capacity building. 96As noted by an independent expert, critical to its success is that buyers make concrete commitments to suppliers: So it's a partnership approach where they commit from the beginning to work with a certain amount of suppliers for a longer period of time and they make up for that commitment where they maybe sometimes have to accept slightly higher prices by helping those suppliers to improve productivity and training them on human rights … and by sticking around longer the chances for establishing improved working conditions are much better. 93Interview with a senior representative of Anti-Slavery International (4 March 2019). 94Interview with representative of Worker-driven Social Responsibility Network (n 11). 95Governance Apparel MSIs face a number of additional legitimacy challenges, including insufficiently diverse participation, a failure to demonstrate impact over time and a failure to move beyond the first tier.A number of reforms in this regard are also needed.The suggestions here are focused on addressing the limits of social audit. 96Locke and others, 'Virtue out of necessity' (n 29); Riisgaard and others (n 41) who emphasise the role of co-operation.
To this end, Apparel MSIs should more actively regulate the buyer-supplier relationship, impose more stringent controls in respect of purchasing practices (including clear consequences for a failure to progress), formalise opportunities for leverage among its membership base, and incorporate suppliers as active members with the ability to raise concerns about buyer conduct.
Further, Apparel MSIs must make greater strides to address the underlying social and political context of production.It is unrealistic to expect private transnational regulatory initiatives to, on their own, solve the complex ramifications of what is ultimately the failure of states to protect.However, Apparel MSIs must enrol both the states in which buyers are domiciled (home states) and the states in which production occurs (host states), to provide for better systems of human rights protection. 97Many scholars have written about the need for greater interaction or 'complementarity' between public and private regulation, 98 the UNGPs refer to a 'smart mix' of regulation, 99 and the UN has noted that weak involvement from governments may be one reason that MSIs have not reached their full potential. 100Accordingly, Apparel MSIs should undertake more sustained political advocacy, and also better connect to regulatory frameworks.
At the home state level, as mandatory human rights due diligence laws gain momentum (particularly in Europe), the ETI and Fair Wear have already actively lobbied the European Parliament in favour of such laws. 101The European Parliament's position on the draft EU Corporate Sustainability Due Diligence Directive (currently under negotiation) explicitly refers to multi-stakeholder initiatives, noting that by participating in MSIs, companies may support aspects of their due diligence, including to coordinate joint leverage, achieve efficiencies, scale up best practices, and seek expertise relevant to specific sectors, geographies, commodities or risk issues' (emphasis added). 102e potential for complementarity to advance the regulatory capability of MSIs is an area that will require further empirical testing and research.At the host state level, MSI advocacy should be connected to specific incentives, such as commitment to produce a certain quantity in a country over a period of time.It is vital that such advocacy incorporates the voice and demands of local civil society.The worker-driven initiative, Action Collaboration Transformation, provides an example of how advocacy for strengthened local human rights protections might be linked to the purchasing practices of buyers. 103inally, and most importantly, it is incumbent upon Apparel MSIs to explore alternate forms of monitoring and participation that strengthen the voice and concerns of workers. 104Apparel MSIs have been extensively critiqued on the limited and superficial participation of critical stakeholders-including local civil society, workers and other rightsholders-both in governance structures and implementation activities. 105Strengthened participation can be achieved through, among other things, embracing technological solutions for improving worker voice, setting up local systems of monitoring which empower local actors, and engaging with regulatory innovations such as the workerdriven social responsibility model. 106his article argues that Apparel MSIs must move beyond audit by taking these steps to preserve their legitimacy, effectiveness and relevance.It also acknowledges that there is a real question as to whether even such an evolution would be sufficient to overcome the broader structural issues identified in this paper.As demand for multi-stakeholder solutions to complex business and human rights challenges continues to grow, MSIs will remain an important and fruitful area of study.

Conclusion
This article has argued that ongoing reliance by Apparel MSIs on social audit to perform their regulatory compliance function poses a challenge to their regulatory legitimacy.It is now widely accepted by a range of stakeholders that social audit is not capable of delivering widespread and sustained human rights improvements in global apparel supply chains.First, there are a number of inherent problems in the manner in which social audit is conducted.Second, social audit is not capable of addressing the underlying business model and purchasing practices of buyers, or the broader social and political context in which production occurs.
This article has argued that the extent to which Apparel MSIs rely on social audit has implications for their ability to claim regulatory legitimacy.The legitimacy of the Certification Apparel MSIs-SAI and WRAP-is very low due to their almost exclusive reliance on social audit as a regulatory tool.In contrast, the Governance Apparel MSIs-FLA, Fair Wear and ETI-have each acknowledged the limitations of social audit and sought to move beyond it in at least some capacity.These efforts render Governance Apparel MSIs a significantly more legitimate form of regulation than Certification Apparel MSIs.Nonetheless, Governance Apparel MSIs must strengthen their attempts to move beyond audit by: adopting stronger oversight in respect of member company's purchasing practices; undertaking more sustained political advocacy directed at home and host states; harnessing greater complementarity between public and private regulation; and considering new techniques to incorporate the voice of workers and local civil society stakeholders into their regulatory efforts.In order to preserve their legitimacy, effectiveness and relevance, MSIs must be willing to address these pressing challenges and evolve accordingly.

104
Stefano Ponte and Emmanuelle Cheyns, 'Voluntary standards, expert knowledge and the governance of sustainability networks' (2013) 13(4) Global Networks 459; Emmanuelle Cheyns and Lone Riisgaard, 'The exercise of power through multi-stakeholder initiatives for sustainable agriculture and its inclusion and exclusion outcomes' (2014)31(3) Agriculture and Human Values 409.105Anner (n 10); AFL-CIO (n 11); S Prakash Sethi and Janet Rovenpor 'The role of NGOs in ameliorating sweatshop-like conditions in the global supply chain: The case of Fair Labor Association (FLA), and Social Accountability International (SAI)' (2016) 121(1) Business and Society Review 5. 106 See (n 11).