Horizontal Collaboration in Response to Modern Slavery Legislation : An Action Research Project

Purpose: To investigate how horizontal collaboration aids organisations in responding to modern slavery legislation and in gaining a socially sustainable competitive advantage. Design/methodology/approach: Action research has been conducted in the textiles & fashion industry and a relational perspective adopted to interpret five collaborative initiatives taken to tackle modern slavery (e.g. joint training and supplier audits). The primary engagement has been with a multi-billion pound turnover company and its collaborations with 35 brands/retailers. A Non-Government Organisation (NGO) and a trade body have also participated. Findings: Successful horizontal collaboration is dependent on both relational capital and effective (formal and informal) governance mechanisms. In collaborating, firms have generated relational rents and reduced costs creating a socially sustainable competitive advantage, as suggested by the relational perspective. Yet limits to horizontal collaboration also exist. Research limitations/implications: The focus is on one industry only, hence there is scope to extend the study to other industries or forms of collaboration taking place across industries. Practical implications: Successful horizontal collaborative relationships rely on actors having a similar mind-set and being able to decouple the commercial and sustainability agendas, especially when direct competitors are involved. Further, working with non-business actors can facilitate collaboration and provide knowledge and resources important for overcoming the uncertainty that is manifest when responding to new legislation. Social implications: Social sustainability improvements aim to enhance ethical trade and benefit vulnerable workers. Originality/value: Prior literature has focused on vertical collaboration with few prior studies of horizontal collaboration, particularly in a socially sustainable supply chain context. Moreover, there has been limited research into modern slavery from a supply chain perspective. Both successful and unsuccessful initiatives are studied, providing insights into (in)effective collaboration.


Introduction
Modern slavery is attracting significant media attention (e.g.The Guardian, 2016) while legislation has prompted discussion of this social sustainability issue in the academic literature (e.g.New, 2015).For example, new UK legislation requires organisations with a turnover greater than £36 million to publish an annual statement regarding action taken to combat modern slavery in their supply chains (UK Government, 2015).The following definition of modern slavery has been proposed: "the exploitation of a person who is deprived of individual liberty anywhere along the supply chain from raw material extraction to the final customer for service provision or production" (Gold et al., 2015, p. 487).The complex and global nature of modern supply chains means tackling this exploitation is challenging for firms to do alone (Gold et al., 2010).As a result, firms are looking beyond their boundaries, including by collaborating with competitors.Thus, there is a need to understand how this type of collaborationreferred to as horizontal collaboration (e.g.Touboulic and Walker, 2015a)can assist firms in combating modern slavery.It is also important to understand the impact of horizontal collaboration on competitive advantage in terms of social sustainability performance.
The literature has broadly defined supply chain collaboration as "multiple firms or autonomous business entities engaging in a relationship that aims to share improved outcomes and benefits" (Soosay and Hyland, 2015, p. 613).Thus, collaboration can exist in many formsinternally, externally, vertically, and horizontally (Barratt, 2004) and is often regarded as a deliberate strategy (Fawcett et al., 2010).In their content analysis of the collaboration literature, Soosay and Hyland (2015) found that research has concentrated on dyadic buyer-supplier vertical collaboration.Miemczyk et al. (2012) called for researchers to look beyond dyadic relationships by taking a network perspective that includes horizontal relationships and the roles of non-business actors.
Few authors have explored horizontal relationships in the context of Sustainable Supply Chain Management (SSCM), with Touboulic and Walker (2015a) suggesting the relational view (Dyer and Singh, 1998) would be an appropriate theory for such future work.More generally, supply chain research into modern slavery is limited, with Gold et al. (2015) calling for more empirical work.Further, although previous studies have considered the impact of environmental regulation (e.g.Sharfman et al., 2009;Ramanathan et al., 2014) there is a need to also examine the impact of social regulation, i.e. modern slavery legislation, on collaboration.
In the light of the above, this paper uses relational theory to investigate how horizontal collaboration aids organisations in responding to modern slavery legislation and in gaining a socially sustainable competitive advantage.It asks: How can horizontal collaboration, including the involvement of nonbusiness actors, aid organisations to gain competitive advantage in terms of social sustainability performance, in response to modern slavery legislation?
An action research approach has been undertaken where the primary engagement was with the evolving modern slavery related practices of a major international company in the textiles and fashion industry and its horizontal collaboration with 35 brands as well as two non-business actors (a Non-Government Organisation (NGO) and a trade body).Five collaborative initiatives were studied, leading to two novel contributions.First, we provide empirical insights into retailers' collaborative responses to modern slavery legislation.Second, we provide a theoretical contribution using a relational perspective.In particular, the concepts of relational rents, relational capital, and governance further our understanding of how horizontal collaboration enables firms to gain competitive advantage in terms of social sustainability performance.
The paper continues in Section 2 by reviewing the literature and then by outlining the research method in Section 3. Sections 4 and 5 respectively present the findings and the discussion.Section 6 contains the conclusions, including implications for research and practice.

Literature Review and Theoretical Background
An overview of the key horizontal supply chain collaboration literature is provided in Section 2.1 below.Section 2.2 then discusses horizontal multi-stakeholder collaboration, including collaboration with non-business actors, before the relational view is discussed in Section 2.3.

Horizontal Supply Chain Collaboration
Most extant literature concerning horizontal collaboration has focused on logistics collaboration, such as competing suppliers or retailers sharing containers and warehousing (e.g.Hingley et al., 2011) or non-competitive disaster relief logistics (e.g.Schulz and Blecken, 2010).In terms of collaborating rival organisations, an early example comes from the aviation industry, where competing airlines formed an alliance to enable entry into new markets and increase their global transportation network, attracting more customers (Oum and Park, 1997).Thus, research to date has shown that both competing and non-competing organisations can collaborate horizontally and that this can take place at different supply chain stages.However, there is a need to understand the benefits that can arise from horizontal collaboration in contexts other than logistics (and airlines).Chen et al. (2017) recently found that few researchers have studied horizontal collaboration in the context of SSCM.There are however a few papers that have considered horizontal collaboration between suppliers for a common buyer (Lim and Philips, 2008;Touboulic and Walker, 2015a) and between buyers (Nidumolu et al., 2014).For example, Touboulic and Walker's (2015a) action research in the food industry provided evidence of horizontal supplier-to-supplier relationships being formed during supplier meetings facilitated by the buyer.Although the main focus was on the vertical collaborative relationships between a large multinational buyer and each of its eleven small agricultural suppliers, the suppliers share their achievements and frustrations with each other during meetings, and this helps to shape the buyer's strategy.But it was also suggested that suppliers may become unwilling to share environmental information to retain a competitive advantage.Similarly, Lim and Phillips (2008) studied Nike's collaborative compliance model, which facilitated best practice sharing amongst competing suppliers resulting in improvements for the entire global value chain.This arguably could not have been achieved in their previous arm's length approach.
Horizontal collaboration between buyers has been illustrated by Nidumolu et al. (2014) using a case study of 'The Sustainable Apparel Coalition' (SAC) formed by Walmart and Patagonia.This alliance brought together competing brands, retailers, and manufacturers to improve sustainability performance within the industry by developing The Higg index.This index allows environmental indicators to be compared at a company, product, and factory level; and it encourages firms to compete on their sustainability ranking.In some cases, buyer collaborations have developed into 'meta organisations' (MOs), i.e. organisations made up of many members, which are becoming increasingly important for addressing corporate social responsibility issues that require collective action (Berkowitz and Dumez, 2016).Berkowitz et al. (2017) for example considered how corporate social responsibility (CSR) issues can be tackled at the industry/sectoral level through MOs.In their study of the oil and gas industry, a MO made up of 18 cross-sectoral major corporations improved CSR through standards setting, reporting guidelines, and capability building.

Horizontal Multi-Stakeholder Collaboration
In addition to collaboration between competitors, there has also been research into collaboration with external stakeholders such as NGOs or non-business actors (Bäckstrand, 2006;Mena and Palazzo, 2012).Hahn and Gold (2014)  (2012) confirmed that both vertical collaboration within the supply chain (customers and suppliers) and outside (NGOs and governments) were effective for enhancing environmental performance, reducing emissions and establishing innovative initiatives.
In the context of modern slavery, Gold et al. (2015) discussed the multi-stakeholder approach needed to remediate slavery with reference to initiatives in West African cocoa farms and tobacco sourcing in Kazakhstan involving buyers, suppliers, government, NGOs, communities, etc.The examples further highlighted the complementary resources, including local knowledge that non-business actors can offer to the relationship.However, there is scope to build on these insights by using first-hand empirical evidence to further understand the impact of non-business actors being present during horizontal collaboration.In particular, no prior literature has looked at this empirically in the context of modern slavery.

The Relational View
Although Dyer and Singh (1998) drew upon vertical buyer-supplier collaboration for illustrative purposes, relational theory has been extended to horizontal relationships by Walker et al. (2013), Hahn andGold (2014), andTouboulic andWalker (2015b).This theory is deemed appropriate as it takes an inter-organisational perspective, viewing the linkages between firms as important sources of competitive advantage.We focus on three key elements of the theory below: relational rents, relational capital, and governance.

Relational Rents
According to relational theory, resources and capabilities are more valuable when combined in unique ways, resulting in relational rents, i.e. supernormal individual firm profits (Dyer and Singh, 1998).The theory postulates that there are instances when this competitive advantage can only be generated through joint idiosyncratic contributions specific to the collaborating organisations (Dyer and Singh, 1998).Most of the extant SSCM literature using the relational view has looked at the rents that accrue from vertical collaboration alone and/or collaboration with external parties (e.g. Simpson and Power, 2005;Gold et al., 2010;Paulraj, 2011;Albino et al., 2012;Blome et al., 2014;Theißen et al., 2014).For example, Gold et al. (2010) presented a conceptual framework to show that inter-organisational collaboration on environmental and social issues can develop joint valuable and rare resources and capabilities that are difficult to imitate.As a result, firms can compete with rival supply chains or networks, simultaneously achieving economic, environmental, and social performance.Touboulic and Walker (2015a) demonstrated that this can be extended to horizontal supplier-supplier relationships.Their study however did not use the theory to provide an in-depth examination of horizontal collaboration, given its focus was also on the vertical buyersupplier relationship.
The concept of relational rents, although defined by Dyer and Singh (1998) as a supernormal profit, has been used more recently in the context of non-profit making organisations.For example, Hahn & Gold (2014) suggested that non-business actors such as CSOs (Civic Society Organisations) can generate a supernormal ability to meet their objectives of building public visibility and attractiveness to donors.There is however scope to enhance the concept of relational rents in the context of SSCM.To this end, factors can be identified from the SSCM literature that demonstrate how collaboration can contribute to relational rents, including leverage and risk mitigation (Nidumolu et al., 2014), supply chain transparency (Carter and Rogers, 2008), improved manufacturing performance, inter-organisational learning, knowledge sharing and expertise, resource sharing, and capability building (Vachon and Klassen, 2008;McDonald and Young, 2012;Blome et al., 2014;Touboulic and Walker, 2015a).In addition, Carter and Rogers (2008) showed how collaboration can reduce the costs of actions taken to improve sustainability, such as collaborative audits that lower transaction costs.

Relational Capital
The extant literature has demonstrated that effective collaboration is dependent on relational capital.According to Kale et al. (2000, p. 218) "relational capital refers to the level of mutual trust and friendship that arises out of close interaction at the individual level between alliance partners."Similarly, Touboulic and Walker (2015a) referred to trust and relationship history as examples of relational capital.In this paper, we expand on this by including other factors that impact the collaborative relationship.These include communication, commitment (Simpson & Power 2005;Verghese and Lewis, 2007;Vachon and Klassen, 2008;Theißen et al., 2014), and the role of absorptive capacity during knowledge transfer (Vachon and Klassen, 2008).

Governance
The relational view has also helped authors to consider governance for SSCM (Vurro et al., 2010).According to Dyer and Singh (1998), effective governance is important to the creation of relational rents with the authors distinguishing between third party enforcement (e.g. in the form of legal contracts) and self-enforcement.Similarly, in their study of BoP partnerships, Hahn and Gold (2014) considered both formal and informal governance mechanisms.Formal mechanisms included formal contracts used for strategic alliances supported by informal mechanisms, such as trust and mutual goals.Touboulic and Walker (2015a, p. 185) also identified support from top management as an effective governance mechanism; and they referred to the negative impact of a lack of effective governance, including a "misalignment of time frames for achieving sustainability goals".
In conclusion, research has applied the relational view to buyer-supplier collaboration and is profit oriented.Touboulic and Walker (2015a) is the only study to have considered horizontal collaboration in a SSCM context, and none of the papers have focused on social sustainability in relation to modern slavery.Thus, there is a need to further understand how horizontal collaboration, including the involvement of nonbusiness actors, can aid organisations to gain competitive advantage in terms of social sustainability performance.

Action Research Justification
Action research has become increasingly prevalent in the study of organisations (Easterby-Smith et al., 2012;Coughlan and Coghlan, 2016).It aims to influence practice and encourage change whilst providing contextual insights that facilitate theory building (Touboulic and Walker, 2015b).The researcher simultaneously takes action and creates knowledge (Coughlan and Coghlan, 2016).It therefore allows an in-depth understanding to be developed from participant observations and sometimes 'unspoken information' (Schoenherr et al., 2008).It relies on a collaborative approach where practitioners and researchers become 'co-researchers'; there is a mutual dependency on each other's skills, facilitated by a high level of trust enabling access to information (Hult and Lennung, 1980;Näslund et al., 2010).
There have been calls for SSCM researchers to undertake action research, with Touboulic and Walker (2015b, p. 309) arguing that the case study and survey methods "are not sufficient to provide an integrated view of SSCM phenomena".Modern slavery is a complex, sensitive issue, but adopting an action research approach has allowed us to work closely and develop trust with multiple brands, facilitating access to rich information.Furthermore, Coughlan and Coghlan (2016, p.237) stated that action research is "applicable to the understanding, planning and implementation of change in operations".The approach is therefore appropriate for understanding the change process involved as organisations look beyond their boundaries to tackle modern slavery.

Collaborating Organisations and the Unit of Analysis
The research team has engaged with the evolving modern slavery related practices of Fashion and Sports Co.a multi-billion pound turnover company in the textiles and fashion industryand its horizontal collaboration with 35 brands/retailers.This is an industry characterised by complex, global supply chains and high labour intensity (Bruce et al., 2004), making it vulnerable to modern slavery (Gold et al., 2015).The brands and retailers vary in size and sell a range of items from UK high street clothing to sports brands and luxury fashion.Fashion and Sports Co. is hereafter referred to as Brand 1 while the other collaborators are referred to as Brand 2 to Brand 36.In addition, two external parties that collaborated with the brands have also participated in this research: NGO and Trade Body.The researchers established the project scope and research purpose with Brand 1, which was outlined in a formal agreement, including the protection of all sensitive information.The core participants have been anonymised and mainly comprised of Corporate Responsibility (CR) managers within each brand and general managers for NGO and Trade Body.
The nature of the relationship was such that, in effect, one of the researchers worked part-time for Brand 1 over a 20 month period.An excellent level of trust was consequently built with Brand 1, facilitated by the background of one of the researchers who had seven years of relevant industry experience.This level of trust enabled access to detailed plans involving horizontal collaboration to address the issue of modern slavery whilst developing their first modern slavery statement.This trust in turn led to introductions to other brands at events where responses to modern slavery legislation were discussed.At these events, a number of joint initiatives were developed and one or more of the researchers were engaged actors in five of these initiatives, as listed in Table I.Thus, these initiatives are the embedded units of analysis in this study, where the primary unit of analysis is the response to modern slavery legislation.
[Take in Table I] Throughout the research project, the research team were mindful of the common pitfalls associated with action research.For example, as discussed by Näslund et al. (2010) and Coughlan and Coghlan (2016), it is important to ensure knowledge creation takes place so a contribution to academic theory is made as well as a contribution to the practice of the collaborating organisations.This was addressed by ensuring rigorous documentation and adopting a cyclical approach to the research involving reflection, as discussed in sections 3.3 and 3.4.Engaging in reflexivity is a key mechanism through which quality in action research can be ensured (Marshall and Reason, 2007).Other key elements include ensuring the researcher remains impartial (Koplin, 2005) and that there is transparency amongst researchers with regards to the choices being made during action research (Reason, 2006).Thus, regular meetings with all researchers and quarterly meetings between the researchers and a key representative from Brand 1 were held.

Data Collection
Multiple data collection methods have been used to provide triangulation and rich qualitative data (Näslund et al., 2010).Key data sources have included: participating in day-to-day activities, meeting minutes, discussion documents, observations, and interviews.Thus, the data takes a number of formats, including interview transcripts, meeting notes, and documents produced by the researcher and other participants.Diaries have, for example, been used to record key aspects of horizontal collaboration, including key observations and reflections (Coughlan and Coghlan, 2016).During meetings with multiple brands, 'Chatham House rules' applied, whereby participants discussed issues openly but their comments could not be attributed to them.Under these circumstances, quotes have not been recorded by brand or individual.
These multiple methods of data collection were possible given the action research approach and that the first author was actively engaged in the organisations' response to modern slavery legislation and the embedded initiatives.For example, the researcher participated in the planning and pilot stages of Initiative 4 (modern slavery training).This involved collecting data by conducting pre and post pilot training interviews with Brand 1 attendees, analysing the results, and sharing these with all collaborating brands.
The researcher also attended the pilot training session and de-brief meetings afterwards.This process of reflection therefore involved all collaborating brands participating in this initiative and led to changes in the materials used at subsequent iterations of the training.Key data collection methods during this initiative included interview transcripts, meeting notes, diary entries, and documentation such as training materials.
In addition to studying change within the organisations, the researchers also collected data on the nature of the horizontal collaboration taking place, including the relational rents generated, relational capital built, and governance mechanisms put in place.Thus, the data collection process also included tracking the commitment of brands as the initiatives evolved through a process of reflection that was recorded in diary entries.

Action Research Cycles
The action research framework outlined by Coughlan and Coghlan (2016) has been adopted to address rigour by engaging in multiple cycles of action.Each cycle contains a pre-step that involves understanding the rationale for action and four main steps involving constructing, planning action, taking action, and evaluating action.It is this process of evaluation or reflection involving the questioning of all aspects of the research that is distinctive to the action research process (Näslund et al., 2010).
Additionally, 'meta learning' ensures monitoring and reflection throughout.As illustrated in Figure 1, the action research project can be considered one major cycle (i.e.representing the main unit of analysis: the response to modern slavery legislation) with minor cycles (i.e. the five initiatives as the embedded units of analysis) taking place within the project (Coughlan and Coghlan, 2016).Initiatives 1 and 2 took place continuously and simultaneously throughout the engagement with Brand 1 whereas other initiatives followed on from each other, feeding into the next cycle (e.g.Initiative 3 took place before Initiatives 4 and 5).There has therefore been a spiral of action research cycles (Coghlan and Brannick, 2014).Detailed discussion of the steps involved in these cycles are given in Section 4.
[Take in Figure 1]

Data Analysis
Given that the five collaborative initiatives are used as embedded units of analysis, each one was first analysed individually through a within-initiative analysis followed by a cross-initiative comparison.This is akin to the within-case/cross-case analysis in case study research (Eisenhardt, 1989).It is applicable in the context of action research given that it is considered by some to be a specific form of case study research (Näslund et al., 2010).Nvivo software has facilitated data coding in two stages to identify themes and categories (Yin, 2009;Miles and Huberman, 2013).Themes either emerged inductively from the data or deductively from the literature, as shown in Table II.Each code was discussed and a final categorisation agreed amongst the research team.Tabulation aided the analysis (Coughlan and Coghlan, 2016), as further described in the findings below.
[Take in Table II]

Findings
As an example, Table III summarises the empirical evidence for Initiative 4, divided into relational rents, relational capital, and governance mechanisms.Similar evidence is available for the other initiatives and the major action research cycle (see Figure 1).Key aspects of the evidence are discussed below for the major and minor research cycles in sections 4.1 and 4.2, respectively before Section 4.3 provides cross-initiative analysis.
[Take in Table III]

Major Action Research Cycle: Collaborative Response to Legislation
In February 2016, 22 UK-based brands and NGO attended a meeting initiated by Brand 24 to discuss a shared response to the new modern slavery legislation.Brand 24's motivation for the meeting was recent media coverage that exposed their involvement in an instance of modern slavery in the UK.The head of CR had to give evidence in a court of law for 4.5 hours, during which he was asked to explain why his company audits did not identify modern slavery within their supply chain.He summarised that this was both a UK problem, "we are dealing with criminals" and an overseas "ingrained cultural problem".This exposure brought momentum, highlighting the need for the industry to come together to tackle modern slavery before others also found themselves in court.
Thus, this meeting was the catalyst for the five initiatives that followed.
NGO led the meeting and shared their expertise of the 'problem' and 'causes' of modern slavery.This centred on the 'Bait and Switch' concept, i.e. the offer of a great job in another country (the bait).This persuades the potential worker to pay a fee to a labour broker, but they are being miss-sold the job and end up in forced labour (the switch).There was broad acknowledgement that modern slavery is taking place in all of the brands' supply chains.Brands were for example already aware that workers' passports were being taken from them and that the presence of migrants increased modern slavery risk.It was evident that the brands were eager to tackle modern slavery but were also anxious as they were responding to modern slavery legislation for the first time.NGO promised to propose some actions and there was agreement amongst the brands that they needed to raise awareness within their businesses, such as through training and risk identification.After this meeting, two of the researchers and Brand 1 reflected on the discussions and worked with the CR team over several months to determine collaboration opportunities and the scope of a series of five initiatives.
As the initiatives progressed, Brand 1's strategic response to modern slavery began to take shape.This resulted in many drafts of their modern slavery statement, which included information on the initiatives being reviewed internally and with one of the researchers.Their first statement was published in September 2017 and illustrated that they perceived relational rents to have been derived from the horizontal collaboration.
Other brands also published their statements around this time, similarly referring to joint initiatives. Brand 36 sharing their 'Modern Slavery First Steps', which included a modern slavery workshop conducted in India for local staff, suppliers, and auditors.Focus was given to encouraging transparency and having conversations rather than formal interviews with workers regarding fees, deposits, and living arrangements.A key message was the need to emphasise to global suppliers that the brands are not in competition when tackling ethical trade issues such as modern slavery.

Minor Action
 A revelation of recent exposure to modern slavery in the UK "we got raided by customs and exile -it makes it an easy way to sell to the business that we need to do this [investigate modern slavery]".
 Brands sharing how to appropriately engage with high risk countries.One company showed videos to Indian factories of modern slavery in the UK.They explained "We didn't just say India is ranked high risk for slavery, we emphasised that this is a UK issue too".
The brands also frequently discussed limited resources and budget constraints particularly during tough trading, claiming that they "can't take resource into all of the tiny factories".
Collaborating therefore provides a platform to discuss how to share resources.A representative from Trade Body explained "Our approach is country wide risk assessment and working with companies to address collectively where their own leverage and resource can't achieve beyond managing risk in their own supply chain".
Relational capital was also both demonstrated and developed further at these meetings and workshops.In particular, it was demonstrated through their common mind-set as it was evident that CR representatives were very passionate, with one expressing "We [CR teams] live and breathe it [ethical trade]".The meetings also provided opportunities for CR teams to build informal relationships and foster mutual trust.Despite the intent to share supply chain information to aid in tackling modern slavery, confidentiality concerns amongst executives limit formal knowledge sharing.
Thus, opportunities to speak to each other informally at these meetings was important.
Further, some brands explained that they will informally call one another to update them on any instances of malpractice identified during visits to shared factories.It also became apparent that some brands have formed close relationships because of regularly collaborating in Trade Body working groups.A representative from Brand 1 explained "If you don't join one [Trade Body] working group they go off and create more so you can easily get left behind.It is therefore best to be in them all even if they are not completely relevant to your business." In terms of governance, brands acknowledged that there is heightened awareness and interest in modern slavery within their firms.This is particularly due to the legal requirement to issue a modern slavery statement, generating executive attention.
However, honest conversations took place regarding how to get executives 'on board'.only then do they start to pay attention."Thus, common problems were established during these meetings that were ripe for a collaborative response.This particular issue fed into Initiative 4 as training was tailored to the needs of different types of employees.
After Trade Body meetings, one of the researchers would de-brief and reflect with Brand 1's CR team to decide if any action was needed.Examples include meetings to discuss risk assessment and training, which led to the development of Initiatives 2, 3, and 4; and weekly team meetings with one of the researchers to evaluate progress.

Initiative 2: Purchasing Practices Project
Following ongoing Trade Body meetings and workshops, Initiative 2 was established in which a group of European brands reviewed their purchasing practices, thereby aiming to produce a set of guidelines to assist company buyers.As raised in the initial modern slavery meeting in February 2016, the brands are aware that their purchasing practices impact their suppliers leading to poor working conditions and increased modern slavery risk.Brand 1's Head of Ethical Trade argued "Most of our CR issues are created by us as an industry -how on earth can a factory plan capacity".Given that factories work with multiple brands, a collective step-change is needed to make an industry-wide difference, enabling factories to compete without having to compromise on working conditions.This initiative is possible due to high levels of relational capital as the brands are all known to one another and have previously worked together on other projects.As a result, there is a high level of trust, which enables an in-depth analysis and discussion of each other's purchasing practices in conjunction with the researchers.Thus, relational rents have been generated as each brand has provided their knowledge, expertise, and time to formulate a mutually agreeable set of guidelines that have been collectively refined through multiple iterations.This includes guidelines on planning/forecasting, price negotiations, production, and supplier relationships.Two of the researchers have been involved in refining the guidelines, followed by the iterative process of evaluating and refining them further.An example refinement was putting a greater focus on the specific demands buyers make that conflict with CSR requirements, such as short lead times that force suppliers to hire temporary/casual labour, which heightens modern slavery risk.

Initiative 3: Risk Matrix Project
Initiative 3 involved NGO producing a shared country risk matrix to help brands assess their current and future sourcing location decisions.This would also help brands prioritise their audit efforts as limited resources make it difficult to conduct in-depth investigations in every factory.The main purpose of the collaboration was to generate relational rents by spreading the cost of the risk matrix development amongst all collaborating brands, resulting in each brand having access to a risk profile of vulnerable worker populations across 22 countries.This would also free up resource as NGO would produce the matrix.A series of meetings took place between the different brands, NGO, and one of the researchers to scope out and plan the project.Yet, although there was enthusiasm for the initiative, communication problems and a lack of commitment resulted in time delays and the initiative ultimately not taking place.Initiative 3 could therefore be argued to have been unsuccessful given that each brand developed their own, separate risk matrix tailored to their individual needs.
Nonetheless, there were relational rents accrued from participating in the initial discussions, e.g. it helped each collaborator determine what was needed in their own organisation.Indeed, whilst developing their own risk matrix in collaboration with one of the researchers, Brand 1 regularly referred to the risk matrix information acquired from the initial collaboration with NGO and the other brands.Despite this individual development, a representative from Brand 1 claimed they would be willing to share the risk matrix with other brands for them to either individually or collectively use and adapt: "I want us to be confident our risk matrix is fit for purpose and donate it to our peers for their free adoption."This further demonstrates relational capital in terms of noncompetitiveness and the mind-set to work together to make industry improvements.

Initiative 5: Targeted Modern Slavery Audit
The feasibility and development of Initiative 5 was a direct result of Initiatives 3 and 4, as the risk matrix identified the priority factory to visit and the trained employees were equipped in understanding how to audit to uncover modern slavery.Therefore, in November 2016, Brand 1 partnered with NGO and travelled to Thailand to conduct a pilot modern slavery audit, accompanied by one of the researchers.Brand 1 invited another of the factory's key customers (Brand 18) to observe the audit process and increase pressure/leverage.This relational rent was important in this context, as explained by a representative from Brand 1: "there are not many factories in which we could influence policy without the support of other customers … The more leverage we have the more likely the factory is to engage".However, neither company had previously conducted a targeted modern slavery audit, and they were therefore heavily reliant on the expertise of NGO.Thus, knowledge sharing was with the NGO rather than between brands.It is also important to note that Brand 18 did not want to contribute financially to the audit.Brand 1 therefore agreed to let them observe on the condition that they "engaged [financially] in remediation".Brand 1 explained "we have opened their eyes to it [modern slavery issues] and we have then got a joint responsibility to change, to fix".This collaboration between Brand 1 and Brand 18 was successful despite a lack of initial relational capital as they did not have a prior relationship.They also had different mind-sets in terms of supplier development as Brand 18 were more willing to switch supplier.A representative from Brand 1 explained "Brand 18 have got a much wider sourcing network and they could feasibly and very quickly pull out".Developing relational capital during the audit was also difficult due to power struggles caused by the collaborating parties not having an equal investment in the initiative.As stated by a representative from Brand 1 "we don't want to cause an issue [create panic] if anything is found, any issues are on our terms".Despite these difficulties, trust was developed, as highlighted by Brand 18 adopting the role of translator during interviews with the factory's management and document review.Initially, this was done by one of the factory's managers but as a representative from Brand 1 claimed "The factory wouldn't have translated it all.[The representative from Brand 18] was good and picked up on issues."As the representative from Brand 18 was experienced and based in Asia, he had local insight and could use his expertise to further probe the factory's management.He also noticed anomalies and openly discussed these with Brand 1.Thus, relational rents were generated (in terms of the audit findings) at the same time as relational capital being built in terms of trust.
The audit took place over four days and was a reflective process throughout with a feedback loop built into the audit that involved a series of de-briefs between Brand 1 and NGO.This helped cross-reference findings and identify any changes to the audit process to follow up on key issues raised.Further, this reflective process helped to evaluate the pilot process and develop a modern slavery audit protocol for future audits.For example, Brand 1 concluded that if they were to repeat this exercise with Brand 18 or another brand, "we would need to agree in advance a lot more of the scope of it [the audit] and tell them exactly what our approach was and what we thought we were going to find."

Cross-Initiative Analysis
Table IV summarises the relational rents, relational capital, and governance mechanisms affecting each collaborative initiative studied.For each factor, an 'X' signals relevance (not relative importance) to a specific initiative, as shown for example for Initiative 4 in Table III.It therefore highlights how common the factors are across the five initiatives.
[Take in Table IV] In terms of relational rents, cost reduction was achieved in Initiative 4, which was one of the costliest initiatives involving a direct payment to NGO.Had Initiative 3 taken place, brands would have also benefited from a cost saving by dividing the expenses paid to NGO.It is also anticipated that a cost saving will be made at a later stage for Initiative 5 during collaborative remediation.Where a rent was not accrued, this was due to it not being an intended outcome of the initiative.In terms of relational capital, common factors such as trust, communication, commitment, and previous collaboration history impact success.The only differences appear for confidentiality concern, which was most prominent in Initiative 1 due to Trade Body meetings and workshops involving sensitive brand-specific information sharing.Finally, in terms of governance, all initiatives had both third-party involvement and top management support demonstrating these were of common relevance regardless of the nature of the initiative.

Discussion
In comparison to the prior literature, this paper makes four key contributions, which are discussed in turn below and lead to four propositions: 1.It finds that relational rents are generated through horizontal collaboration to achieve competitive advantage.
2. It provides empirical evidence to highlight how relational capital is developed in horizontal collaboration.
3. It provides empirical evidence to demonstrate the impact of both formal and informal governance mechanisms.4. It considers relational rents, relational capital, and governance in both successful and unsuccessful horizontal collaborative relationships.
Firstly, the study advances our knowledge on the creation of competitive advantage through horizontal collaboration.Although the concept of relational rents was initially defined by Dyer and Singh (1998) as a supernormal profit, it has been used more recently in the context of non-profit making organisations (e.g.Hahn and Gold, 2014).
In the context of our research, although the main focus is on business actors, the outcomes relate to social sustainability rather than directly to profits.Thus, the findings extend the literature by furthering our understanding of the benefits that can arise from horizontal collaboration in contexts other than logistics (e.g.Schulz and Blecken, 2010;Hingley et al., 2011).The contributions to relational rents vary by initiative, as shown in Table IV.
In two initiatives (initiatives 3 and 4), a key purpose of collaboration was to spread the cost of employing the NGO across brands.Similarly, previous studies have identified how collaboration can lead to overall cost savings (Verghese and Lewis, 2007;Vachon and Klassen, 2008).Relational rents have been further generated by increasing supply chain transparency and leverage (Carter and Rogers, 2008;Nidumolu et al., 2014), and by sharing knowledge, expertise, and resources.This has resulted in brands benefiting from capability building in each initiative, most notably developing internal capabilities for understanding risk and detecting modern slavery (Vachon and Klassen, 2008;Touboulic and Walker, 2015a).This is particularly important as the brands were dealing with the unknown, having not previously investigated modern slavery.Modern slavery is a complex issue, requiring a deeper level of investigation than many other social issues.
In addition to the above, relational rents are also being built in terms of reputation for appropriately responding to modern slavery legislation as organisations are undertaking more initiatives via horizontal collaboration than they could undertake alone.Thus, irrespective of the other outcomes of the initiatives, the majority of collaborating brands have publicised their involvement in joint projects in their modern slavery statements.
Thus we also build on previous literature regarding the influence of legislation on collaboration (Sharfman et al., 2009;Ramanathan et al., 2014) as the findings demonstrate how legislation can provoke horizontal collaboration between business actors.Further it is anticipated that modern slavery legislation will increase general awareness and create extra media and NGO attention; and that the media and other nonbusiness actors will study the brands' modern slavery statements thereby further incentivising them to undertake initiatives that reduce reputational risk and generate relational rents.This discussion leads to our first proposition: P1: Relational rents can be generated through horizontal collaboration in response to legislation and other external forces, thereby leading to competitive advantage for the business actors involved.In the case of modern slavery, new legislation is combined with media scrutiny and NGO pressure, leading to a collaborative response that generates relational rents in the form of cost savings, knowledge sharing, new capabilities, and enhanced reputation, thus achieving socially sustainable competitive advantage.
With regards to relational capital, the literature has identified trust as being an important mediating factor (Cheng et al., 2008;Sharfman et al., 2009;Theißen et al., 2014;van Hoof & Thiell, 2014) and suggested it emerges as a result of relationship history (Touboulic and Walker, 2015a).This work however is largely in the context of vertical collaboration.Our study suggests that trust is similarly an important component of a successful horizontal collaboration; however, in horizontal relationships, this is generally developed informally via relational aspects of networking and a relationship history from previous collaborative efforts.Yet the findings also provide an example of trust being formed quickly despite no prior collaboration, as demonstrated by an alliance between Brands 1 and 18 during Initiative 5 (Modern Slavery Audit).Trust is found to be of particular importance when collaboration is between competitors due to confidentiality concerns as firms seek to protect their individual competitive advantage.
For example, actors had concerns over sharing supply chain data, particularly during Trade Body meetings and workshops (Initiative 1).This is similar to Touboulic and Walker (2015a) who found that suppliers may be unwilling to share information to retain a competitive advantage in the context of environmental sustainability.Therefore, when competitors came together they generally attempted to put their individual commercial agendas to one side.This however was more difficult when the competitive and sustainability agendas overlapped.Consequently, gaining and maintaining trust is a very delicate issue when collaborating horizontally.
Many of the studied brands are attempting to integrate sustainability into their business rationale, and the initiatives described are contributing to placing sustainability at the forefront of decision making.For example, the purchasing practices initiative (Initiative 2) is helping to ensure procurement decisions take potential social sustainability consequences into account.Similarly, the modern slavery training (Initiative 4) equips employees with modern slavery knowledge to support their day-today commercial decisions.Thus, the findings confirm the common claim in the SSCM literature that the sustainability agenda needs to be integrated into the business rationale (Beske and Seuring, 2014).The findings therefore also demonstrate the importance of relational capital in terms of the brands' collective mind-set to work together to make industry improvements.This is particularly importance when forming horizontal collaborations and encourages commitment.Although there are differences in the brands' individual responses and modern slavery statements, their separate and collective actions represent the competency of the industry, especially to the media and other pressure groups.This discussion leads to the second proposition: P2: A successful horizontal collaboration is dependent upon building relational capital, underpinned by establishing trust and commitment between business actors.In the case of modern slavery, for trust and commitment to be established it becomes important for competing business actors to decouple their commercial and social sustainability agendas.
In terms of governance, the extant collaboration literature has found that formal mechanisms are supported by informal mechanisms such as trust (Dyer and Singh, 2008;Hahn and Gold, 2014).Likewise, our findings demonstrate that governance can be achieved through formal and informal mechanisms.However, in the context of horizontal buyer-buyer collaboration, there are some significant differences.For example, an informal governance mechanism was achieved via the involvement of third party, non-business actors.Initiatives 1, 3, 4, and 5 involved either the Trade Body or the NGO while Initiative 2 involved a global industry union that has not participated in the research.Although some of the initiatives involved a direct payment to a third party, brands were at liberty to pay for access to knowledge and resources and not use them.
The payment therefore does not lead to any formal governance.Instead, the third parties adopted an informal mediating rolefacilitating the collaboration and acting as a central point of contact.Our paper therefore also contributes to the literature on multistakeholder collaboration (Albino et al., 2012;Hahn and Gold, 2014;Rodriguez et al., 2016) by providing evidence of buyers that have collaborated with non-business actors.
Further, it provides evidence of a multi-stakeholder approach being adopted to address modern slavery in particular, as suggested by Gold et al. (2015).The non-business actors were of particular importance to bridging the gap in modern slavery knowledge.
Modern slavery is a criminal issue leading to complex repercussions while there was also concern regarding how the media would portray brands if any issues were uncovered in the published statements.NGO are experienced independent experts in modern slavery and their support and resources were imperative.
In addition, our study has provided evidence of legislation being a formal governance mechanism.As in previous studies (e.g.Touboulic and Walker, 2015a), top management support generally provided effective governance, but this was as a direct result of the modern slavery legislation that stipulated statements must gain boardroom approval.This ensured CR teams received business support to collaboratively tackle modern slavery within their supply chains.The findings also provided additional evidence of self-enforcement achieved through collaborating brands appointing a lead from each company.This helped, for example, during the development of the training curriculum in Initiative 4. Therefore, the following proposition (3a and 3b) is presented: P3a: Informal governance mechanisms can be created by involving non-business actors that facilitate and support horizontal collaborations.In the case of modern slavery, non-governmental organisations and trade bodies can play an important role in facilitating, mediating, and monitoring collaboration between business actors, including by sharing knowledge and resources.
P3b: New legislation can create formal governance mechanisms that drive firms to make improvements.In the case of modern slavery, publically released statements require boardroom level approval, and this prompts top management involvement that acts as an effective governance mechanism for improved standards and transparency within the supply chain.
The findings from our study use the relational view to understand relational rents, relational capital, and governance based on evidence from successful and unsuccessful horizontal collaborative initiatives.Touboulic and Walker (2015a) called for a greater understanding of the difficulties of achieving collaboration.Hence, the unsuccessful initiative, i.e. the risk matrix project (Initiative 3), provides a further contribution.In particular, it was apparent that the results emerging from this initiative were too broad signifying the importance of collaborating when brands have similar needs.This extends the previous literature that has considered this from a buyer-supplier perspective only (Simpson and Power, 2005;Touboulic and Walker, 2015a).In addition, although the joint training initiative (Initiative 4) was a success, some brands left the scheme as time progressed whilst others augmented the material.Despite the erosion of relational rents, findings from Initiatives 3 and 4 have, for example, demonstrated that the brands were able to develop absorptive capacity to generate relational rents within the collaborative relationship, providing empirical evidence to support Vachon and Klassen (2008).This allowed the brands to create additional value by working independently to develop individual risk matrices and training programmes.
Further, in all four successful initiatives, the level of commitment varied across brands, with significant fluctuation over the course of the action research.This provides additional insights into generating relational capital for effective collaboration.
The evidence highlights the impact that a lack of effective governance can have even in those initiatives deemed successful.This adds further evidence to the conclusion of Touboulic and Walker (2015a, p.185) who identified the negative impact of a lack of effective governance mechanisms, such as "misalignment of time frames for achieving sustainability goals".In our findings, a lack of communication could be seen as a lack of commitment towards a given initiative, causing delays, leading to frustration amongst collaborating parties.This is a result of multiple parties collaborating coupled with the fact that each firm is working at its own speed to progress issues internally.Our findings also show that a lack of effective governance mechanisms can result in power struggles in the absence of a formal agreement.This was evident in Initiative 5 (Modern Slavery Audit) as a result of the collaborating parties not making an equal investment.
The discussion now closes with our final proposition (4a to 4c): P4a: Horizontal collaborations that fail to meet a collective objective can still generate relational rents at the firm level.In the case of modern slavery, the knowledge gained from an initial collaboration can be used to inform and enhance the response of individual business actors for improving standards and transparency within the supply chain.
P4b: Horizontal collaborations can fracture over time, thereby eroding relational capital.In the case of modern slavery, business actors that initially collaborated may work individually when their goals diverge, they seek to maintain individual competitive advantage, or to tailor the initiative to reflect their own specific business requirements and supply chain characteristics, undermining the trust that has been developed through their relationship.
P4c: Horizontal collaborations can fail or be delayed when there is a lack of effective governance.In the case of modern slavery, a lack of effective (informal and formal) governance can create power struggles and delays as business actors seek to meet their own targets for the collaboration and bring about change internally within their own organisations.

Conclusions
Few prior studies have explored horizontal collaboration in the context of SSCM, with the majority of this work being on environmental sustainability.Meanwhile, there is a lack of research into modern slavery from a supply chain perspective.Thus, this study has adopted a relational view using relational rents, relational capital, and governance to explore horizontal collaboration in response to modern slavery legislation.Action research has been used to understand how horizontal collaboration, including the involvement of non-business actors, has helped organisations gain a socially sustainable competitive advantage in this context.The findings reveal that firms are using collaborative initiatives to build new capabilities that improve social sustainability performance by generating relational rents in terms of the organisations' reputations for appropriately responding to modern slavery legislation.Further, the findings have highlighted the relational capital and governance mechanisms supporting or hindering successful horizontal relationships thereby extending the SSCM literature that has focused on vertical relationships.In terms of relational capital, trust is of particular importance between competitors and, in collaborating, competitors have to put their individual commercial agendas to one side to improve sustainability.The paper also highlights the role of non-business actors as an effective informal governance mechanism, facilitating horizontal collaboration whilst also providing much needed knowledge and resources for tackling a complex social issue.

Managerial Implications
This research provides managers with examples of how collaborative relationships can be formed in response to new legislation.Prior to modern slavery legislation, firms have not had this level of exposure to a criminal issue or had the threat of court appearances, which adds another dimension to social sustainability practices and reporting.
Collaborating with other firms, particularly when there is uncertainty, can therefore help to develop an initial response, share expertise, and distribute costs.
When forming horizontal collaborations, it is important that firms have a similar mind-set and can decouple their CSR and commercial agendas, especially when collaboration involves direct competitors.Additionally, firms need to foster trust; hence, prior relationship history can be important to project success.It is also key that horizontal collaboration has top management support.This can enable information sharing, which can result in industry level improvements, and promote self-governance by ensuring CSR has heightened exposure throughout the business.It is important that CSR and specifically modern slavery is central to business decisions, particularly relating to purchasing practices.Modern slavery awareness therefore needs to be filtered throughout the organisation, which can be achieved by training all levels and departments.
Collaborative relationships can be facilitated by non-business actors.Involvement in trade bodies, for example, can help identify potential collaborators.Both NGOs and trade bodies can act as a central point of contact and share resources and expertise.
Enlisting the help of experts in areas such as modern slavery can also prove beneficial.
Once potential collaborators have been identified, time needs to be spent scoping collaborative initiatives.During this initial stage, it is important to consider each firm's individual needs.Collaboration can become counter-productive when firms are not sufficiently committed and do not communicate effectively.This can lead to projects being too time-consuming without sufficient pay-off, leading to frustration and delays.It is therefore important that firms communicate with one another throughout.
Companies can also benefit from appointing a lead for each initiative to aid communication and progression.It should however be noted that even if a collaborative initiative is unsuccessful, the initial scoping and involvement can help with sensemaking, providing a foundation for a firm's own individual development or response.

Limitations and Future Research
We have focused on horizontal collaborations that involved Brand 1 only.There may have been other collaborations within the group that we were not aware of or able to study.Further, our research is focused on responses to the UK modern slavery legislation.Future research could consider how organisations are responding to legislation in other countries.Other industries could also be studied while there is an opportunity to consider whether firms should engage in cross-industry collaboration where competition is not a factor.Finally, further analysis could be undertaken of the involvement of third parties such as trade bodies and NGOs to explore the development of relational capital, their governance role, and their contribution to developing relational rents.
for example have considered collaboration amongst business and non-business actors when implementing Base of the Pyramid (BoP) projects.These non-business actors contribute non-tangible resources (such as local market expertise, information, and know-how) and facilitate trust amongst other BoP actors.Other studies have researched collaboration with NGOs; for example, McDonald and Young (2012) investigated the evolving relationship between Greening Australia (NGO) and Alcoa (an Australian mining company) where collaboration improved the reputation of both organisations.Similarly, Rodriguez et al. (2016) researched the benefit of six firms collaborating with an NGO to implement supplier development programs to alleviate poverty.Likewise, Albino et al.
Research Cycles: The Five Collaborative Initiatives 4.2.1 Initiative 1: Trade Body Meetings & Workshops Many of the brands that attended the first meeting are members of an ethical trade body (Trade Body).Attendance at Trade Body meetings and workshops (Initiative 1) facilitated the four other initiatives described below, with four key meetings focusing on modern slavery during the research project, whilst the meetings themselves contributed to relational rents through knowledge sharing.Examples of knowledge sharing include: One brand explained "They [executives] glaze over [on ethical trade issues] and don't have time for training to be brought up to speed.They [executives] only have awareness of modern slavery due to the law and that's due to the company secretary [panicking],

A
group of 21 brands initially expressed an interest in this initiative following Trade Body meetings (Initiative 1), but this reduced over time to just six brands.Five brands in particular that were all members of Trade Body and had strong relational capital (given their relationship history) drove the training.Key motivators for collaboration included factors associated with relational rents such as cost saving and a lack of resource at the individual organisational level.For example, the CR Manager from Brand 1 explained "Training with [NGO] is costly and in Indonesia, for example, we only have a few employees [too few to hold a training session].We need to get other brands involved with training [to consolidate training in regions]".The companies collaborated from April 2016 onwards during both the development of the training programme curriculum and the delivery itself.A key action during this initiative was piloting the training scheme, which took place in July 2016 and one of the researchers was heavily involved at this stage.This was the most reflective part of the process and involved three stages: (1) a large group sharing initial thoughts; (2) individual brands reflecting internally to suggest amendmentsin the case of Brand 1, this included pre and post pilot training interviews conducted by one of the researchers; and (3) the formation of a smaller group to make the final revisions.Evaluations from the reflective process were to make the training more interactive and include more practical tools, exercises, and videos.It was also concluded that there needed to be a clearer structure and that the training should demonstrate the impact of modern slavery for both the victims and businesses.This led to the first rollout of the training scheme in Asia in September 2016 attended by 40 employees from four brands.It included two half-day buyer risk-flagging sessions ('Avoiding Modern Slavery in Supply Chains') designed to help buyers identify and respond to risk indicators in the systems and practices of suppliers.The impact of the collaboration and reflective process on relational rents was summarized by an attendee of the first training session who claimed "Having other brands involved gave the training a different dynamic.Attendees had different levels of experiencethis gave a lot of richness to the discussion and we gave each other advice".Although this joint training was a success, some brands left the scheme as time progressed while others augmented the material.For example, in May 2017, Brand 1 conducted further individual training.An internal project team was created to plan and produce this training material using the generic material as a starting point but tailored to the company's own supply chain and business requirements.A lack of communication regarding individual plans caused delays to the joint training scheme development, to some extent eroding relational capital.