Southern sustainability initiatives in agricultural value chains: a question of enhanced inclusiveness? The case of Trustea in India

Recent studies have shed light on the emergence of Southern sustainability initiatives in commodity-based value chains. These initiatives position themselves as countering the exclusionary nature of many global multi-stakeholder initiatives (MSIs), as critically analysed by previous studies. However, a common theoretical perspective on the inclusiveness of MSIs is still lacking. By drawing on the theory of regimes of engagement, we develop a theoretical framework which helps understanding the overt and subtle practices of including or excluding different stakeholders in MSIs. We apply this framework to ‘Trustea’, an Indian MSI for sustainable tea production, to further investigate the ‘inclusiveness paradigm’ of Southern sustainability initiatives. Our findings highlight the complexity of stakeholder engagement across the different phases of Trustea’s development, from initiation and code development to formal launch and roll-out. We find that different types of engagement can exist in parallel. Justifiable engagement is expressed in the participating organisations’ reference to a shared common goal of Trustea, whereas familiar engagement is linked to the rootedness of the individuals involved in the Indian tea industry. Strategic engagement, in turn, comes to the fore through the strong emphasis on developing an Indian code for sustainable tea production and promoting adoption among tea growers. The importance of reaching this objective has led to the replication of exclusionary patterns also noted for global MSIs, as small-scale producers and other weak actors have been largely excluded from decision-making processes, despite informal efforts aimed at substitution of representation.


Introduction
Over the last two decades, multi-stakeholder initiatives (MSIs) have become important governance arrangements to respond to pressing, often global sustainability challenges (e.g. Fransen 2012; Mena and Palazzo 2012;Zeyen et al. 2016). MSIs typically involve businesses, non-governmental and other stakeholders with divergent interests working together in a collective process to develop rule-making authority which is not supported by governmental regulation but relies on market mechanisms. MSIs have proliferated widely, particularly in commodity-based value chains where they often aim to regulate the social and environmental impacts of production processes, e.g. in coffee, palm oil, sugar cane, cotton and biofuel production, by means of standards and certification schemes.
The signature design feature of MSIs is inclusiveness, as they involve a variety of stakeholders in their governance and standard-setting activities (Schleifer 2019, p. 52). By using the label 'multi-stakeholder', these initiatives are keen to claim legitimacy based on balanced representation and participation of "all categories of stakeholders", especially during the standard-setting process (Cheyns and Riisgaard 2014, p. 40). This is critical in the absence of formal rulemaking authority (Schouten and Glasbergen 2011;Mena and Palazzo 2012). Stakeholder participation also serves as a means to empower marginalised stakeholders, especially smallholder producers (Auld et al. 2015). Yet, the extent to which MSIs can be inclusive of all relevant interests and overcome unequal power relations in global trade is heavily debated (Bitzer and Glasbergen 2015).
While a common theoretical framework on the inclusiveness of MSIs is still lacking, research has addressed different elements of the inclusiveness debate. Critically, many studies note a lack of equal access to and influence in decisionmaking for groups representing the interests of producers and local communities in the global South (Cheyns 2014;Bennett 2017;Renckens and Auld, 2019). In addition to the fact that participation in MSIs demands considerable resources, exclusion of Southern producers and local communities also results from specific forms of engagement, knowledge and debate practices in MSIs (Cheyns 2011(Cheyns , 2014Ponte and Cheyns 2013). Higgins and Richards (2019) refer to this as the 'invisibilisation effect', in which particular actors and their claims and interests are marginalised or excluded from MSIs. This has evoked substantial criticism against a type of multi-stakeholderism which ignores power imbalances between actors and privileges the interests of transnational corporations (McKeon 2017).
Realising the need for enhanced inclusiveness, prominent schemes, such as the Roundtable on Sustainable Palm Oil (RSPO), have recently developed targeted strategies to engage smallholders more effectively and to achieve a better connection with local actors and contexts (e.g. Johnson 2019). Yet, the tendency to marginalise or exclude Southern actors through uneven participation and unequal representation of interests in MSIs has also prompted the emergence of Southern sustainability initiatives as an explicit countertrend to Northern MSIs. In different sectors where Northerndriven MSIs have tended to dominate, including palm oil and soy, actors from Southern producer countries have developed their own 'home-grown' sustainability initiatives and standards to challenge the hegemony of Northern MSIs (Hospes 2014;Schouten and Bitzer 2015;Hidayat et al. 2018). Thus far, most attention has been paid to the sustainability standards developed in Indonesia and Malaysia in response to the RSPO (e.g. Hospes 2014; Hidayat et al. 2018;Higgins and Richards 2019). Here the exclusive dynamics have been most evident since the RPSO's beginnings in 2004, as the majority of the RSPO's members stem from the European private sector, whereas actors from Malaysia and Indonesia, where 85% of global palm oil production takes place, are underrepresented (Ruysschaert et al. 2019).
The rise of Southern sustainability initiatives can also be observed in other important export producer countries or regions, such as Brazil (for soy and coffee; Hospes 2014), South Africa (for fruit and wine; Schouten and Bitzer 2015), Southeast Asia (aquaculture; Sun and van der Ven 2020) and India (tea; Langford 2019). Initial research indicates that Southern sustainability initiatives aim to develop standards which focus on creating buy-in with local producers, facilitating context-specific applicability and ensuring economic sustainability of production (Schouten and Bitzer 2015;Sun and van der Ven 2020). By developing standards that are relatively easy to meet, 'inclusion' of smallholder producers is seen as an important outcome of Southern sustainability initiatives (Wijaya and Glasbergen 2016;Hidayat et al. 2018). Thus, Southern sustainability initiatives could address the exclusion generated by MSI-developed standards and broaden stakeholder engagement, especially by smallscale producers (Higgins and Richards 2019).
However, since the emergence of Southern sustainability initiatives is still a relatively new phenomenon, our understanding of these initiatives is limited. Little research has been conducted to assess the actual degree of inclusiveness of Southern sustainability initiatives and possible implications for broader sustainability governance. This paper therefore investigates the 'inclusiveness paradigm' of Southern sustainability initiatives, researching the dynamics of inclusion and exclusion of small-scale producers.
The empirical case for this paper is 'Trustea', an Indian multi-stakeholder initiative for sustainable tea launched in 2013, which aims at the domestic tea market (Langford 2019). By drawing on 'regimes of engagement' of pragmatic sociology (Thévenot 2007(Thévenot , 2011(Thévenot , 2014, the paper analyses how Trustea has engaged with different actors within a multi-stakeholder setting. Particular attention is paid to the expression and consideration of the interests of smallholder growers. Our paper contributes to existing literature in the following manner. First, we contribute to the debate on MSIs by unpacking the theoretical notion of 'inclusiveness'. While regimes of engagement theory has been utilised previously in research on MSIs (Cheyns 2011(Cheyns , 2012Cheyns and Riisgaard 2014), what is lacking thus far is an operationalisation of how different forms of engagement give rise to dynamics of inclusion or exclusion of specific stakeholder groups. Our aim is not to reproduce the normative assessments which seem to dominate studies on MSIs and which often result in one-dimensional criticism of these arrangements (Glasbergen 2013). Rather, we attempt to build and apply a theoretical framework to help understanding the overt and subtle practices of including or excluding different stakeholders in MSIs, based on how actors 'engage' in a specific setting. This complements recent literature which discusses inclusiveness in MSIs in relation to institutional design and structural power (Bennett 2017;Schleifer 2019;Renckens and Auld 2019), but offers a new perspective by looking at the socially constructed types of engagement which shape and are shaped by institutional structures.
Second, we connect to the emerging literature which explores how practices of global MSIs are received, modified, contested and mimicked by 'Southern' actors (e.g. Hospes 2014;Schouten and Bitzer 2015;Wijaya and Glasbergen 2016;Hidayat et al. 2018;Sun and van der Ven 2020). As global MSIs face increasing legitimacy and effectiveness challenges, responses by Southern actors, especially by governments and industry associations, are likely to increase. This is also connected to the growing importance of South-South trade, which creates new market opportunities for producers of agricultural commodities that are not subject to the sustainability demands from European or American buyers (Schleifer and Sun 2018). The present paper on the Indian code for sustainable tea can therefore offer relevant insights on the inclusiveness of sustainability standards developed in the global South. As such initiatives are considered providing an important response to the invisibilisation and exclusion generated by many global MSIs (Higgins and Richards 2019), detailed analysis is necessary to understand their level of inclusiveness.

The inclusiveness of multi-stakeholder initiatives in agricultural value chains
Multi-stakeholder initiatives (MSIs) have been defined as "initiatives governing social and/or environmental standards of production that have participants from both business and society interest groups as members and governance structures allowing for an equal possibility of input among the different partners in steering the initiative" (Fransen 2012, p. 166). The legitimacy of MSIs is crucially based on balanced representation of and participation by all categories of stakeholders. The embodiment of inclusiveness is particular relevant when it comes to setting voluntary standards to govern the sustainability of agricultural production, which lies at the heart of many MSIs (Cheyns and Riisgaard 2014). Inclusive processes of standard development and enforcement do not only serve to establish legitimate decision-making power, but they can also contribute to the empowerment of peripheral actors in the global economy (Auld et al. 2015). For these reasons, inclusiveness has been labelled the key design feature of MSIs (Schleifer 2019). However, other studies have criticised that the rhetoric of MSIs does not match their practical implementation and have exposed various aspects of 'exclusion' in MSIs (Cheyns 2011(Cheyns , 2012Bitzer and Glasbergen 2015;Wijaya et al. 2018). This has led to claims that Southern stakeholders, their perspectives, needs and issues are marginalised in global MSIs (Higgins and Richards 2019).
While this debate has prevailed since the early emergence of MSIs (e.g. Bäckstrand 2006;Dingwerth 2008), little theorising has been done on what inclusiveness actually entails in the context of MSIs. To disentangle the 'inclusiveness paradigm' of MSIs, we therefore focus on three questions which have so far been critically addressed in the scientific literature. These questions loosely follow Quack's (2010) normative dimensions of rule-making in global governance: (1) How are actors able to participate in MSIs?; (2) how are different principles of justice and voices recognised in MSI processes?; and (3) how are different forms of knowledge drawn upon, discussed and employed in MSIs? Inclusiveness thus carries a clear procedural connotation, essentially focusing on whether and how actors can contribute to MSI processes in a meaningful way.

Inclusion of stakeholders
First, inclusiveness of MSIs can be approached from the perspective of inclusion, i.e. the extent to which stakeholders affected by the issue can participate in the structures and processes of MSIs (Mena and Palazzo 2012). Although many MSIs have set up governance structures to ensure representative inclusion of different stakeholder categories (e.g. Mena and Palazzo 2012), many studies show the shortcomings of stakeholder participation. The majority of MSIs in agricultural value chains have been initiated and driven by actors from the global North, outweighing the participation of actors from Southern producing countries, both quantitatively and qualitatively (Dingwerth 2008;Fransen 2012;Bitzer and Glasbergen 2015;Moog et al. 2015). Even as formal barriers to joining may be low, de facto participation is often tied to the availability of resources and capabilities (Bennett 2017;Dentoni et al. 2018). This creates a wave of exclusion of those actors who are not able to organise themselves, e.g. due to lacking resources or unfamiliarity with multi-stakeholder settings. Particularly vulnerable groups, such as smallholder farmers and local communities, are often found to be poorly included in MSIs (Ponte 2008;Bitzer and Glasbergen 2015). In some MSIs, it has therefore become common for NGOs to represent the interests of stakeholders who are not able to participate-although the legitimacy and effectiveness of such representation has been questioned (Cheyns 2011).

Plurality of principles of justice
The second element of inclusiveness relates to the acceptance of critical and disputed discourses within MSIs, independently of how they are formally represented (Arenas et al. 2020). This goes beyond recognising a plurality of voices, but equally denotes facilitating a plurality of principles of justice, i.e. considerations of 'what is just' in view of MSIs' ambition to promote universal norms (Cheyns 2011;de Bakker et al. 2019). While MSI processes are widely portrayed as deliberative and consensus-oriented, studies have questioned the empirical grounding of this claim. On the one hand, time constraints and urgency can lead to negotiating pre-existing solutions rather than creating a moral basis around a shared common good (Cashore et al. 2007). On the other hand, participation formats can be restrictive in the discourses that are considered acceptable. Studies have shown that subjects which are not perceived as 'unpolitical' or 'neutral' are quickly dismissed as off-topic (Schouten et al. 2012;Elgert 2012Elgert , 2012, and discourses which are not considered having the appropriate rhetorical form are often excluded (Cheyns 2011).

Different forms of knowledge
Finally, inclusiveness refers to the forms and sources of knowledge allowed and employed in MSI processes. From a problem-based perspective, a major advantage of MSIs is their ability to involve actors who cut across different knowledge domains (Dentoni et al. 2018). However, integrating diverse knowledge input appears to be challenging for MSIs; importantly, because knowledge from international experts with technical expertise enjoys higher credibility than the experiences of affected stakeholders who may not be able to present their views in a format that is considered adequate in an MSI setting (Cheyns 2011;Wijaya et al. 2018). This leads to the dominance of technical expertise rather than local experiences (Ponte and Cheyns 2013). In their study on the Roundtable on Sustainable Palm Oil, Offermans and Glasbergen (2015) found that knowledge input included different global and local sources; yet, the use of these different inputs was rather restricted. Knowledge from smallholder farmers, in particular, did not influence decision-making due to procedural constraints and a lack of knowledge management within the MSI (Offermans and Glasbergen 2015). For the Roundtable on Responsible Soy, knowledge was purposefully filtered to centre on "facts that are difficult to dispute" and to show its utility in exploring possible solutions for win-win situations (Elgert 2012, p. 300).

Inclusiveness and engagement theory
Understanding inclusiveness therefore involves examining how MSIs operate not only within a plurality of stakeholders, but also within a plurality of principles of justice and different forms of knowledge. Engagement theory argues that this depends on the regimes of engagement that predominate in a given setting, i.e. the way that actors 'engage' in MSI processes and which elements enable them to make themselves heard (Thévenot 2007;Ponte and Cheyns 2013). Thévenot has defined three forms of engagement, based on different 'cognitive formats' of the environment where actors feel 'ease' and 'self-confidence'. These shape the communication formats used and the degree to which actors feel capable to engage.
Firstly, in justifiable engagement, actors engage in participation in the name of the common good and recognise themselves as moral subjects within a pluralist setting (Thévenot 2007). The concerns of a single actor are therefore related to a common good and go beyond the specific interests of the actor (Cheyns 2011). The recognition of multiple perceptions of the common good allows a form of 'critical participation' where inequalities between actors become evident and are purposefully addressed, often by drawing on scientific knowledge. According to Boltanski and Thévenot (1991), justifiable engagement is the cornerstone of the modern public sphere, where different principles of fairness and justice are key to establishing legitimacy of people's engagement.
Secondly, in strategic engagement, actors assess a situation and engage in participation according to their interests. As interaction is limited to those with a stake (Centemeri 2015), managerial and technical capacities of actors become critical to accomplish a desired action. Contrary to justifiable engagement, differences between actors are not expressed by calling upon a common good. Rather, the common good tends to get lost in the idea of an accomplished action (Thévenot 2007). This is thus a highly functional form of engagement, where actors use strategies that are instrumental to their own interests and draw on technical knowledge. As the focus lies on utility, i.e. realising (mutual) plans, public legitimacy of this type of engagement is much more limited.
Finally, in familiar engagement, actors interact with their environment through personal relations and attachments. Participation is limited to those who are familiar with a specific environment. Actors engage in affected and emotional participation, sharing their personal experiences and attachment to their environment (Thévenot 2014). Knowledge exchange is based on practical experiences, perceptions and emotions, which explains that the forms of knowledge used do not lend themselves to extended communication (Thévenot 2007). As this type of engagement is based on shared exposure to a concrete material environment (Thévenot 2007), it can be used to mobilise local action to solve the problems and challenges that these actors encounter. Public legitimacy is restricted to those familiar with the environment.
These three forms of engagement therefore correspond to how actors interpret and respond to a situation, which enables them to coordinate with other actors who share their evaluative judgement of the situation (Centemeri 2015). Engagement theory refers to this as 'creating commonality' among actors. However, coordination and agreement with actors who do not share this judgement is challenging. Disagreement is likely to emerge on who should be involved in engagement processes, how these processes should take place and which weight and voice to grant to specific actors, values and norms, and knowledge and expertise.
Referring back to the three elements of inclusiveness identified above, different regimes of engagement thus have fundamentally different implications for the ability and power of actors to become involved, for the principles of justice drawn upon, and for the forms of knowledge used as underlying the cognitive formats that characterise actors' access to reality (Thévenot 2007) (see Table 1).
This does not imply that different formats of participation are necessarily more or less 'successful' according to some pre-set ideal-instead they represent different ways of organising participation for the expression of stakeholder voice (Meilvang et al. 2018). At the same time, there is an implicit hierarchical order of regimes of engagement based on their level of generality (i.e. what creates commonality between stakeholders) and their level of publicity (i.e. quantitative participation rates of stakeholders) (Eranti 2018). Participation of stakeholders then ranges from proximate in familiar engagement (not beyond closely affected stakeholders) to preference-based strategic engagement (opt-in participation by relevant stakeholders) to finally, justifiable engagement, as the most general and publicly open format (Brandl et al. 2014;Eranti 2018). From this perspective, justifiable engagement has the most potential for inclusiveness, as there is no space for the great majority of details, personal attachments and pure self-interest (Eranti 2018).
However, this simple hierarchical order does not capture the inclusive and exclusive tendencies present in all forms of engagement (Thévenot 2011). For example, studies have shown that actors relying on familiar modes of engagement find it difficult to express their voice in strategic or justifiable engagement formats, as they lack the required means of communication. This has been noted to be the case with smallholder farmers in global MSI settings, where emotional attachments to specific environments and lived experiences seemingly did not resonate with the engagement format of the MSI (Cheyns 2011(Cheyns , 2012. Thus, it is rather the "structural tyranny of one regime of engagement relative to another" that seems to be most problematic, as it leads to exclusion of specific participation formats, principles of justice or forms of knowledge drawn upon by different actors (Thévenot 2011, p. 53). By shaping how actors interact with one another, the forms of engagement can turn into mechanisms of domination. As actors' capacity in a particular context is linked to a specific form of engagement, power is exercised by enforcing a specific cognitive format and by suppressing other formats (Thévenot 2014). Power then turns into acts of inclusion and exclusion based on prevailing evaluative formats captured by individuals or social classes (Thévenot 2014). Evaluative formats are also not equally powerful: Engagement based on realising plans Justification: any concern has to be formatted in relation to a characterisation of the common good Engagement leads to compromising between a plurality of principles of justice and objectives tends to suppress more cooperative modes of coordination and personal familiarisation (Thévenot 2014). This is because, as Centemeri (2015) argues, engagement based on quantified expression of value (price, efficiency) has progressively reduced the space for valuations based on the common good and familiarity.

Case study selection
The paper focuses on a single case study (Trustea), to allow for an in-depth exploration and analysis of the processes of inclusion and exclusion in a Southern multi-stakeholder setting (Yin 2009). We based the selection of Trustea on the following considerations. The case had to be a Southern sustainability initiative with a general ambition of inclusiveness by means of involving different stakeholder groups. Trustea satisfies these requirements: It was launched in 2013 to develop and implement a sustainability code for the domestic Indian tea market through a multi-stakeholder partnership between Tata Global Beverages, Hindustan Unilever, the Dutch Sustainable Trade Initiative IDH, Rainforest Alliance, Solidaridad, the Ethical Tea Partnership, and the Tea Board of India (a state agency regulating the Indian tea industry). Producer organisations and smallholder representatives are also included in the multi-stakeholder programme committee. Moreover, Trustea is an interesting case of a Southern MSI that aims at the market for domestic tea consumption in India rather than following South-to-North value chains (Langford 2019). This contrasts with the experiences of many global MSIs and their certification schemes, which often target commodities for international consumption, while neglecting domestic consumption in producing countries. This is particularly relevant for the tea industry, where China and India, the main producing countries, also have the biggest rate of domestic consumption. Additionally, Trustea is a sustainability initiative that targets small tea growers in India, who had previously been neglected by international certifications operating in the country (Vermeulen and Dengerink 2016). This carries the potential for addressing the exclusion generated by Northern-based sustainability initiatives.

Data collection
Qualitative methods formed the basis of data collection. First, a preliminary analysis of the case study was conducted through web-based desk research, including an analysis of Trustea documents (reports, presentations, guidelines, newsletters, code of conduct), reports by participating organisations (websites, annual reports, news items) and media coverage (news items). This set the context for the second step: field research in the form of semi-structured interviews with key informants.
Interviewees were selected based on purposive sampling, targeting (1) key stakeholders involved in the development and implementation of Trustea (direct participants) and (2) certified producers and producer associations. The first category of interviewees was approached to deepen our knowledge on Trustea and the involvement of different actors, from initial development to the current management of Trustea. Interviewed organisations include the implementing partners of Trustea (Solidaridad and the Ethical Tea Partnership), the technical partner (Rainforest Alliance), and one of the founding partners (IDH). None of the commercial partners, namely HUL, Tata Global Beverages and Gujarati Tea Processors and Packers agreed to an interview due to corporate communication policies.
The second category of interviewees was selected to understand the perspective of producers' and producer representatives on Trustea, how they are able to participate and how they perceive their role in Trustea. One of the authors of this paper contacted several certified tea companies based in Kolkata, out of which five companies agreed to an interview. The choice of producers took into account both the geographical location of the offices to allow a face-to-face interview and the direct link with Trustea. Additionally, the two main tea producers' associations of India and representatives of producers within Trustea have participated in the research.
In total, 20 interviews were conducted and transcribed, including 16 interviews in the area of Kolkata, India, during June 2017 (see Table 2). All interview transcripts were anonymised and given codes to allow for direct quotes to be used in the paper.

Data analysis
We analysed the data collected through desk research and field research through thematic analysis. This is a particularly helpful method to identify and analyse patterns of meaning within (textual) data (Braun and Clarke 2006). We first engaged in open coding of the data collected which helped us to identify first-order concepts related to the dynamics of inclusion or exclusion of stakeholders in Trustea. During this inductive phase, we listed these concepts by actor type to compare, categorise and identify differences and commonalities across interviewees and, where possible, sources of archival data.
In the second step, we collated the codes into potential themes through axial coding, looking for relations between the first-order codes that we had identified. In this phase, we approached the data using our guiding questions to further operationalise the inclusiveness framework and categorise first-order codes under 'representation of actors', 'principles of justice' and 'forms of knowledge', which we then ordered chronologically along the key phases of Trustea's development and early implementation.

Research background
The production of tea in India dates back to the British colonisers who introduced the cultivation of tea in 1833 and established the plantation system which still exists at present times (Mishra et al. 2011). Tea is a labour-intensive crop and employs more than one million workers, the majority of which are female (Makita 2012). According to Indian law, tea workers and their dependents should be provided with housing and other facilities on the estate. Plantations supply about 50% of Indian tea production. The other half is produced by small-scale growers, which typically sell their tea to estates or bought-leaf factories (Seetharaman 2018). The tea market is dominated by four leading branded manufacturers (Hindustan Unilever, Van Rees, James Finlay and Tata Global Beverages) (Larsen 2016).
India is the second largest country in terms of tea production and consumption, but its tea industry faces several sustainability challenges (van der Wal 2008). Reports of labour rights' violations, child labour and gender discrimination are widespread, in addition to environmental problems associated with mono-plantations, such as deforestation, soil degradation and overuse of agrochemicals (Mishra et al. 2011;van der Wal 2008). Yields have severely decreased in recent years due to changing rain patterns, which has increased the cost of production and decreased the profitability of tea production (van der Wal 2008).
International attention to these challenges increased in the mid-2000s, when a number of initiatives were started  (2007). International certification schemes also started looking at India's tea sector and by 2008, four schemes were certifying Indian tea producers against their standards: organic, Fairtrade, Rainforest Alliance and UTZ certified (the latter two were initially separate, but merged in 2018) (Vermeulen and Dengerink 2016). All schemes target the export market, which has been criticised to lead to a narrow focus on already better-performing farms and leaving out the large domestic market, which absorbs 80% of national tea production (Vermeulen and Dengerink 2016). Therefore, Trustea explicitly focuses on the domestic tea market, having developed a sustainability code in order to verify the social, economic and environmental performance of Indian tea estates, smallholders and bought-leaf factories. By October 2019, Trustea had verified almost half (48%) of the total tea produced in the country annually, including 630 estates and bought-leaf factories with more than 557,000 workers, and 51,000 smallholder growers (Trustea 2019).

Emergence of Trustea: a Dutch-inspired initiative
The initial trigger for the emergence of Trustea can be found in the campaigns by Northern NGOs and civil society organisations in the mid-2000s, aiming to expose the adverse social impacts of tea production in India (e.g. child and forced labour, gender inequality). While the entry into tea production of the aforementioned sustainability standards, such as Rainforest Alliance and Fairtrade, can be seen as a direct response to increased public scrutiny, working and living conditions on the large majority of tea estates remained precarious and damaged the image of the Indian tea industry (Interviewees I001, I005).
Hindustan Unilever (HUL) was the first main tea buyer in India who decided to act on this. In 2010, the company declared that it would buy 100% of its tea from sustainable sources by the year 2020 (Interviewees I001, I002, I008, I009). At that time, HUL was using the Unilever Standard for Sustainable Agriculture and had established linkages with Rainforest Alliance to discuss sourcing of certified tea in India. However, it quickly became clear that Rainforest Alliance would not able to certify sufficient estates to cover the supply needed by HUL (Interviewees I007, I008, I009 and I012). Four constraints were identified which impeded quick upscaling of the Rainforest Alliance standard in India. Firstly, Rainforest Alliance's definition of child labour followed the guidelines of the International Labour Organization for developed countries and set the minimum age for workers at 15 years. In India, however, the minimum legal age is 14, which meant that many estates could not be certified. Secondly, the Rainforest Alliance standard banned a specific weed control chemical widely utilised in India and few less toxic alternatives were available to growers at that time. Thirdly, producers could only receive Rainforest Alliance certification if they had not cut forests since 2005 rather than banning deforestation from the day of certification. Finally, Rainforest Alliance had so far focused on certifying large tea estates, but had hardly any certification experiences in the increasingly important segment of small tea growers (Langford 2019).
HUL initially proposed to adapt the Rainforest Alliance standard on these points, but as the NGO refused, HUL decided to develop a new sustainability standard for Indian tea-one which was more suited to Indian conditions and could reach more tea estates, not only those producing for the export market (Interviewees I001, I002, I005, I007, I009). "These international standards were developed based on different logics of different realities, not the realities of the [Indian] domestic market. They are not going to be very effective in a market like India" (Interviewee I004).
HUL discussed this matter with the Dutch Sustainable Trade Initiative (IDH), with whom they had worked together before on a sustainable sourcing programme. The two organisations chose to involve Solidaridad, a Dutch NGO, for its technical expertise and experience in sustainable agriculture (Interviewees I005, I007). Solidaridad had already created a domestic tea standard for Indonesia (Lestari Code) and contributed to writing the section related to tea in the UTZ standard (Interviewees I005, I007). It had also worked on a local sustainability scheme for Southern India in collaboration with Unilever ten years earlier (Interviewee I012). Being the founding and funding partners of Trustea, IDH and HUL were in the position to choose the actors to involve, which they did based on strategic considerations, initially with the main purpose of developing a code of conduct for sustainable tea production (Interviewees I007, I008, I012). In addition to Solidaridad Asia, they invited the Rainforest Alliance to draw on its complementary technical expertise, experience with certification and its international credibility (Interviewee I009). According to one of the creators of the Trustea code (Interviewee I012), these actors were "ahead of time" in discussing sustainability issues in the Indian tea industry. One year later, also Tata Global Beverage and the UK-based Ethical Tea Partnership joined the initiative.

Code development: an expert-driven process
The development of the Trustea code started before the programme's official launch and was initially driven by two individuals with recognised expertise: a sustainability expert of HUL, with a professional background in tea plantations, and the head of Solidaridad Asia. "Initially, we sat down in a room and created the first draft… then we consulted with stakeholders" (Interviewee I005). The knowledge employed to develop the standard drew largely from the technical expertise of these individuals, complemented with input by tea experts from the Rainforest Alliance and IDH. "This initial team, which created Trustea, was composed by seniors well-known in the industry. They were the best people to give back to the industry, exploiting their knowledge and experience" (Interviewee I002).
Code development started from the existing Unilever Sustainable Agriculture Standard and the Rainforest Alliance standard. This provided legitimacy, according to an individual involved (Interviewee I009). Together, the initiating members of Trustea designed a code which promoted sustainability in a way deemed suitable and feasible for the majority of tea estates and the smallholder producer segment. "The Trustea code is unique for different reasons: it is only applicable to tea; it includes safety standards besides sustainability indicators; it includes smallholders; […] and it pays attention to local realities" (Interviewee I001).
In line with international guidelines for standard creation, field tests were conducted to assess the feasibility of the code on the ground, and public consultations on the draft code were organised with different stakeholders (Interviewees I008, I009, I016). Specific attention was paid to ensuring alignment with the extensive set of government regulations in the Indian tea industry. Therefore, Trustea held several meetings with governmental bodies to get 'official' buy-in and support for an Indian sustainability agenda in tea (Interviewees I005, I011). This also shows in the code that was eventually agreed upon: nearly all Trustea code chapters refer to different areas of Indian legislation, such as the Food Safety and Standard Act and key regulations of the Plantation Labour Act. This was necessary to make the code suitable to Indian conditions, while acknowledging that the code would not be useful to support tea exports (Interviewee I003).
Despite the emphasis on local ownership of the code, tea producers were not involved in the code development phase (Interviewees I002, I013, I017, I08). The founding organisations justified this by referring to the lack of experience of producers with sustainability and growing resistance to existing certification schemes at that time. "None of the producers were part of the creation of the code. There were only experts involved. It would have been biased otherwise" (Interviewee I017). Since all experts that were part of code development had a professional history in tea production, it was assumed that they would be able to include the perspective of producers in the code. This was necessary, as "many producers could not think about sustainability in a structured way at that time" (Interviewee I018). Stakeholders therefore emphasised the importance of having experts from the tea industry driving the code development process, which brought in experiences with tea production and provided credibility and good reputation (Interviewees I001, I002).

Formalisation of Trustea: multi-stakeholder governance
Once the code was finalised, Trustea was formally launched in 2013 as a partnership between IDH, Solidaridad, Rainforest Alliance, HUL and Tata Global Beverages. Solidaridad was appointed as implementing partner, together with the Ethical Tea Partnership. The Tea Board of India was explicitly invited to join as the regulatory body of the industry. According to interviews, the involvement of a governmental player was heavily debated among the funding partners who were strongly committed to establishing Trustea as a voluntary and private standard-setting scheme (Interviewee I008). At the same time, governmental support was deemed critical for local buy-in (Interviewees I009, I011).
Trustea's governance structure consists of three main bodies. The first one is the Funders' Steering Group, comprised of HUL, Tata and IDH, which is responsible for strategic and funding decisions. The Programme Committee, officially the main decision-making body of Trustea, consists of the founding members mentioned above, plus the Tea Board of India, Gujarat Tea Processors and Packers Ltd, the Indian Tea Association, United Planters Association of Southern India and the National Federation of Small Tea Growers. Furthermore, the Advisory Committee convenes a large group of industry actors, such as brokers and traders, research associations and producers' representatives, but does not have decision-making power.
According to stakeholders interviewed, this governance structure represents the entire Indian tea industry (Interviewees I001; I009). The only relevant group not included are labour unions, despite an initial consultation. NGOs involved believed that the unions would have brought an important contribution to Trustea; however, Trustea's funders did not consider them strong and unified enough to be able to participate constructively (Interviewee I007). "We did consult them unofficially and they were supportive of it, but somehow we could not bring them on board […] There was a bit of a feeling that it would have become a kind of, you know, it may not go in the direction which we were looking at, considering that the Indian movement in the tea industry is not exactly what it used to be" (Interviewee I005). Moreover, involving all the workers' representatives was thought to overcrowd the Programme Committee. Therefore, the founding members of Trustea agreed to proceed without direct representation of tea workers in the programme.
Tea producers, by contrast, are directly represented in Trustea, especially in the Programme Committee which includes two associations for larger producers and one dedicated to small-scale growers. These organisations were invited to Trustea's meetings as soon as code development was concluded, according to an interviewee (I001). "The producers are all there, in one or another committee" (Interviewee I005). While several stakeholders emphasised that this gives everyone a voice in Trustea, interviewees also recognised the challenges of involving producers in the governance structure. On the one hand, the ability of the National Federation of Small Tea Growers to adequately represent the perspectives of small growers was questioned, as many small growers were considered to be unaware of the existence of Trustea (Interviewee I011). On the other hand, one of the NGOs involved also indicated that the inclusion of producers had increased the level of discussion in Trustea -which other Trustea members tried to carefully manage to limit the scope of discussion: "There is usually a lot of discussion at the meetings since a lot of producers have been taken on board. So the day before the meetings there is a closed-door meeting among the main decision actors to discuss the most pressing issues. After this, there is free discussion during the meeting, where everyone's ideas are taken into account. This is done to save time, otherwise the discussions would be very long" (Interviewee I009).
This suggests that the three funding organisations and two implementing organisations are able to steer the discussion toward their own agenda, thereby retaining high control over Trustea. This was considered essential to keep the system functional and create a robust management structure (Interviewees I009; I001). Thus, producer organisations mainly perform an advisory role in Trustea, with low decision-making power in comparison with the funders of the programme and the implementing partners.

Roll-out of Trustea: focus on including small-scale growers
Although producers were not part of the code development process and do not feature prominently in the governance of Trustea, they quickly became the focus of attention as soon as the code was adopted and roll-out across the industry started. This pertained especially to small-scale growers, whose share in national production has increased continuously over the years, from 34% in 2014/15 to 47% in 2017/18, due to stagnating estate production (Seetharaman 2018). This made them a major force in tea production, already at the time of Trustea's launch, but one which had largely been ignored by established export-oriented sustainability standards (Interviewee I005). At the same time, stakeholders were concerned that small-scale growers were mostly unorganised and often disregarded the legal framework that regulates the tea sector in India (Interviewee I011). "The challenge with them is that they usually do not keep records and the lower wages they pay to the people they involve" (Interviewee I012). As such, Trustea also positioned itself as a guardian of Indian tea regulations, with the aim of creating consistency in the industry.
Because Trustea targeted the domestic tea market where offering a price premium was not an option and hence, could not act as incentive (Interviewee I019), Trustea started an extensive producer outreach programme. This involved providing free training in compliance for estates, bought-leaf factories and small-scale growers through Farm Support Centres, located in the main production areas. The field staff attached to the centres also educated small growers in good agricultural practices and basic safety rules, encouraging the formation of self-help groups and cooperatives (Interviewee I004). "There was a lot of training for them offered by Trustea about awareness on chemicals and awareness of the law" (Interviewee I018). To avoid that producers would perceive the standard as being too demanding, Trustea explicitly emphasised the value of producers' experiences and traditional agricultural practices. For example, the conservation of native vegetation as buffer and natural pesticide, has been included in the latest version of the Trustea code, as a result of pilot tests in small plots of land, especially established for small growers. "The Trustea code incorporates traditional agricultural practices. There's a strong sense of local" (Interviewee I001). Furthermore, all important documents and manuals were translated into local languages and pictorial guidelines to facilitate understanding by small-scale growers (Interviewee I003).
In addition, having the Tea Board of India as a partner turned out as a critical advantage in increasing outreach towards producers. Due to its active and enthusiastic chairman, the Tea Board quickly emerged as the 'face' of Trustea and actively promoted its uptake (Interviewee I001). The official acceptance of Trustea by the Tea Board enhanced its credibility and increased the demand for verification by producers (Interviewees I005, I008, I009). "Once the Tea Board recommended that producers in India should be certified, it had the force of the law, the force of the government" (Interviewee I011).
Finally, while producers expressed concern that Trustea verification would increase their costs of production, the market pull exerted by HUL and Tata also became a motivating argument. These two companies buy 20% of the tea produced in India, collectively, and expressed the commitment to buy only Trustea verified tea by 2020. "They said to producers, if you don't do the certification by 2019, we will stop buying from you" (Interviewee I005). While there were still many buyers on the market who would purchase non-verified tea, the commitment of the largest tea buyers in India was considered a pivotal driver establishing acceptance of Trustea among producers.

Regimes of engagement in Trustea
When analysing our findings, different regimes of engagement in Trustea become visible, which have shaped the opportunities for the inclusion of smallholder growers and other marginalised stakeholders, especially plantation workers.
First, elements of justifiable engagement can be identified, particularly in the early stages of Trustea's development when the idea of an Indian code for sustainable tea emerged. Actors justified their participation by referring to the common good of promoting sustainability in Indian tea production, often captured as 'sector transformation'. "The reason for joining is the possibility to transform Indian tea production" (Interviewee I006). This was similarly expressed by different participants. Because the circle of initial actors was so small, engagement could be based on shared visions and objectives for the Indian tea industry. This also reduced the importance of strategic interests, according to interviewees. "[…] everyone just tries to do their best. Differences between organisations don't really matter too much, because there is general agreement on the goals of Trustea" (Interviewee I002). The strong commitment to shared goals seems to somewhat contrast with the findings on global MSIs, where studies indicate that the qualification of the common good was avoided in the face of conflicting interests of stakeholders involved (e.g. Schouten et al. 2012;Marin-Burgos et al. 2015). By contrast, Trustea was launched based on common values shared by participants, carefully delineated to the context of India. This made the common good relatively concrete and relatable to the actors involved.
At the same time, the analysis of Trustea also shows that those stakeholders who were perceived not to share those objectives or not yet, and those who were not able to think of the common good, were excluded. Smallholders, who were considered unable to discuss sustainability in a structured way, were not formally included in the development process of Trustea-although they supplied nearly 50% of Indian tea production and thus represented a major stakeholder group. Similarly, labour unions were viewed as not sufficiently constructive, although they were recognised as important actors in the tea industry. Concerns that were considered as too deeply personal or too emotional were not deemed appropriate for Trustea. This resembles findings from global MSIs (Cheyn 2011;Schouten et al. 2012). As a result, the point of view of smallholders and plantation workers was mainly expressed by the founders of Trustea, performing an informal substitution of representation. Even when Trustea was formalised, smallholder growers only received one seat and unions were once again excluded. This suggests that, in addition to justifiable engagement, strategic engagement is present in Trustea, as organisations were included based on what they were able to bring to the table. International NGOs-Solidaridad, IDH, Rainforest Alliance and the Ethical Tea Partnership-were acknowledged for bringing international experiences, technical knowledge, financial support and credibility to the initiative. The two big buyers HUL and Tata were not only important funders of Trustea, but deemed critical actors in sector transformation due to their significant market power in India. Similar forms of strategic engagement by large companies and NGOs based on functional utility were also found in global MSIs (Cheyns 2011(Cheyns , 2012. In the case of Trustea, while inclusion of local producers and unions was minimal, the Tea Board of India, as a governmental actor, was explicitly involved with a view to its influence in the tea industry and its legitimising function. Several interviewees confirmed that endorsement by the national government is an important prerequisite for legitimacy in the Indian tea industry. This also showed in Trustea's rapid uptake by producers which was largely attributed to the promotion by the government. "[…] the government power through the Tea Board of India, it was a major force in the adoption of the Trustea code" (Interviewee I018). At the same time, Trustea explicitly referred to the self-interest of producers to be able to sell their produce, by publicly reinforcing the sustainability targets of HUL and Tata to only purchase Trustea-verified tea by 2020.
Finally, Trustea exhibits elements of familiar engagement. This is based on the accentuation of Trustea as an Indian initiative-despite the fact that all the NGOs involved were Northern NGOs (Langford 2019). Yet, seeing that their staff members were (mostly) Indian, all actors that have driven Trustea's development share a common national identity. Also the Trustea implementing team is composed of local actors who are well-known in the tea industry, with several years of experience in the field. Their involvement hinges on their extensive knowledge of local conditions, attachment to the environment and personal connections with the industry's key players. This also explains why Trustea was able to combine technical expertise with local agricultural knowledge and practices to create proximity to producers' realities. Engagement with producers, especially small-scale growers, was explicitly sought through an outreach programme in the form of Field Support Centres. This is different from the findings for global MSIs, where technical knowledge from experts seems to prevail, with limited knowledge of, and only indirect links with, local working and living conditions of smallholders and communities (Ponte and Cheyns 2013;Offermans and Glasbergen 2015).
Furthermore, Trustea emphasised that it does not introduce new regulations for the Indian tea industry, but enforces and facilitates compliance with the existing-familiar-legal 1 3 system. The prominence of local rules is supposed to enhance the acceptability of Trustea among producers (Langford 2019). This aligns with prior studies that note how Southern sustainability initiatives explicitly aim to position themselves as 'home-grown' initiatives (Schouten and Bitzer 2015; Sun and van der Ven 2020).

Co-existence of different forms of engagement and implications for inclusiveness
The co-existence of different regimes of engagement positions Trustea as a rather different type of MSI, compared to global MSIs, where research indicates the dominance of strategic engagement based on self-interest (Ponte and Cheyns 2013). This is not necessarily because actors in global MSIs are unable to go beyond self-interest but rather because settings are limited in their capacity to promote diverse types of engagement (Schouten et al. 2012). Cheyns (2014, p. 442) noted that when "the majority of the international community have neither put down any roots nor have any attachments" to the sites of production, global MSIs face difficulties in promoting 'care' and finding commonality among actors. Where there is no placed-based connection between actors, familiar engagement cannot arise. Justifiable forms of engagement are also challenging in large-scale international settings. Conflicts among actors as to what the common goal of an MSI should be fuel conflicts, restrict deliberative capacity and lead to exclusion or self-exclusion of actors (Elgert 2012;Schouten et al. 2012). The local character of Trustea and the involvement of a limited number of actors thus seems to give rise to a more diverse engagement regime than what has been observed for many global MSIs.
While reaching small tea growers is one of Trustea's main objectives, this has not promoted high levels of smallholder inclusiveness (see Table 3). Justifiable engagement was present during the initial stages of Trustea's development and is supposed to constitute the most publicly open engagement format, but in this case the construction of the common good was explicitly sought through restricted participation. Smallholders were not recognised as sufficiently capable of contributing to this process. Familiar engagement can also be identified in Trustea. Yet, with regard to smallholder growers, this mostly referred to integrating local, placed-based knowledge of tea production in the Trustea Code. This was achieved by resorting to Indian tea experts from international NGOs, rather than increasing smallholders' participation in Trustea's processes.
Instead, inclusiveness in Trustea is mostly understood as producers' adoption of the Trustea code. This makes Trustea-in outcome terms-more inclusive than previous sustainability initiatives in the tea sector which largely focused on tea plantations, but it does not match the procedural nature of inclusiveness employed in this paper.

Implications for inclusiveness of Trustea processes
Justifiable engagement Participation in Trustea, especially in the early phases, was based on the common good of 'sector transformation' to make tea production more sustainable.
Smallholder producers were not deemed ready for structured deliberation on sector transformation and sustainability, and thus excluded from the early phases of Trustea, especially code development. Workers were not yet identified as relevant stakeholders.
Strategic engagement Actors were invited to join based on their capacity to achieve the objectives of Trustea (promoting sustainable tea production in India), including specific knowledge, experiences, financial support and legitimacy.
Smallholder growers were not part of code development, as other stakeholders already brought relevant expertise and capacity to the table Smallholder growers were later included in Trustea's governance structure. Labour unions were not included, as they were not considered strong enough to represent their constituencies.
Familiar engagement Trustea is explicitly positioned as an Indian initiative, referring to Indian regulations and local, placed-based knowledge.
Smallholder knowledge and traditional farming practices were explicitly included in the Trustea Code. Small growers were also targeted by local experts through Field Support Centres. However, this is did not produce higher levels of smallholder participation in Trustea, as tea experts were able to 'represent' smallholder knowledge.
The focus on producer adoption aligns with a strategic regime of engagement, where careful consideration is given to link participation to realising specific goals, thereby also overriding considerations of other modes of engagement. Smallholder growers were recognised as legitimate stakeholders, but they were only included in Trustea's governance structures once their perceived lack of understanding of sustainability could no longer jeopardise Trustea's overall goals. But even here, their opportunities for participation remain limited and key decision-making power lies with the funding and founding partners of the initiative. Workers and union representatives were also identified as relevant stakeholders, but have remained outside of Trustea due to a perceived lack of organising capacity.
As such, the dominance of strategic engagement can be observed for Trustea's processes. While elements of other engagement formats were present, these were not sufficiently impactful for 'enhanced inclusiveness' of Trustea as a Southern MSI.
The case of Trustea also reiterates the importance of the initiating actors of MSIs-in this case, European NGOs (via their Indian offices) and two of India's largest tea buyers with considerable market power. These actors emphasised the urgency of creating a standard for sustainable tea production, which overrode considerations of wider stakeholder participation. As recognised by Renckens and Auld (2019), early choices around the structure and nature of participation in MSIs can turn into forces of structural influence that are difficult to overcome.

Conclusion
This study investigated the inclusiveness paradigm underlying many MSIs by researching the dynamics of inclusion and exclusion of small-scale producers in Trustea, as an example of a Southern MSI. Such initiatives are increasingly emerging in countries of primary production, as a reaction to the spread, and perceived inefficacy, of Northern-driven MSIs.
We first developed an operationalisation of MSI inclusiveness by drawing on the theory of regimes of engagement, focusing on the dimensions of plurality of actors, principles of justice and forms of knowledge. Our findings highlight the complexity of stakeholder engagement in Trustea. Rather than confirming the existence of a single type of engagement, as has been indicated for many global MSIs, we find that all three types of engagement can exist in parallel: Justifiable engagement is expressed in the participating organisations' reference to a shared common goal of Trustea, whereas familiar engagement is linked to the national identity of the individuals involved and the national ownership of the initiative drawing on the involvement of the Tea Board of India. Strategic engagement, in turn, comes to the fore through the strong emphasis on developing an Indian code for sustainable tea production and promoting adoption among tea growers. The importance of reaching this objective largely limited participation to those organisations deemed critical in reaching this objective whilst those considered to potentially jeopardise the shared objective, were not included. This has led to the replication of exclusionary patterns also noted for global MSIs, as smallholder producers and other weak actors have been largely excluded from decision-making processes, despite informal efforts aimed at substitution of representation.
These findings reiterate the difficulty of achieving inclusiveness in MSIs. This is not only manifest in global MSIs, but also present in the investigated example of a Southern sustainability initiative, showing that marginalised actors-such as small-scale growers but also workers-do not become de-marginalised through a shift in geographical scale of MSIs. Instead, familiar and justifiable forms of engagement need to be strengthened to exist in parallel, and not in subordination, to strategic engagement.