Understanding regional value chains through the interaction of public and private governance: Insights from Southern Africa’s apparel sector

Regional value chains (RVCs) and South–South trade are increasingly considered key features of 21st-century globalisation. This article investigates how RVCs are shaped by the interaction of private and public governance. It evaluates how this interaction unfolded in Southern Africa’s apparel RVCs, exploring trade, investment and labour regimes across three levels of analysis: national, regional, and global. The paper draws on trade data, secondary literature, and interviews with suppliers and institutions in Eswatini and Lesotho (the largest exporters to the region), and lead firms in South Africa (the largest regional importer). The findings underline the critical role of public governance in shaping retailers’ and suppliers’ participation in RVCs through: (i) regional ‘trade regimes’ protecting regional exporters from global competitors, and recent shifts in global trade regimes; (ii) national and regional ‘investment regimes’ facilitating investment flows from South Africa to Lesotho and Eswatini, and the more recent shift of US-oriented suppliers towards regional markets; and (iii) ‘labour regimes’, including lower wages, less comprehensive labour legislation and weaker trade unions in Lesotho and Eswatini compared to South Africa. The article concludes by considering the policy implications of the interaction of private and public governance for existing and future RVCs in Sub-Saharan Africa.


INTRODUCTION
Recent studies show that South-South trade now exceeds North-South trade, and that intra-regional commerce accounts for a large share of the global South's improved trade performance (Horner & Nadvi, 2018).This is leading to a growing interest in the role of The online version of this article is available Open Access 'regional value chains' (RVCs) within the academic and policy communities (Barrientos et al., 2016;McKinsey & Co., 2019;Pasquali, 2019).Yet our understanding of the dynamics behind RVCs remains limited.
Historically, research on global value chains (GVCs) has given primacy to the role of lead firms and the governance of their interactions with suppliers (Gereffi et al., 2005).This emphasis is also seen in much of the emerging research on RVCs (Morris et al., 2016;Staritz & Morris 2017;Suder et al., 2015;Rugman et al., 2009).Recently, various scholars from across the GVCs, global production networks and international business fields have argued for a better understanding of the influence of public policy in shaping value chain dynamics (Coe et al., 2008;Horner & Alford, 2019;Van Assche, 2017).There is growing evidence, for example, that industrial, trade and labour policies have significant impacts on the geographies of GVCs (Curran et al., 2019;Whitfield et al., 2020).However, investigating how public and private governance interact to inform lead firms' and suppliers' strategies in RVCs in the global South remains limited (Krishnan, 2018;Pickles et al., 2015).This gap cuts across disciplinary boundaries of international business and economic geography, and addressing it is relevant for policy actors at national, regional and global levels.
In seeking to fill this gap, we first consider how the interaction of public and private governance at the national, regional and global levels might shape firms' participation in RVCs.On public governance, we investigate the interaction across three distinct regimes: 'trade regimes' influenced by regional and global preferential trade agreements; 'investment regimes' under national investment strategies; and 'labour regimes' defined by wage differentials, national labour regulations and the effectiveness of trade unions.The term 'regime' recognises that the rules and incentives that underlie public governance involve a combination of policy regulation, policy implementation and the role of public and private institutional actors that influence this process (Smith et al., 2018).Second, we empirically evaluate this framework through a study of the apparel RVCs in Southern Africa.The Southern African apparel industry, including its participation in GVCs (Gibbon, 2008;Staritz, 2011), and the implications of emerging RVCs for local suppliers have been previously studied (Morris & Staritz, 2014;Morris et al., 2011;Staritz & Morris, 2012, 2017).We build on this earlier work, in order to assess how the interaction between private and public governance helps explain the ways in which apparel RVCs have developed in Southern Africa.Our particular focus is on South Africa as the main regional importer of apparel, and its nearest regional neighbours -Lesotho and Eswatini as the largest apparel exporters to the region.
We note two critical features of apparel RVCs in Southern Africa.First, the dominance of South Africa as the regional end market for apparel.Second, the power and influence of South African retailers, in particular six lead firms, which together account for 90% of the South Africa apparel market and have an increasing retail presence across the subcontinent.In this context, public governance, involving the interaction between trade regimes, investment regimes and labour regimes, played a critical role in shaping both retailers' sourcing decisions and suppliers' participation in South African-led RVCs.This came about through: (i) duty-and quota-free trade under the Southern African Customs Union (SACU) and Southern African Development Community (SADC) rules of origin, and the recent exclusion of Eswatini from the United States' (US) African Growth and Opportunity Act (AGOA); (ii) attractive national investment policies in Eswatini and Lesotho, which favoured the inflow of South African investments and (in Eswatini) the shift of US-oriented suppliers towards regional markets; and (iii) more comprehensive labour legislation and a stronger trade union presence in South Africa, resulting in significantly higher minimum wages compared to Eswatini and Lesotho.
Our paper makes two important contributions.First, existing scholarship on the dynamics of value chains in both the economic geography and international business disciplines has tended to focus on the private governance of North-South networks (Mayer & Phillips, 2017;Alford & Phillips, 2018).Although the role of states and civil society in GVCs has been increasingly acknowledged (Alford et al., 2017;Langford, 2019), we push this crossdisciplinary literature forward by conceptualising not only the dynamic interaction of public and private governance across different geographical levels of analysis, i.e. national, regional, and global, but also the significance of understanding the interaction between trade, investment and labour regimes in explaining how RVCs evolve over time.Second, from an empirical perspective, previous research on apparel RVCs in Southern Africa mostly focused on aspects of trade, lead firm's sourcing strategies, ownership and social embeddedness, and suppliers' upgrading (Staritz 2011;Morris et al., 2011Morris et al., , 2016;;Morris & Staritz, 2012;2014;2017).To the extent that public governance was included, it was largely limited to considerations of trade agreements and rules of origin within intra-SADC and SACU trade (Staritz 2011;Morris & Staritz 2017).Comparatively lower labour costs were also indicated as a determinant of South African investments in Lesotho's apparel (Godfrey 2015;Morris & Staritz, 2017).In this article, we build on this literature to argue that understanding the dynamics of apparel RVCs in Southern Africa requires unpacking the interactions between private and public governance.This includes analysing trade, investment, and labour regimes, which are shaped by global, regional and national policies, as well as examining how these policy regimes are critically interwoven.
Methodologically, we draw on an extensive reading of the literature, analysis of country-level trade data from UN-COMTRADE, primary research over a number of years in Lesotho, and primary fieldwork undertaken in 2019/20 in Eswatini and South Africa.Interviews were conducted with five of the six major South African retailers as well as with a range of South African key informants.The interviews in Lesotho and Eswatini included four with government officials, three with trade unions and employers' associations, and 31 manufacturers (21 in Eswatini and ten in Lesotho).The focus on South Africa, Eswatini and Lesotho for our primary data is warranted by their respective positions as the largest buying and manufacturing countries in the regional apparel value chain.
The article is structured as follows: section two presents the study's analytical framework; section three describes the growth of apparel GVCs and RVCs in Sub-Saharan Africa; section four investigates the role of private governance in explaining the dynamics of RVCs linkages; section five explores the role of public governance, focusing on trade, investment and labour regimes; section six concludes and considers the policy implications of our findings.

CONCEPTUALISING REGIONAL VALUE CHAINS
RVCs connect lead firms and suppliers within a single world region, defined by 'common regulatory regimes (such as the European Union), preferential trading rules for their regional members (like ASEAN), or have a notional regional identity (like Latin America)' (Horner & Nadvi, 2018, 222).The determinants of links between lead firms and suppliers in RVCs in the global South are poorly understood (Horner & Nadvi, 2018;Krishnan, 2018;Morris et al., 2016).This is in spite of recent data pointing to a 'regionalisation trend ' in GVCs (McKinsey & Co., 2019).Below we summarise key arguments on the evolution and governance of RVCs.These are grouped into two categories: (i) private governance, where lead firms are the main orchestrator of RVCs; and (ii) public governance, where state policy and interventions by public authorities influence the formation and organisation of RVCs.

Private Governance
To understand the role of private firms in establishing and governing their interactions across RVCs, we distil three sets of arguments.
First is what we call RVCs as spillovers from GVCs.Barrientos et al. (2016a) and Krishnan (2018) have associated the growth of regional horticultural value chains in Eastern and Southern Africa with a process of 'strategic diversification' from global towards regional production networks.In some cases, lead firms in RVCs have also replicated the governance structure of GVCs -i.e.'learning by imitation' -including the use of suppliers' exclusivity contracts and private standards (Das Nair et al., 2018;Reardon et al., 2007, 251).Concerning the apparel sector in Southern Africa, Morris et al. (2016) argued that a number of suppliers previously exporting overseas successfully shifted to serving the regional South African market.Yet, they maintain that RVCs remained separate from GVCs, with the former characterized by comparatively smaller batches with a higher fashion content, and speedto-market production (Morris et al., 2016;Staritz & Morris, 2017).
Second, RVCs benefiting from firm-and countryspecific advantages.Studying the determinants of RVCs in East Asia, Suder et al. (2015) conclude that the proximity of countries at different stages of economic development, featuring different labour costs and capabilities, is likely to facilitate intraregional production networks between lead firms in more advanced economies and suppliers in comparatively lower income countries.In relation to Japan first and China later, this dynamic has been observed to produce advantages for lead firms as they benefit from lower labour costs, and for suppliers in less-developed countries as they integrate into their neighbours' production networks, generating revenue and employment (Suder et al., 2015).This pattern reflects the renowned Asian flying geese paradigm, whereby production continuously shifts from more to the less advanced nations in response to changes in national comparative advantages (Kojima, 2000).In the context of the regional apparel value chain in Southern Africa, Staritz (2011, 85) also argued that territorial proximity and lower labour costs in Lesotho and Eswatini played a role in motivating South African manufacturers to relocate production to these neighbouring countries.
Third, RVCs arising from the embeddedness of domestic and regional investors (Morris et al., 2016;Staritz & Morris, 2012).Relative to the apparel sector in Sub-Saharan Africa (SSA), Morris et al. (2016) show that variations in firms' ownership are critical to explain participation and upgrading across global and RVCs: 'Ownership characteristics of supplier firms shape the ability to shift between different end markets and respond to lead firm requirements' (Morris et al., 2016(Morris et al., , 1244)).According to these scholars, the evolution of apparel RVCs in Southern Africa was enabled by local (in Mauritius) and regional (in Lesotho and Eswatini) investors.In addition, East Asian-owned firms located in Southern Africa supplied buyers in the global North.For such transnational firms, ties with buyers and strategic decisions were finalised at their East Asian headquarters and not by the Southern African manufacturing plant.In Lesotho and Eswatini, for instance, transnational firms exporting almost entirely to the US were mostly owned by Taiwanese investors, while regional exporting firms were owned by South African and the few Asian investors who had developed 'stronger ties to South African buyers [did so] based on proximity and ability to meet smaller order sizes' (Morris et al., 2016(Morris et al., , 1255)).Similarly, in Madagascar, geographical proximity and cultural ties with Mauritius played a crucial role in the formation of regional production networks connecting Mauritian apparel producers to Malagasy suppliers (Morris & Staritz, 2014;Morris et al., 2016).
These three arguments are not mutually exclusive but highlight different approaches to explaining private determinants of RVCs.In our view, a focus on private governance is insufficient to understand the evolution of apparel RVCs in Southern Africa.We also need to grasp how public governance influences firms' behaviour.

Public Governance
The broad scholarship on value chains is underpinned by 'persistent firm-centrism' (Alford & Phillips, 2018).Consequently, the term governance has historically 'retained a strongly firm-centric character', with lead firms exercising governance over suppliers through varying levels of power (Mayer & Phillips, 2017, 135).The international business literature also overwhelmingly focuses on the private governance of inter-firm networks and knowledge flows between buyers and suppliers (Buckley & Strange, 2015;Kano et al., 2020).More recently, scholars across both traditions have begun to explore the role of public governance in shaping labour conditions and innovation in GVCs (Locke, 2013;Van Assche, 2017).For instance, in the context of GVCs, scholars have explored the role of states and civil society organisations in determining product and labour standards among suppliers in the global South (Alford et al., 2017;Langford, 2019;Horner & Alford, 2019).Yet, empirical literature analysing the interaction between private and public governance of RVCs remains 'surprisingly very limited' (Slany, 2019, 327).In the following paragraphs, we summarise arguments that relate to the role of investment, trade, and labour regimes in shaping trade in RVCs and GVCs.
First, state regulations can be used to enable domestic firms to link rapidly, efficiently and reliably with foreign value chain partners (Van Assche, 2017).In particular, investment regimes are critical to attract foreign investors, facilitating learning and 'filling the gap' in countries' exporting sectors (Blyde et al., 2014).State-developed export processing zones (EPZs) with preferential taxation and rebates on fabric imports have crucially contributed to attracting apparel foreign direct investments (FDIs) in Mauritius, Madagascar, Ethiopia and Kenya (Farole, 2011;Morris & Staritz, 2014).Previous research on Asian FDIs in Lesotho's apparel industry has also highlighted the importance of financial incentives to attract overseas' investors (Bennet, 2006).Second, regional and global trade regimes can significantly influence lead firms' geographies of sourcing (Curran et al., 2019).Using intra-regional trade in value-added as an indicator of participation in RVCs, Slany (2019) finds that trade liberalization with a reduction of import tariffs on capital goods favoured countries' participation in RVCs across SSA.In the apparel sector, multilateral trade agreements granting preferential access to the EU and the US markets have been crucial in developing apparel GVCs across Africa, Asia, and South America (Bair, 2017;Pickles et. al., 2015;Kaplinsky & Morris, 2008).In SSA, duty-free trade under the SADC Trade Protocol and SACU facilitated intraregional apparel trade (Staritz, 2011, 85-87).Conversely, recent research by Chang and Lohre (2016) in Ethiopia and Rwanda shows how protectionist trade policies introduced by national governments can screen local manufacturers from global competition, effectively enabling firms' participation in global export markets.Similarly, according to Staritz and Whitfield (2017, 3;Whitfield et al., 2020), what sets Ethiopia's apparel sector apart is the parallel presence of a protected national market along with liberalised global exports 'as firms make profits in the protected domestic market while they are learning how to meet the cost, quality and delivery standards of export market.'Third, comparative labour regimes matter.National labour legislation falls within the realm of public governance.Such legislation can set minimum wages and standards as well as regulate health and safety, but it also enables organisation and collective bargaining by trade unions.Labour rights and agency in GVCs have been a growing focus of research.However, scholars have adopted different approaches to labour, ranging from the issue of labour control regimes (Smith et al., 2018) and the labour process (Taylor et al., 2013), to worker agency and resistance (Coe & Jordhus-Lier, 2011;Selwyn, 2012), to labour standards linked to social upgrading (Barrientos et al., 2011).Moreover, and intersecting with aspects of private governance, the creation and enforcement of labour codes by NGOs such as Clean Clothes Campaign as well as multi-stakeholder initiatives, such as the Ethical Trading Initiative or the International Labour Organisation's (ILO) Better Work Programme, bring together government, trade unions, suppliers and buyers (Bair, 2017).To our knowledge, however, few studies have had a regional focus with respect to labour regimes in the global South (Godfrey, 2015); and none have considered their interaction with other aspects of public and private governance in RVCs.

Private and Public: A Dynamic Interaction Approach
As they outsource production tasks, lead firms face a variety of risks connected to product quality, management, labour, and the environment.According to the economic geography and international business literatures their responses to these pressures varies depending on context-specific economic, political, social, and geographical factors (Yeung & Coe, 2014;Kano et al., 2020).Scholars have used concepts such as 'multi-scalar', 'hybrid', and 'synergistic' governance to indicate how suppliers' participation in GVCs is inevitably influenced by a confluence of lead firms' strategies, civil society initiatives, and public policies (Alford et al., 2017;Bair, 2017;Gereffi & Lee, 2016).In this domain, Pietrobelli and Staritz (2018, 570) advocate a 'multi-scalar framework that goes beyond traditional approaches that either focus on the nation state or the firm'.A key contribution, they argue, is to unravel the interactions between private and public actors while 'integrating the global with the local, and the firm (micro) with the meso and macro levels.' We support such an integrative endeavour and draw on Gereffi and Lee's (2016) taxonomy for 'synergistic' governance, which we adapt by replacing their categories of horizontal and vertical governance with multi-scalar geographical dimensions.Our adaptation accounts for an increasing overlap between GVCs and RVCs in a context of shifting end markets (Horner & Nadvi, 2018), requiring consideration of private and public governance interactions at the national, regional and global levels.While Gereffi & Lee's (2016) concept is specific to the relationship of clusters with GVCs and models six potential upgrading trajectories, our adaptation is used primarily to organise and analyse empirical data with the focus on RVCs.Notably, Gereffi & Lee (2016) use 'social governance' to indicate the action of civil society organisations (including NGOs and trade unions).However, trade unions are generally established and act within frameworks prescribed in countries' labour legislation, which effectively blurs the distinction between public and social governance.Hence, our preference for the use of the term 'labour regimes' as this more accurately captures how rules in respect of labour are framed and implemented, and the impact that labour organisations and trade unions play in shaping the resulting practices (Smith et. al., 2018).
In sum, we recognise that lead firms exercise private governance to coordinate value chain linkages, including compliance by suppliers with codes of conduct.At the same time, we note that public governance, through national laws and inter-governmental bilateral, regional and global regulations and trade agreements influences these dynamics, and we explore how public policies shape trade, investment and labour regimes.Hence, we address three major concerns related to economic geography and international business research: the firmcentric approach that leaves public institutions 'lurking in the background' (Locke, 2013;Alford & Phillips, 2018, p. 99); the limited attention to labour regimes and regulations in shaping RVCs in the global South (Godfrey, 2015;Barrientos et al., 2016); and the overwhelming focus on North-South GVCs, overlooking local and regional actors in the global South (Buckley & Strange, 2015, 246;Horner & Nadvi, 2018).

APPAREL RVCS IN SUB-SAHARAN AFRICA:
A SOUTH AFRICAN STORY Sub-Saharan Africa (SSA) is one of the least trade integrated regions worldwide.Whilst in Asia, Europe, and America intra-regional trade accounts for respectively 61.1%, 67% and 47.4% of these regions' total trade share, in Africa this figure drops to 15.2% (UNCTAD, 2019).Nevertheless, the apparel sector represents a notable exception.Since 2005, intra-SSA trade in apparel has grown more than fourfold to account for about 34% of the region's total apparel exports.Comparatively, exports to the developed North have decreased by 30% over the same period, shifting from 92% to 66% of the sector's export share (UN-COMTRADE).
Excluding Ethiopia (Staritz & Whitfield, 2017), today's SSA apparel exporters emerged in the 1990s and early 2000s, initially as a result of quotahopping FDIs arising from the Multi-Fibre Arrangement, and subsequently benefitting from duty-free access to the United States (US) and European Union (EU) markets under the US' AGOA and EU's Everything-But-Arms trade regimes (Gibbon, 2008;Staritz, 2011).Whilst the dynamics of the sector's emergence vary across different countries, Lesotho and Eswatini both benefitted from an early inflow of Taiwanese FDIs organised through triangular networks linking head offices in Taiwan, where orders are processed, to retailers in the US and production units in Southern Africa.As argued by a number of studies that address this period in more detail, these were essentially 'footloose' investments with very limited local and regional linkages (Staritz, 2011;Staritz and Morris, 2017).Some two decades later, although the global North remains an important market for most of these countries, a significant regionalisation trend has emerged (Staritz & Morris, 2012).In line with previous findings (Morris et al., 2016), export data indicates a growth in regional exports for Eswatini, Lesotho, Mauritius, and Madagascar (Fig. 1).Moreover, each of these four countries experienced a diversification of end markets from the North to the region (Fig. 2).Between 2010 and 2018, Mauritius increased its regional export share from 10% to 22% of the total, Madagascar from 5% to 17%, Lesotho from 13% to 32%, and Eswatini from 39% to 98%.This dynamic has not gone unnoticed.Staritz (2011, p. 57) argues that following the 2008 financial crisis: 'regional end markets, in particular South Africa in the case of SSA, also may become central to substitute for reduced exports to developed countries' markets.'More recently, Morris et al. (2016Morris et al. ( , 1245) ) indicate that: 'regionalizing exports […] are important for Lesotho, Swaziland, Madagascar and Mauritius, but not for Kenya.'But the extent of this regionalisation trend has been underestimated, and the role of South African retailers in conjunction with that of public institutions, and specifically the interaction between trade, investment and labour regimes, has not been adequately investigated.
Import data confirms the centrality of South Africa to regional apparel trade: the only country whose regional apparel imports grew is South Africa (Fig. 3).Whereas Staritz (2011) noted that regional clothing imports by South Africa 'remained relatively marginal, reaching only 9%', by 2018 regional sourcing constituted about 31.5% of South Africa's total apparel imports.Of this, 33.3% originated from Eswatini, 23.9% from Lesotho, 23.2% from Mauritius, and 16.5% from Madagascar.Conversely, in the same year over 95% of regional clothing imports by Namibia, Botswana and Zambia originated from South Africa, mostly in the form of re-exports.Furthermore, South Africa's apparel market generated the second highest revenue in SSA (Figure 4) and was second in terms of per-capita expenditure on apparel (after Mauritius), about five times larger than the SSA average.
In what follows, we concentrate our analysis on Eswatini and Lesotho, and their linkages with the South African end market.Importantly, we are not comparing between Lesotho and Eswatini and their ties with South African retailers, but rather considering what factors help explain the emergence of Lesotho and Eswatini collectively as the leading exporters of apparel to the regional market.

DRIVERS OF APPAREL RVCS IN SOUTHERN
AFRICA: PRIVATE GOVERNANCE Historically, South Africa's apparel sector has been dominated by six national retailers (Top-6), whose sales in 2015 accounted for over 90% of the nation's total (Fig. 5). 1 Their sourcing strategy up to the early 2000's, was largely restricted to a number of large full-package local manufacturers (Staritz, 2011).Over the last two decades, however, two factors significantly reshaped the Top-6 regional sourcing and retailing strategies: (i) increasing access to large volumes of cheaper garments made in China and South Asia; and (ii) the more recent entrance of global apparel retailers like Zara and H&M into the South African market.These developments indicate an increasingly competitive domestic market.We analyse these in turn.
Following South Africa's democratic election of 1994 and its membership of the World Trade Organization (WTO) in 1995, import tariffs and quotas on a number of goods including textile and apparel were significantly reduced (Staritz, 2011).Throughout the late 1990s and early 2000s, this scaled down the cost of garments sourced from overseas, with the result that South African retailers increased their share of global sourcing from Asia sixfold between 2000 and 2005, with concomitant reduced internal production (Fig. 6).In the process, the Top-6 significantly improved their bargaining power vis-a `-vis local manufacturers, leading to a fundamental restructuring of production (Staritz, 2011).The number of South African full-package apparel manufacturers declined, the average size of firms decreased, employment dropped by at about 30% from 2008, 2 and a large proportion of apparel production started shifting to neighbouring countries with lower production costs (Godfrey, 2013).Our data suggest that between 2000 and 2018/19, the number of South African-owned apparel production plants in Eswatini and Lesotho rose from zero to 9 in Eswatini and 18 in Lesotho, employing about 19,000 workers (11,500 in Lesotho and 7500 in Eswatini).
From a retail perspective, South Africa's internal market has become more competitive following the entrance and expansion of large international retailers, including Cotton On and Zara in 2011,

Regional value chains in Southern Africa
Importing from places like Lesotho and Eswatini is an easier and cheaper option…' (Cape Town, 15/ 04/2019), and the divisional executive of a Top-6 argued: 'We are facing strong competition from new retailers and cheaper imports.To be competitive, we need to broaden our market base and look at Africa not just for sourcing, but increasingly for retailing…' (Johannesburg, 05/07/2019).Another retailer also stressed that, as international brands enter the market, the pressure on cutting costs increases: 'carefully expanding our presence in Africa and globally will allow us to increase scales and revenue.' (Johannesburg, 04 /07/2019) Overall, from a private governance perspective, RVCs emerged via a two-way process: first, the regionalisation of production as suppliers sought to retain orders with retail clients by reducing prices through relocation to neighbouring Lesotho and Eswatini and, second, the regionalisation of retailing as the Top-6 gradually expanded their presence and sales across the sub-continent.However, as we shall see in the next sections, regional sourcing by South African retailers goes well beyond the relocation of South African suppliers, and it includes an increasing number of global suppliers shifting from GVCs to RVCs.

DRIVERS OF APPAREL RVCS IN SOUTHERN
AFRICA: PUBLIC GOVERNANCE In this section, we examine how trade, investment and labour regimes at national, regional and global levels, have influenced the dynamics of apparel RVCs in Southern Africa.These 'regimes' are primarily a function of public policy, but firms and trade unions also play an important role in shaping them.The interaction of these 'regimes' creates a framework of rules that critically enables and incentivises private actors' decision.

Trade Regimes
Southern African apparel value chains are strongly influenced by global and regional trade regimes.At the global level, AGOA has provided most countries in SSA with preferential, duty-and quota-free access to the lucrative US market since 2001. 3For most SSA countries, these preferences come with single transformation rules of origin, which allows them to use fabrics imported from third countries in their apparel exports to the US. 4 AGOA further requires member countries to fulfil a long list of compliance regulations in order to be eligible -including the enforcement of private property rights, the protection of worker rights through the ratification of ILO core conventions, the protection of human rights, the elimination of barriers to US investments, a system to combat corruption, and economic policies to reduce poverty (AGOA, 2020).While the Multi-Fibre Arrangement encouraged the emergence of apparel GVCs in the early 1990s and AGOA provided the incentive to maintain them later on, regional trade agreements have played a critical role in the dynamics of RVCs.SACU is a long-established customs union with free trade between Lesotho, Eswatini, Botswana, Namibia, and South Africa (with no specified rules of origin).Furthermore, the SADC Trade Protocol of 2008 allows duty-and quota-free trade of apparel under 'double transformation' rules of origin, requiring fabrics to originate from within SADC (Staritz, 2011, 95).Unlike AGOA, SACU and SADC do not impose any other major conditionalities.
The critical role of global and regional trade agreements in shaping firms' participation in RVCs is illustrated by the case of Eswatini.Starting as early as 2003, the country faced increasing criticism for failing to comply with AGOA's requirements on workers' freedom of assembly, including the free functioning of trade unions.Eventually, this led to it losing its AGOA status from 2015 (The Economist, 2014).With a rapid and sharp drop in US orders, all transnational suppliers in Eswatini had shifted their end market to South Africa by 2017.This contradicts Morris et al.'s (2016Morris et al.'s ( , 1259) ) finding on the 'disembeddedness' of 'transnational producers focused on the US market'.Instead, managers of well-established Taiwanese suppliers, part of triangular manufacturing networks serving the US market, exerted a degree of autonomy to develop a regional strategy.The manager of a large plant in Eswatini stated: 'At the beginning, my boss in Taiwan was not interested in exporting to South Africa.We tried to relocate to Lesotho, but it did not work […] Now, finally, they are sending more managers from Taiwan to help me cope with the new structure… In the long run, even if we go back to the US market, we want to continue exporting to South Africa… We cannot rely only on one market.'(Matsapha,19/07/19) As noted above, it was the South African buyers that took the initiative to tap into the capacity of Eswatini's large transnational exporters.Of seven Eswatini firms actively exporting to the US in 2011, four were directly contacted by regional buyers: 'We didn't approach the South African side because we are not familiar, but somehow they came to us, and asked us to do woven, which we could.We started with small orders… They helped us a lot…' (Nhlangano, 12/07/2019).A South African buyer stated: 'We knew their market was gone… There are not many factories of that size and capability in South Africa… It was an opportunity.They've got a large contingent of mechanics and supervisors from outside the country.You know, they can go to Taiwan and interview people, I can't, I don't have that skill…' (Matsanjeni, 22/07/2019) The situation in Lesotho is different.The country was never threatened by exclusion from AGOA.Consequently, shifting of foreign-owned transnational suppliers from the US to the South African market was limited compared to Eswatini.Yet, it did occur.In early 2017, out of a total of 28 Taiwanese and Chinese-owned factories employing 31,943 workers, six firms had switched entirely to supplying South Africa (total employment 4369 workers), and another two firms were supplying both the US and South Africa (total employment 1032 workers). 5 Whilst most Southern Africa apparel producers can access both global and regional markets dutyand quota-free, differential tariffs for other global exporters critically shape the value chain.In the absence of a trade preference arrangement, countries import under nationally set 'most favoured nations' (MFN) tariffs.This is especially significant for the South African market, but also for the US where Chinese-made apparel, which often competes with SSA apparel exports, faces MFN tariffs.In 2018, China, Bangladesh, Viet Nam and CAFTA region 6 were among the largest apparel exporters into the US.In the same year, China, Bangladesh, and India were the main non-SADC apparel exporters to South Africa.Table 2 summarises the tariffs these countries face to access the US and SACU markets and further compares them with the tariffs faced by SSA apparel exporters.
All but AGOA and CAFTA countries face up to 32% MFN tariffs when entering the US market.However, for the products that SACU countries export most to the US (i.e.men's trousers, women's trousers, and men's shirts) MFN tariffs range between 0% and 14.9%.Therefore, SADC and SACU countries' advantage over their global competitors ranges from 0% to 14.9% for the US market.This is considerably lower than the advantage they enjoy when exporting to South Africa, where all major non-SACU/SADC competitors face a 40-45% MFN tariff.
To get a better sense of how tariff regimes affect sourcing, we compare unit values for the top two products exported by Lesotho, Eswatini, Madagascar and Mauritius that are common to the US and South African markets.With the exception of Mauritius, unit values are consistently higher for the same products when exported to South Africa than to the US (Fig. 7).On the one hand, the relatively higher unit values characterising regional exports reflect a demand for short-run products with a higher fashion content.Here, it is argued, regional production networks allow for more flexibility and increasing speed-to-market via a rapid turnaround (Morris et al., 2011(Morris et al., , 2016)).On the other hand, the outcome of Fig. 7 is in line with the statements provided by suppliers, which suggests that the South African market is less price competitive.The manager of an Eswatini factory explained: 'To do business with the US, you need to be a huge company.When they place the order, they tell you ''this is Bangladesh and Cambodia's price.What price can you give to me?''If Bangladesh or Ethiopia charges USD 4, we have not only to do to the same, but they say we need to do it at USD 3.5 because we have the duty advantage… The US market is all about big volumes and low prices.South Africa instead [demands] smaller volumes but pays better prices.' (Nhlangano, 13/09/2019) Whilst Eswatini was officially reinstated into AGOA in 2018, as of December 2019 no firm had shifted back to the US market.Crucially, when asked why they are not reverting to the US market following reinstatement, five firms previously engaging in overseas exports mentioned 'competition and lower profit margins' as major factors (Fig. 8).The other main factor was the complex bureaucracy and auditing required to fulfil AGOA requirements.The managing director of a large Taiwanese firm, indicating tension between his own regional preference and his head office, stated: 'I'm not very interested in going back to the US, but I get pressure from my main office in Taiwan.The problem is the price… US buyers are looking for an AGOA country to supply their product, but they're still asking us to compete with the Bangladesh on price, but it's not fair!My main office still has to consider this… Maybe we will do some business with US buyers, if we select the better price customers only…' (Matsapha, 11/07/2019).
The manager of another factory in Eswatini was also sceptical of a potential return to AGOA, especially on stricter conditionalities: 'You need to have documentation for everything.They [US buyers] have more requirements than the South Africans.They want the company to be like a 5-star hotel… They need to check everything, doors, waste, workers… With South Africa we just follow the local laws…' (Nhlangano, 17/07/2019)

Investment Regimes
As discussed in section four, the number of South African-owned apparel production plants in Eswatini and Lesotho rose significantly over the last two decades.Along with trade preferences, a critical motivation for South African FDIs was attractive investment packages.In both Lesotho and Eswatini, investment regimes have had three main pillars: (i) establishing investment promotion agencies (i.e. the Lesotho National Development Corporation (LNDC) and the Eswatini Investment Promotion Agency (EIPA)) to reduce times and costs of setting up factories in the country; (ii) developing a package of tax incentives available to foreign investors; and (iii) granting access to industrial infrastructure, e.g.factory 'shells' in industrial parks with subsidised rentals.
LNDC and EIPA were set up in 1967 and 1988, respectively, to promote foreign investments.Their scope has been to help investors fast-track compliance with tax and labour laws, provide information on standards and trade compliance across export markets, and manage the provision of factory shells.For instance, as of 2019, LNDC and EIPA were allocating ten shells per year over the next 4 years in Eswatini and 11 shells at a new site in Lesotho (Dlamini, 2019).
Tax incentives have also been a major plank in the strategy to attract FDIs.In Lesotho, since 2004, foreign investors benefit from a reduced 10% corporate income tax, no withholding tax on dividends, and tax-free repatriation of profits.In the early 2000s, Eswatini introduced a special tax dispensation for exporters under AGOA: an income tax on profits of only 10% (compared to the normal rate of 27.5%), and a 15% withholding tax on dividends paid to foreign shareholders.In addition, since 2007, both countries enjoy a SACU provision for duty-and VAT-free access to capital equipment and the possibility for firms to operate through 'deferred accounts', which allow them to effectively discount VAT and import duties on fabrics imported from outside SACU.Lesotho and Eswatini introduced double taxation agreements with South Africa in 1997 and 2004, respectively, exempting South African investors from paying taxes in South Africa for income generated in the two SACU countries.
In Eswatini, EIPA crucially facilitated links with South African buyers for three transnational producers and helped them shift to the South Africa market when suspension from AGOA loomed.The facilitation included the extension of fiscal benefits previously restricted to overseas exporters.Since then, all large firms previously exporting to the US have renegotiated their preferential treatment to cover their South Africa exports.The manager of a large Taiwanese-owned plant explained: 'I asked the government to allow me to shift to the local market.I had to ask, because when we came, we had preferential fiscal treatment granted because of AGOA.I asked, why don't you allow me to do SACU production, I can get better prices and workers can get better salaries?So, eventually they agreed.Other factories followed me […] I'm the leader of this idea'.(Nhlangano,13/09/19) The majority of South African investors in Eswatini also report benefitting from the local government's investment dispensation, particularly the lower corporate tax rate, but this is secondary to the attraction of low labour costs.Furthermore, as argued by an official at EIPA, there is no need to assist South African investors with trade because they know the market, but 'lower taxes in a system that allows duty-free exports' are an attraction (Mbabane, 08/07/2019).The manager of the first South African plant to relocate to Eswatini also explained: 'It is very easy to start a business in [Eswatini]… This happens via EIPA.EIPA has improved… It is now a more welcoming and helpful environment' (Johannesburg, 04/07/2019).

Labour Regimes
Regional trade regimes and investment regimes facilitated the growth of apparel RVCs between South Africa and Lesotho and Eswatini.Yet, sharp differences between the labour regimes in these countries critically contributed to this process.
While minimum standards regulation can be quantified and relatively easily compared, labour regimes are more complex and are for the most part the result of enabling provisions.The system that emerges is therefore determined largely by how trade unions and employers use the national legislation.This caveat aside, there is no doubt that South Africa has a much more developed labour regime than those in Lesotho and Eswatini.In South Africa almost all workers are covered by labour legislation, there is a national minimum wage, legislated minimum conditions of employment and some sectoral minimum wages and conditions, freedom of association and organisational rights, a well-established centralised collective bargaining system, protection against unfair dismissal, effective dispute resolution institutions, a health and safety statute, workmen's compensation and unemployment insurance, and an extensive (but poorly functioning) training system.The labour regimes in Lesotho and Eswatini share only some of these features and do not match the levels at which standards are set and rights protected in South Africa.
In the apparel sector, it is collective bargaining that sets South Africa apart from Lesotho and Eswatini.Legislation in South Africa includes a framework for the voluntary establishment of sectoral bargaining structures called bargaining councils and provides a mechanism through which collective agreements reached in the bargaining councils are extended to apply to all employers and employees in the sector.Agreements reached at the National Bargaining Council for the Clothing Industry (NBCCI) currently regulate minimum wages, conditions of employment, and social benefit funds for all workers in the apparel sector.
Lesotho does not have equivalent collective bargaining provisions.Its legislation protects freedom of association but provides no support for organisation by trade unions or a framework for collective bargaining at either firm or sectoral levels (Godfrey, 2013;2015).Eswatini, on the other hand, has a framework for the establishment of joint negotiation councils which is similar to that of South Africa, including a mechanism to extend agreements.However, trade unions in Eswatini have not yet managed to establish such a council for the apparel sector, effectively rendering these provisions irrelevant at this point in time.
The differences between the labour regimes become sharper when one shifts the focus from the 'enabling provisions' to the way in which the apparel trade union in South Africa has used NBCCI to improve sectoral wages and working conditions.The Southern African Clothing and Textile Workers Union (SACTWU), was formed in 1989 out of a series of amalgamations of textile and clothing unions, in a period when progressive trade unions burgeoned in South Africa.The amalgamations left SACTWU well positioned: it was the only trade union in the apparel sector and it had a very high level of representation amongst workers (Maree & Godfrey, 1995, 131-133).From its inception it sought to increase historically low wages and to merge the regional industrial councils into a national council.This was achieved in 2002 with the creation of the NBCCI.In the years that followed, the labour cost differentials narrowed between the major clothing manufacturing centres in South Africa as well as between these centres and rural areas (Godfrey, 2013).
Bargaining by SACTWU to raise wages, while tariffs came down and local manufacturers rapidly lost market share to cheap imports, saw the sector shrink and restructure.The solution for some manufacturers was to relocate to neighbouring countries that had weaker trade unions, limited or no collective bargaining, and much lower labour costs.One such firm explained the rationale to relocate to Lesotho as follows: 'The main emphasis of why we started in Lesotho was purely from a cost point of view, because the labour rates are much cheaper in Lesotho than in South Africa.The second thing is you can work a full 45-h week here.And the third issue is you are not so hamstrung with the unions and all these fancy regulations that they have in South Africa.Those are the three main criteria.'(Maputsoe, 29/06/2011) The South African owner of another firm in Lesotho stated: Our fieldwork in Eswatini found that six of the seven (still active) suppliers that initially relocated from South Africa identified 'union pressures and higher wages' as the primary motive for relocating (Fig. 9).Other reasons for moving include: Eswatini's attractive investment policy (4), AGOA single transformation rules of origin (3), better security (2), and direct support from the government (2).As the manager of a South African production unit that moved to Eswatini explained: 'The main reason why we left [South Africa] is because of the bargaining council and the union… They were hammering us and we just couldn't run our business.Eswatini offered a competitive edge into South Africa…' (Matsapha, 24/07/2019) Wages are the major component of labour costs.While sector-wide wage rates are negotiated annually at the NBCCI, unions in Eswatini and Lesotho are less well organised and collective bargaining is limited.In Lesotho, there are five rival trade unions in the apparel sector, none of which is well organised.In the absence of wage bargaining at the firm level, unions engage in annual multipartite consultations facilitated by the Wages Advisory Board, which then recommends an increase for the sector.The situation is better in Eswatini: the major trade union in the sector, the Amalgamated Trade Union of Swaziland (ATUSWA), is recognised as the representative of their workforce by five of the 21 apparel manufacturers, but as at 2019 only two had concluded a collective agreement with it.ATUSWA, furthermore, does not have sufficient representation in the sector to push for the establishment of a joint negotiation council.In its absence a minimum wage is periodically set by a Wages Council together with the Ministry of Labour and Social Security.
Apparel employers in Eswatini are however less accommodating than those in Lesotho: by informal agreement firms pay exactly at the prescribed minimum wage and have made a concerted attempt to avoid collective bargaining with ATUSWA.The spokesperson for the Eswatini Textile and Apparel Traders Association (ETATA) stated it 'will deal only with trade-related issues.[It] will not engage in collective bargaining with unions or in negotiations over labour regulations' (Matsapha, 11/07/2019).Our interviews also revealed considerable dissatisfaction with one employer who had allegedly 'broken ranks' to negotiate higher wages with the union.
In light of the above, the steep regional wage differentials, which motivated South African manufacturers to relocate, remain, and with their close proximity make Lesotho and Eswatini attractive sourcing options for South African retailers.In August 2019, the weekly minimum wage of a trainee and a qualified machinist in South Africa ranged between ZAR 806-857 and ZAR 909-1164, respectively, in Eswatini and Lesotho they were less than half, ranging from ZAR 284 to ZAR 391 for the former and from ZAR 418 to ZAR 492 for the latter (Fig. 10).If one were to add social security contributions to the minimum wages, the gap is even wider.
Differences with regard to private governance provide a further factor favouring regional supply.While exporters to the US must comply with US retailers' codes of conduct, the same is not true within RVCs (Godfrey et al., 2019).These codes and the auditing that accompanies them is one of the reasons suppliers in Eswatini do not want to revert to the US market.Conversely, with few exceptions, such codes do not exist or are comparatively less demanding in RVCs.As reported by a factory manager in Eswatini: 'US audits were nonsense.They ask us to observe human rights, stop bag searching, reduce overtime… If you export to the US market you have to comply.[…] South African buyers are more understanding.They don't touch these issues because they know what's going on.They are not worried with this as long as you deliver.' (Nhlangano,17/07/19) Not much has therefore changed since Staritz's (2011, p.188) survey found that 'only two [South African] retailers stated that labour compliance is important in their sourcing criteria'.To date, only one of the Top-6 has committed to ethical sourcing by joining the UK-based Ethical Trading Initiative (ETI), thereby binding suppliers to ETI's base code.The other five retailers vary from making only vague statements of intent in their annual reports to taking tentative in-house steps towards greater scrutiny of labour conditions at suppliers.Further, the practice of sourcing via intermediary agenciesknown as design houses -allows South African retailers to screen themselves from association with labour conditions at supplier factories (in 2019, 12 out of 21 plants in Eswatini were contracted via design houses) (Pasquali & Godfrey, 2020).The low concern for labour standards among South African lead firms is further evident in Lesotho, where in 2010 the ILO launched the Better Work Programme to enforce compliance in country's apparel industry.Yet, while most suppliers to the US, under pressure from their buyers, joined the programme, South African retailers would not endorse the programme with the result that almost all South Africa-oriented suppliers refused to join it (Godfrey et. al., 2019).

CONCLUSION
The growth of South-South trade and the increasing importance of RVCs is now widely recognised.What is less well understood is what factors determine the dynamics of RVCs.Traditional international business and economic geography scholarships have historically focused on the centrality of lead firms who exercise private governance to shape global and regional value chain linkages.More recent arguments have underlined the need for a greater understanding of the role of public governance, through various forms of local, national, regional and global policy interventions.
Building on this agenda, our focus has been to understand how particular forms of RVC dynamics observed in the Southern African apparel sector have unfolded.While Sub-Saharan Africa is poorly integrated in terms of intra-regional trade, the apparel sector is an exception.Growth in intraregional trade in apparel in SSA is marked, most pointedly between South Africa, Lesotho and Eswatini.We have sought to explain this development by unpacking interactions between private and public governance.In particular, we focused on how firms' behaviour was affected by trade, investment and labour regimes, which were themselves framed by distinct global, regional and national policies and institutional settings.
The growth of apparel RVCs in Southern Africa was clearly driven by the agency of major South African retailers who expanded their sourcing and retailing presence regionally.Their actions were critically influenced by trade, investment and labour regimes -as summarised in Table 3. Whilst the Multi-Fibre Arrangement and AGOA created important opportunities for SSA countries to access the US market duty-and quota-free, the level of protection within SACU implies that the South African market is better screened from foreign competition than its global alternatives.This facilitates regional sourcing.As argued in previous studies, a number of suppliers with regional linkages have successfully leveraged this trade advantage and engaged in RVCs (Morris et al., 2016;Staritz & Morris, 2012;2017).However, the active pursuit of regional suppliers by South African retailers and the loss of AGOA membership by Eswatini saw increasing shifting towards RVCs also of those 'transnational producers focused on the US market' (Morris et al. 2016(Morris et al. , 1259)).
Investment incentives in Lesotho and Eswatini further facilitated relocation by South African apparel manufacturers and, in the latter case, the shifting of transnational overseas exporters from GVCs to RVCs.In addition, a strong trade union in South Africa, sector-wide collective bargaining, and much higher labour costs were a major catalyst that propelled the foundation for RVCs.The role of differential labour regimes in this regional dynamic unfolded over time, running parallel with declining tariff protection in South Africa.In this context, SACU's long-standing duty-and quota-free trade regime encouraged relocation and regional sourcing.
On the one hand, the centrality of South African retailers and suppliers supports the notion that heterogeneity in firm-and country-specific advantages (i.e.labour costs) favours regional integration through RVCs (Suder et al., 2015;Staritz, 2011).Moreover, territorial proximity and ownership characteristics, including the embeddedness of suppliers in local and regional networks, have also facilitated the formation of inter-firm regional networks, especially via South African retailers and manufacturers (Morris et al., 2016;Staritz and Morris 2017).On the other hand, lead firms' geographies of sourcing and retailing are crucially shaped by public governance.This includes international and regional trade regimes along with their rules of origin and tariff barriers (Curran et al., 2019;Staritz, 2011), as well as national investment policies and labour regimes (Godfrey, 2015).In Eswatini, and to a lesser extent Lesotho, we also observe a spillover from GVCs (Krishnan, 2018), as transnational investments originally serving US buyers shifted to RVCs following changes in global trade regimes and national investment policies.There are however few similarities cutting across GVCs and RVCs.Private governance through audits and codes of conduct is not as pervasive in RVCs as it is in GVCs, a point reinforced by both transnational suppliers in Eswatini who recently shifted to the South African market, and by South Africanlead retailers.
Our findings add to calls for bringing new perspectives to traditional GVC (and emergent RVC) analysis.The traditional notions of governance in value chains need to be further conceptualised.Public governance clearly influences firms' private governance.Moreover, we argue that not only is this context specific, but that distinct forms of policy, framed by different levels of geography (national, regional and global), intersect.It is the interweaving of these that drives specific lead firm practices and results in particular forms of value chain linkages, be they global, regional, or national.We therefore concur with other economic geography and international business scholars in calling for a better understanding of the role of public policy in the study of value chains (Gereffi & Lee, 2016;Mayer & Phillips, 2017;Staritz & Whitfield, 2017;Van Assche, 2017).Further, our study shows that this interaction is not necessarily complementary and thus 'synergistic', therefore leading to firms' upgrading, as Gereffi & Lee (2016) argue.The interaction can be mutually supporting or contradictory, leading to contingent outcomes and value chain architecture.
The promotion of RVCs is attracting greater attention by the policy community and especially so in Sub-Saharan Africa (World Bank, 2019, 234;De Melo & Twum, 2020).Our findings lend weight to these discussions.It is clear that for RVCs to be viable there has to be a commercially viable and growing end market.South Africa's relatively affluent middle-class consumer base provides that, and the presence of well-organised South African retailers has helped developing the necessary supply chains, distribution systems and retail platforms for RVCs in clothing to emerge.However, for these to grow, regional trade preferences were a key enabling factor.Yet, as we showed, trade regimes are only part of the policy framework that facilitated the growth of RVCs.Investment policies and labour regimes also matter.Public policy actors need to consider more carefully the nature of the interactions between these distinct policy levers in promoting RVCs linkages.Hence, understanding the ways in which trade, investment and labour regimes are interwoven is critical to the policy agenda.Together, these policy frameworks underlie the need for more carefully considered regional, and in some cases sector specific, industrial policy strategies.
As RVCs grow, there also needs to be a closer interrogation of how labour governance is brought into regional trade preferences.As we show, the rise of apparel RVCs in Southern Africa is also linked to the fact that neither SACU nor SADC integrate minimum labour standards in the respective trade agreements.This is in sharp contrast to the ways in which core labour standards are incorporated into the trade preferences for the US (AGOA) and EU (EBA) markets.Here, compliance with labour provisions is a prerequisite, and failure to do so risks exclusion from the agreements, as it was the case for Eswatini in 2015.Our evidence indicates that comparative labour regimes, marked by weak and poorly enforced labour legislation and weaker trade unions in Eswatini and Lesotho, ensure significantly lower wages and thus a potential 'race to the bottom' in terms of sourcing practices by lead retailers.This finding is particularly important for the African continent at a time when it is on the brink of implementing the African Continental Free Trade Area (AfCFTA).To date, AfCFTA has given no consideration to differential labour standards across the continent, and there is no social clause in the agreement.Whilst its implementation will likely benefit South African apparel retailers as they can source from Ethiopia where labour costs are amongst the lowest in the world (Whitfield et al., 2020), it will also put enormous pressure on labour market institutions as well as employment, not only in South Africa but in Eswatini and Lesotho as well.
To avoid this outcome, there is an urgent need to consider how labour provisions can be integrated within the AfCFTA protocols, thus allowing public governance to shape the architecture of existing and future regional value chains in a way that will promote more equitable forms of regional integration.This will be a significant policy challenge given the nature of uneven development in the African continent and it is one that policymakers appear to have baulked at to date.

ACKNOWLEDGMENTS
We would like to thank the ESRC-GCRF for the ES/ S000453/1 grant to which this paper is linked as part of the project 'Shifting South: decent work in regional value chains and South-South trade'.We are also extremely grateful to the three anonymous referees and the editors for the insightful comments which have significantly improved the paper.Finally, the completion of this work has benefitted from the useful comments of our colleagues of the GPN Research Group at the University of Manchester and the participants at the SASE 2020 (Network O) where an earlier version of this paper was presented.

NOTES
1 Figure 5 presents the outcome of five of the Top-6 retailers, as no data is available for the largest retailer Pepkor.Including it would raise the total market share by about 20-25% to over 90% (Barnes & Hartogh, 2019).
2 Employment data retrieved from South Africa Department of Trade and Industry (Barnes & Hartogh, 2019).
3 Analogous provisions extend duty-free access to the EU market under the Everything-But-Arms regime for least developed countries and the Economic Partnership Agreements for SADC countries.However, we focus here on AGOA, since Lesotho and Eswatini's exports to the EU are negligible.
4 Under AGOA, only South Africa does not qualify for single transformation.Its apparel exports Regional value chains in Southern Africa require 'triple transformation' -including spinning, weaving/knitting, and clothing production (Staritz 2011, 60).

Figure 1
Figure1Exports to SSA by top SSA exporting countries (trend) -thousand USD.Notes: South Africa is omitted since 96.4% of its exports is constituted by re-exports(Barnes and Hartogh, 2019).Source UN-COMTRADE.

Figure 2 Figure 3
Figure 2 Exports to SSA and the North by top SSA exporting countries (share).Source UN-COMTRADE.

Figure 7
Figure 7 Unit values for the top three exported products to the US and South Africa for Eswatini, Lesotho, Madagascar and Mauritius (2017).Notes: data for 2017 is the most up-to-date across countries.Since Eswatini was excluded from AGOA in

Figure 10
Figure 10 Weekly minimum wages for trainee and qualified sewing machinists in 2019 (ZAR).Notes: In South Africa minimum wages for sewing machinists vary between metropolitan areas (e.g. in Western Cape and KwaZulu-Natal), Giovanni Pasquali is a research associate at the Global Development Institute, University of Manchester.He completed his DPhil in 2018 at the University of Oxford, Department of International Development, with a dissertation titled 'When Value Chains Go South: Governance and Upgrading of the Kenyan Leather Sector'.Giovanni's research focuses on the governance of global value chains and south-south trade, with a particular interest on aspects related to innovation and upgrading in small and medium enterprises in Eastern and Southern Africa.Shane Godfrey is the director of the Labour, Development and Governance Research Unit at the University of Cape Town.He has a PhD in Industrial Sociology from the University of Cape Town and a law degree from the University of South Africa.He has spent the last 30 years conducting research on labour relations issues and labour standards as well as the interface between labour regulation and development.His main research interest is how to achieve decent work in labour intensive manufacturing sectors in developing countries.He has published widely on labour regulation in South Africa and is the co-author of a number of books.Khalid Nadvi is a Professor of International Development and Managing Director of the Global Development Institute, University of Manchester.He has a DPhil in Development Studies from the Institute of Development Studies, University of

Table 1
The Top-6 retail footprints in SSA Notes Since 2015, The Foschini Group, Pepkor and Woolworths have expanded beyond SSA; *Including grocery stores.Source Authors' compilation based on retailers' annual reports and interviews.

Table 2
Tariffs for major apparel exporters to the US and SACU

Table 3
Public governance of the apparel RVCs in South Africa, Eswatini and Lesotho 'At the end of the day if you're trying to sell this product for R100 and your wage cost is R35 you are on a hiding to nothing [referring to South Africa].You have to have Swaziland and Lesotho… And it is purely because you can come here [Lesotho] and compete globally… And all you do [in South Africa] is get hounded by the bargaining council the whole time, who are the ''Gestapo'' for the unions…' (Maseru, 06/12/2018)