Out with the old, in with the new : the role of the Rwandan state in pursuit of the development of the domestic textiles and garment (T&G) industry and its effects on the embeddedness of T&G firms and second-hand clothi

(1)

OUT WITH THE OLD,

IN WITH THE NEW

The role of the Rwandan state in pursuit of the development

of the domestic textiles and garment (T&G) industry and its

effects on the embeddedness of T&G firms and second-hand

clothing importers and traders

(2)

Charlie Robertson

Out with the old, in with the new: The role of the Rwandan

state in pursuit of the development of the domestic textiles and

garment (T&G) industry and its effects on the embeddedness of

T&G firms and second-hand clothing importers and traders

MSc International Development Studies 2018-19

Graduate School of Social Sciences

Charlie Robertson

12117536

15th_{August 2019}

Words: 28,336

Supervisor: Dr. Yves van Leynseele Second Reader: Dr. Niels Beerepoort

(3)

Abstract

In 2016, Rwanda launched the ‘Strategy to Transform the Textile, Apparel and Leather Sectors’ to catalyse growth in this sector as Rwanda seeks to industrialise. The Rwandan state, with an ambitious vision to become an upper middle-income country by 2035, is firmly in the driving seat and is contesting conventional notions of the role of the state in a global production network. Most often, literature points to the power asymmetries between large firms and non-firm actors, where a lead non-firm holds more sway over non-non-firm actors (the state, labourers and civil society organisations) because of the potential revenue-, employment- and knowledge-generating benefits a large transnational firm can bring. This thesis challenges this traditional conceptualisation of the state.

Using a mixed methods approach including interviews, a survey, policy analysis and trade data, this research analyses the role of the state in the development of the textiles and garment (T&G) industry and the effect this is having on the embeddedness and value creation of T&G firms and dis-embedding effect on second-hand clothing importers and traders in Rwanda. This paper found that, firstly, the role of the Rwanda state breaks from previous GPN literature. The state is not just a traditional facilitator of capital providing a conducive enabling environment for private firm activities to thrive but is also a regulator (through the import substitution approach to SHC) as well as a buyer (through public procurement) and an influencer of consumer buying habits. Through these varying roles, the state has created a complex array of fiscal and private

development policies in the T&G strategy that has led to an uneven process of embedding of T&G firms in Rwanda. Secondly, the embeddedness of T&G firms in Rwanda and the influence of the state, depends on what position the firms holds in the production network. Whether a firm is embedded or dis-embedded as a result of the state’s actions depends on the firm’s size, role in the value chain, market orientation and customer base. Finally, and most importantly, the strategy has led to a widespread dis-embedding effect on SHC importers and traders. There has been little support for these groups who have in a sense been ‘forgotten’ by the state at the expense of the promotion of the T&G GPN. This has led to a number of policy

recommendations at the end of the paper of which, most importantly, is the need to introduce re-employment programmes for those who have lost out as a result of the SHC import tariff increases.

Keywords: Global Production Networks, Rwanda, Textile and Garment Industry, Import

(4)

Acknowledgements

This research has been one of the most challenging yet rewarding experiences of my life thus far. I am incredibly proud of the final product which would not have been possible to achieve without the support of some special individuals.

Firstly, to my parents. Thank you so much for always being there for me, providing both emotional and financial support throughout my life and coming to visit me while in Rwanda. Your continued support and love for all your children is something I aspire to emulate in my life. I can only hope to be as loving and generous as you have been.

Secondly to Yves and Edward. Yves, thank you for being my supervisor throughout this journey. You have continued to provide insightful comments and a critical lens to drafts of my thesis that has kept me on my toes. Edward, to put it simply, without you this research would not have been possible. The efforts you went to in order for me to get an intern position at the Ministry of Trade and Industry were very generous. Without this position at MINICOM, I would not have been able to obtain the necessary data or to have had the ‘official’ status that made securing interviews with garment firms extremely easy. Thank you ever so much.

Thirdly, to Sonia, who helped me translate the trader survey into Kinyarwanda and then distribute it in the markets. Once again, without you I would not have been able to reach so many people. I would also like to thank the whole team at MINICOM who provided me with the trade data and answered all my questions. In particular, thank you Dr. Ossiniel for being my supervisor in the ministry and to Jean-Baptiste Ukwizagira, MINICOM’s statistician, who provided me with all the necessary trade data.

Finally, I want to thank all the incredible people I met in Rwanda who made the research trip an unforgettable experience and one I will carry with me for the rest of my life. Rwanda is truly an incredible country and one I plan to visit again very soon.

(5)

Acronyms and Abbreviations

AGOA _{African Growth and Opportunity Act}

APT _{The Association of Professional Tailors}

DMRS The Domestic Market Recapturing

Strategy

DRC _{The Democratic Republic of Congo}

EAC _{East Africa Community}

GCC _{Global Commodity Chains}

GoR _{Government of Rwanda}

GPN _{Global Production Network}

GVC _{Global Value Chains}

ILO _{International Labour Organization}

KGC _{Kigali Garment Centre}

MINICOM Ministry of Trade and Industry

MIR _{Made in Rwanda}

NISR _{National Institute of Statistics of Rwanda}

RPF _{Rwanda Patriotic Front}

RRA _{Rwanda Revenue Authority}

SEZ _{Special Economic Zone}

SHC _{Second hand clothing and footwear}

SME _{Small and medium enterprises}

STAL The Strategy for Textiles, Apparel and

Leather

T&G _{Textiles and Garments}

TNC _{Transnational corporations}

(8)

List of Figures and Tables

Figures

Figure 1: Exports of SHC 1996- 2016 Source: UN COMTRADE ... 10

Figure 2: Who imported SHC in 2016? Percentage breakdown by country Source: UN COMTRADE ... 11

Figure 3: Conceptual Framework ... 22

Figure 4: Convergent Parallel Mixed Methods design – data collection and analysis approach ... 25

Figure 5: Comparison between Industrial growth, target industrial growth and the industrial sector as a percentage of GDP (Source: NISR, 2019a) ... 33

Figure 6: Comparison between imports and exports of goods and services in Rwanda and balance of payments (% of GDP) (Source: NISR, 2019a) ... 34

Figure 7: Percentage split between imports of used clothing and footwear, new apparel and new footwear 2010 to 2015 ... 37

Figure 8: Utexrwa Factory in Kigali ... 43

Figure 9: Haute Baso Shop ... 44

Figure 10: C&H Garments Factory workers ... 44

Figure 11: Tailors in a market in Musanze ... 44

Figure 12: House of Tayo Shop ... 44

Figure 13: Exports of Apparel (2014 constant prices) Source: MINICOM, 2019 ... 45

Figure 14: Import of Intermediate Inputs (2014 constant prices) Source: MINICOM, 2019 ... 45

Figure 15: Mapping of the T&G GPN in Rwanda ... 47

Figure 16: Dyeing plant at Utexrwa ... 49

Figure 17: Import of Intermediate Inputs 2014-2018, Source: MINICOM, 2019 ... 49

Figure 18: Domestic textiles production Source: MINICOM, 2019 ... 49

Figure 19: Police uniform ... 50

Figure 20: Young tailors in training sponsored by APT ... 51

Figure 21: Queues of workers outside C&H Garments waiting to be interviewed ... 53

Figure 22: Destination of Apparel Exports by Value (2014 constant prices) Source: MINICOM 2019 ... 54

Figure 23: Rwandan Fashion Week 2017 Source: Collective Rw ... 56

Figure 24: Sorting and grading warehouse of a SHC importing firm... 61

Figure 25: SHC in Gisozi Market ... 62

Figure 26: Example of a SHC invoice with shipping costs ... 63

Figure 27: A mapping of the SHC production network in Rwanda based on the importers and traders interviewed ... 64

Figure 28: Exports and Re-exports of SHC from Rwanda (USD, 2014 constant prices) Source: MINICOM, 2019 ... 66

Figure 29: Monthly profits of SHC traders pre and post tariff increases (2014 constant prices) Source: MINICOM ... 68

Figure 30: Sourcing strategies of traders pre and phase out ... 69

Figure 31: What will customers do? ... 70

Figure 32: Informal imports of SHC. Source: RRA ... 70

Figure 33: Import quantities of new Apparel, footwear and SHC (metric tons) Source: MINICOM ... 72

Tables Table 1: Typology of the role of the state in GPNs* ... 18

Table 2: Distribution of surveys by location ... 27

Table 3: Key roles of the state in the development of the Textiles and Garment GPN ... 41

Table 4: New Apparel imports (KG and $) into Rwanda 2014 – 2018 (2014 constant prices) Source: MINICOM ... 46

Table 5: Price comparison between Hema Garments and SHC ... 57

Table 6: Total imports of Used Clothing and Footwear 2014 – 2018 (KG and USD, 2014 constant prices). Source: RRA ... 61

Table 7: What will traders do? ... 71

(9)

1. Introduction and context

This introduction will first situate this research within two concurrent contexts that both inspired and informed the undertaking of this work. Firstly, the ‘Africa Rising’ narrative and a recent resurgence and interest in African industrial policy and secondly, the development debate around second hand clothing and footwear. Finally, the problem statement and gap in literature will be identified before outlining the key research question and sub questions.

1.1. ‘Africa Rising’ and Industrial Policy

‘Africa Rising - the hopeful continent’ (The Economist, 2011), ‘The African Growth Miracle’ (Young, 2012) and ‘The New Scramble for Africa’ (Carmody, 2011) are just three titles many news and journal articles have employed to narrativize the high single and double digit growth of many Sub-Saharan African (SSA) countries since 2000. This powerful narrative depicts a

continent brimming with confidence, digitally enabled citizens, increasing numbers of

entrepreneurs, dynamic economies ‘open’ for business and millions of people alleviated from poverty to ‘middle class’ status. The theory of ‘leap-frogging’ has been applied to SSA where the harnessing of technological innovation has seemingly removed the need for countries to develop the sort of heavy manufacturing that catalysed growth in China, South Korea and South East Asia in 1980s and 90s.

However, there has been doubt over the sustainability of this growth. Academics and journalists argue African growth has been bloated and distorted by the steroidal effect of raw commodity exports as well as ‘leap-frogging’ industry in favour of the development of the service sector (Chang, 2015, Behuria and Goodfellow, 2018, Newman et al, 2016). Structural economists argue that without the development of employment-intensive manufacturing, countries cannot expect to achieve the same levels of economic transformation (GDP growth plus a structural shift from agriculture to industry) that will increase productivity and generate higher levels of employment. As such, there have been growing concerns that the astronomical GDP growth in Africa is ‘jobless’ or low-quality growth, with the ILO pointing to youth regularly suffering from under-employment or lack of decent jobs (ILO, 2018a).

Rwanda, a country with a recent dark history with the 1994 genocide, falls perfectly into the ‘African Rising’ narrative. The country has seen extremely impressive levels of GDP growth in the last ten years hitting over eight percent year-on-year with a focus on ‘leap-frogging’ industry and becoming a ‘knowledge-based economy’ (MINICOM, 2012). Consequently ICT, tourism and transport are outpacing manufacturing, with tourism accounting for 30 percent of total export figures (Newfarmer et al, 2018). In 2018, the service and agriculture sectors represented 48 and 29 percent of GDP respectively, while industry was just 16 percent (NSIR, 2019a). However, regardless of Rwanda’s growth, the state is still experiencing high levels of labour underutilisation standing at 57 percent in February 2018 (NSIR, 2019b). As a result, there has been a recent push by the government to focus on employment-intensive manufacturing. The recent release of the National Transformation Strategy is aiming for the industrial sector to reach 21.7 percent of GDP by 2023 (MINECOFIN, 2017), a far cry from the current 16 percent. Rwanda, through the development of its industrial policy, has targeted key sectors for industrialisation and has divided these into short-, medium- and long-term investments. The ‘quick-wins’ include agro-processing, ICT, tourism and most importantly for this research, the textiles and garments industry (MINICOM, 2012). The ‘Strategy to Transform the Textile,

(10)

Apparel and Leather Sectors’ was launched in 2016 and sought to catalyse growth in the industry through a comprehensive range of fiscal and economic development initiatives that will be explored in this thesis. However, the most controversial initiative was the import substitution industrialisation policy to increase tariffs on second-hand clothing and footwear (SHC). In effect, the state hopes to replace the consumption of SHC with domestically produced garments and footwear.

1.2. Second-hand clothing and industrialisation

‘Have a clear-out and beat poverty too’ (Oxfam, 2018) and the ‘Give Up Clothes For Good’ (T.K.Maxx, 2018) are just two examples of how SHC is marketed as an ethical commodity in the Global North. According to a recent study, the average consumer in Europe buys 60 percent more apparel items than 15 years ago and wear them for half as long (Ellen MacArthur

Foundation, 2017). As a result, exports of donated clothing have rapidly increased in the last 12 years from $1 billion in 1996 to $3.7 billion in 2016, peaking at $4.5 billion in 2014 with Europe accounting for the largest export share (UN Comtrade, 2019a). SHC is normally processed in the country where it has been collected before it is sorted and graded to determine the value based on the uniqueness, quality and brand recognition of the garment or footwear (Brooks, 2013). There are then four options, the best clothing is resold in domestic vintage or charity shops, the worst quality is sent to a rag merchant for recycling, the rest is either exported by an NGO or sold by the NGO to a commercial exporter who then exports the SHC to importers normally in

the global south (see appendix VII for a mapping of the value chain).

While donating clothing is seen as an ethical and sustainable activity to rid citizens in the Global North of their overconsumption of apparel and footwear, little attention is given to where the clothing is exported to by those who donate (Norris, 2015, Brooks and Simon, 2012). The largest share of the imports is exported to Africa with Ghana and Kenya accounting for 7.68 percent of total global imports in 2016 (figure 2). However, the importing narrative for SHC differs from

(11)

that of exporting partners insofar as it is seen as both an economic and health hazard by many importing countries. Regarding hygiene, many countries forbid the import of used underwear or demand a certificate from the exporter which verifies that the SHC has been disinfected or fumigated by a licensed agency (Brooks and Simon, 2012).

Vis-à-vis economic concerns, many importing countries argue high levels of SHC depress

domestic demand for local produced garments or footwear and consequently stunt the growth of domestic local industry (Calabrese et al, 2017). This has led certain countries to increase tariffs or

introduce outright bans on SHC imports such as in Indonesia and Lesotho (Hansen, 2004).1

There is debate among academics about whether this stunting effect is in fact true. Indeed, a recent study analysed the link between the increase of SHC and decreased garment production finding that SHC had a negative impact on garment production accounting for 40% of the decline and 50% of the decline in employment (Frazer, 2008). However, other economists argue this study finds correlation and does not explain causation. Instead, they point to the same causal factor, globalisation and trade liberalisation, as the key explanation for both the decline of the African garment sector since 1970s and the rise of SHC (Brookes and Simon, 2012, Balchin et al, 2016).

Regarding Rwanda, in its pursuit of industrial development and catalysing growth in the textiles and garment industry, the state took the decision in 2016 to increase import tariffs on SHC from $0.2 to $2.50 per KG and $5 per pair of shoes. As will be discussed below this was not in

conjunction with its East Africa Community (EAC) peers who, after threats from the U.S. that

1_{For a list of all country tariffs, bans and regulations for the import of used clothing and footwear see this list}

compiled by the Office of Textile and Apparel of the US Department of Commerce:

https://otexa.trade.gov/worn_clothing.pdf

(12)

access to the African Growth and Opportunity Act would be suspended, decided to withdraw from a unified EAC phase out. Rwanda’s decision to phase out SHC has been subject to suspension from AGOA for apparel. President Kagame has been defiant in the face of this arguing that, ‘Africa does not need to be a dumping ground for… second hand anything’ (Rwanda Today, 2018).

1.3. Problem statement and gap in literature

The two sections above contextualise Rwanda as a country ‘on the way up’, reorienting its impressive growth towards industrialisation and targeting the textiles and garment (T&G) sector to catalyse industrial development. As will be shown below it has developed a comprehensive strategy to transform the nascent T&G sector using a mixture of fiscal and policy levers of which import substitution through the phase out of SHC is just one. When I started this research, my intention was to just focus on the effects of the phase out of SHC on domestic garment

production and on the SHC industry itself. But what I realised is that the effects of this strategy cannot be isolated from the context and role of the state. What are the different roles the state is playing in its attempts to create a textile and garment industry in Rwanda and what effect is this having on the T&G firms and the SHC industry?

The state is in control of this transition due to the limited number of firms in the country and is aiming to produce ‘higher quality and increased volumes for both local and foreign markets while reducing imports’ (MINICOM, 2016). As a result, the state needs to be seen more than just a facilitator of capital and or institutional extra to the traditional star of the show –lead firms. As will be highlighted in the theoretical framework, global value and production network literature focuses most centrally on the interaction of lead firms with non-firm actors (Gereffi, Humphrey, & Sturgeon, 2005; Ponte & Sturgeon, 2014). Most often, literature points to the power

asymmetries between large firms and non-firm actors, where a lead firm holds more sway over non-firm actors (the state, labourers and civil society organisations) because of the potential revenue-, employment- and knowledge-generating benefits a large transnational firm can bring. This thesis challenges this traditional conceptualisation of the state.

The Rwandan state, with an ambitious vision to become an upper middle-income country by

2035,2_{is firmly in the driving seat and is contesting conventional notions of the role of the state}

in a global production network. As will be shown, the state is not just a traditional facilitator of capital providing an conducive enabling environment for private firm activities to thrive but is also a regulator (through the import substitution approach to SHC) as well as a buyer (through public procurement) and an influencer of consumer buying habits and centralisation of

production through the development of tailoring cooperatives. Through these varying roles, the state has immense power to manipulate the ‘embeddedness’ of T&G firms in Rwanda while ‘dis-embedding’ SHC importers and traders from their ability to trade effectively which has put them into an economically precarious position. In effect, the state is implementing a form of import substitution industrialisation. However, this form of import substitution industrialisation is not trying to remove large international lead firms to make way for the rise and embeddedness of domestic T&G firms, but instead to remove and dis-embed SHC domestic importers and traders who have relationships with importers from the global north (Brooks, 2012). As such, this thesis takes the concepts of the ‘role of the state’, ‘embeddedness’, ‘value creation’ and ‘agency’ (all of which will be defined below) to analyse the role of the state in the development of T&G industry and what effects their role has had on both the T&G and SHC industries. This thesis will end,

2_{The World Bank defines upper-middle economies as those with a gross national income (GNI) per capita between}

(13)

after analysing the role of the state and the effects of embeddedness on T&G firms and SHC players, with policy recommendations.

1.4. Research question and sub-questions:

What is the role of the Rwandan state in the process of (dis)embeddedness of the textiles

and garment as well as the second-hand clothing and footwear global production networks,

and what effects does this have on value creation and worker agency?

SQ1: How can the role of the Rwandan state be conceptualised in its pursuit of the

development the textiles and garment industry? Are there any tensions between the various roles?

SQ2: How is the role of the Rwandan state creating embeddedness for the textiles and

garment firms and what effect is this having on value creation?

SQ3: How is the role of the state as a ‘regulator’ dis-embedding second clothing (SHC)

importers and traders and disintegrating value in the SHC industry in Rwanda, and

how does the agency of importers and traders affect their ability to cope and respond to

(14)

2. Theoretical Framework

The theoretical framework will first begin with an outline of global value chains and production network literature. Specifically, it seeks to identify the gap in GPN literature that normally focuses on the state as an ‘institutional’ backdrop to all powerful protagonist of the ‘lead firm’. This thesis challenges this notion. Secondly, the key concept of ‘embeddedness’ in GPN literature is outlined and explains how this applies to the Rwandan T&G as well as the SHC industry. Finally, the concept of ‘worker agency’ is discussed specifically in relation to the economically perilous situation SHC importers and traders find themselves in. The conceptual scheme is presented at the end tying all the concepts together. For a detailed operationalisation see appendix III.

2.1. Global production network approach

Over the last twenty-five years, a global commodities and value chain literature has emerged that attempts to explain how global industries are governed and organised as well as explaining how the relationships in these industries affect development and upgrading opportunities for the actors involved (Gereffi, 1994, Gereffi and Sturgeon, 2013). As globalisation fundamentally changed the trade dynamics between firms, industries and nation-states, GCC and GVC analysis provided a framework in which to analyse the role of the lead private firm (Gereffi, Humphrey, & Sturgeon, 2005; Ponte & Sturgeon, 2014). In essence, GVC analysis looks at the value-adding activities of various firms along a globally dispersed input-output chain or network to bring a

product to market, from conception to final use.3

However, the global production network (GPN) framework was developed in 2002 that combined the insights from GCC and GVC literature with actor-network theory to explain the multi-scalar and actor characteristics present in a transnational production system (Henderson et

al, 2002, Coe et al, 2004, Coe et al, 2008).4_{GPN theory’s main point of departure from}

GVC/GCC literature is that it does not view production and distribution processes as linear or vertical in a ‘chain’, instead it views these as a multi-scalar network of functions and operations involving a variety of firm and non-firm actors through which goods are produced and

distributed. A network can be seen as a ‘nexus of inter-connected functions, operations and transactions’ between states, firms, civil society operations, consumers and workers operating at multi-geographical scales (Coe et al, 2008, p.272). GPN literature attempts to go beyond GVC analysis to explain the ‘institutional framework’ around which lead firm operates and a key focus is analysing the complex social processes and relationships between firm and non-firm actors in the production and distribution of goods (Henderson et al, 2002, Coe, 2013, MacKinnon, 2012). This allows for an understanding of the dynamics of power and dissemination of knowledge as multi-directional and non-deterministic at multi-geographical scales.

3_{The key debates on GVCs can be summarised as follows: Sturgeon, 2001 Henderson et al, 2002, Gereffi et al,}

2005, Ponte and Gibbon, 2005, Ponte and Sturgeon, 2014 and Neilson et al, 2014

4_{The conceptual difference between GCC and GVC is small but should be noted. GCC is primarily concerned with}

how global industries are organised and involved mapping the key actors and relationships between them in the production and distribution of a particular good (Coe et al 2008). A similar piece of work for GCCs was the 1994 article by Gereffi who outlined the different between buyer and producer-driven value chains. GVCs, while very similar, are primarily concerned with the governance structures both within and between different global industries and the facilitation of knowledge transfer (Dolan and Humphrey, 2004, Gereffi et al, 2005).

(15)

Underpinning the analysis of a global production network are the key intersecting conceptual categories of power, value and embeddedness that can explain the configuration and

coordination of the actors in the production network:

- Value is analysed by looking at the way firms can create, enhance and capture surplus and economic rent through a number of differing activities, such as investing in skills training of employees to increase productivity. For the purpose of this thesis, only ‘value creation’ will be operationalised as the textiles and garment (T&G) industry is still in its infancy. This key concept will be elaborated upon in the theoretical framework below. - Power in production network is analysed across three main categories: corporate,

institutional (national, international and supranational) and collective (for example trade unions, CSO). This thesis is focusing on the power of the state but looks at the

institutional and collective power that is influencing the state’s role in the development of the T&G GPNs.

- Embeddedness is conceptualised as the social and spatial arrangements in which firms are embedded which influence their strategies in production and distribution.

Embeddedness can be conceptualised under three broad categories: network, territorial and societal. Network embeddedness refers to how firm and non-firm actors interact with each other that is affected by the level of trust between them (Coe and Hess, 2006). Territorial embeddedness refers to how firms ‘anchor’ themselves in a territory and how economic activities play out in a physical space. Finally, societal embeddedness refers to ‘make up’ or cultural background of firms that influences their economic actions outside of their home territory (Alford, 2016). For the purpose of this research, territorial and network embeddedness will be subject to more focus that societal embeddedness. This is largely because the Rwandan textiles and garment industry is in its nascent stages, there are not many foreign firms in the country and this this dimension is less applicable. Looking specifically at the role of the state in GPN literature, it is conceptualised as but one element in the ‘institutional power’ of non-firm actors wielding influence over GPNs

(Henderson et al, 2002). An institutional lens is crucial to understand how ‘vertical’ value adding activities of lead firms interact with the ‘horizontal’ territorial or social relationships that create the ‘enabling environment’ for lead firm activities. Yet theorisation of the state is largely limited in GPN literature to a set of activities in facilitating and regulating to enable production and distribution of goods by private firms (Yeung, 1998). This includes functions such as

guaranteeing private property rights and contracts through national legislation as well as trade roles and frameworks for foreign exchange controls. What is lacking in the literature is a framework to explain when a state goes beyond just an ‘institutional’ actor providing the backdrop to the accumulation process in a production network. This is the case with the

Rwandan state in its pursuit to embed the new T&G production networks while actively seeking to dis-embed used clothing and footwear production networks. The state is taking an active role in manipulating and restructuring the embeddedness of firms and labour via the implementation of a comprehensive textiles and garment strategy.

To theorise why the Rwandan state has taken such a stance and conceptualise the evolving role of the state within global production networks, we need to see this industrial policy decision as not just a form of functionary capitalism but as a result of its changing position and dominance in the formation of global production networks. As such, the emphasis is not on the industrial policy itself but on the configuration of social forces at both national and international levels that underpin state support for specific policy directions. We need to explain why the state is

supporting certain policies over another. In order to do this, a reconceptualization of the state is needed within GPNs. A typology developed by Rory Horner (2017) will be adapted below to

(16)

analyse the role of the Rwandan state and its effect on the textiles, garment and SHC production networks.

2.2. The role of the state in GPNs

GPN literature with its relational approach goes further than GCC or GVC in conceptualising the role of the state with its focus on the institutional landscape that provides a broader

understanding of non-firm actors and governance in the make-up of GPNs (Yang & Coe, 2009, Henderson et al, 2002). Yet the state is seen as but one institutional actor in the GPN

framework.

Recently there has been an advancement in the literature that focuses specifically on the role of the state in GPNs. Glassman as well as Arnold and Pickles have argued the need for more focus on the geopolitical framing in which GPNs are situated (Glassman, 2011, Arnold and Pickles, 2011). Glassman argues that the global economy is dictated by more political than economic imperatives giving the example of the political significance of the US interest in forming GPNs in South Korea. Perhaps the most significant contribution so far to state involvement in GPNs is the work of Adrian Smith (2014, Smith et al, 2018). Applying a strategic-relational approach, Smith shows how the state can operate at different geographical scales to govern GPNs and dictate the terms of economic integration in order to align its geo-political and economic interests (2014). Smith uses the example of Tunisian and EU economic integration, a mutual beneficial move for both institutions. Firstly, for the EU to find new avenues of capital accumulation and competitiveness while attempting to ‘manage’ its southern border and secondly to satisfy the Tunisian political economy and development agenda. As such, Smith is linking the function and governance of GPNs with the wider political economies in which they are embedded.

However, this approach still focuses on the governance of GPNs by a state in its attempt to align geo-political and economic interests as opposed to their active participation in a GPN itself. Looking at Rwanda, the motive to develop the T&G sector is evidently driven by economic

development goals (see chapter 4) but the more interesting question is how the state is using its

position beyond just a governor or facilitator of capital to develop this GPN without the presence of large TNCs. In order to advance the role of the state, a framework is needed to critically assess the activities of the state as a ‘protagonist’ in the formation, development and management of GPNs. Horner, adapting a typology from Evans (1995) on the state roles in developmental states, has developed a four-dimension typology to examine roles of the state (2017). There are four dimensions to the role of the state in his framework: facilitator, regulator, producer and buyer. These will be elaborated upon before indicating how this framework will be adapted for this analysis.

The role of the facilitator is most commonly associated with the state in GVCs and GPNs (Gereffi, 1994, Gereffi & Sturgeon, 2013). The role is typically concerned with promotional activities and policies to assist firms in GPNs in relation to the opportunities and challenges faced by the global economy. This can include the development of export processing zones, tax incentives or subsidies and the upgrading of infrastructure to facilitate trade. It can also include interstate and regional trade negotiations to develop favourable trade policies for local firms. Policies can generally be broken down into ‘horizontal’ economy-wide initiatives, ‘vertical’ industry specific policies or GVC-centred policies that aim to leverage international linkages that lead to firm upgrading (Horner, 2017). Mayer and Phillips argue that this role conceptualises the stage not as a ‘passive’ agent but an active architect in creating the conditions for the

(17)

The role of the regulator involves policies or measures that limit or restrict involvement in GPNs or active deregulation. These can include tariff controls or quotas on imports (very relevant to Rwanda), labour regulations and standards implementation (Horner, 2017, Dicken, 2015). Regarding de-regulation this can include outsourcing certain regulatory functions like standard-setting to private actors (Mayer and Phillips, 2017). The use of restrictive regulatory policies has declined since the era of the World Trade Organisation (WTO) in 1994 which banned

quantitative restrictions such as quotas and reduced policy space for import tariffs as well as subsidies (Chang, 2015). However, there has been a recent rise in protectionism and regulatory policies, for example the on-going trade war between the US and China (BBC, 2019), figures citing 400 new protectionist policies globally since 2009 (The Economist, 2013) and the rise of the ‘smart protectionism narrative’ (The Atlantic, 2016).

Yet, beyond a purely public governance role as the facilitator and regulator, the state can be seen as an active participant in GPNs with the roles of buyer and producer. The buyer role is

associated with large scale public procurement programs from private firms (Horner, 2017). For the OECD countries, public procurement expenditure represented 12% of GDP, ranging from 5.1 percent in Mexico to 20.2 percent in Netherlands (OECD, 2017). For the producer role, state-owned enterprises (SOEs) have been used to kick start production in GPNs where the private sector does not see a competitive advantage (Cuervo‐Cazurra et al, 2014, Horner, 2017). SOEs represented 22.6 percent of the Fortune Global 500 list in 2014 (Augustynowiz,

Kwiatkowski, 2015). However, as Horner notes, there is a considerable gap in GVC/GPN literature on the influence and viability of these two state roles (Horner, 2017).

For Rwanda, the role of buyer is relevant as the government is implementing a large-scale garment public procurement program that will be explored in the empirical chapters. In the adaptation of this framework, the role of ‘influencer’ will be explored as opposed to ‘producer’ as there are no state-run or owned T&G manufacturers. The role of ‘influencer’ is in relation to the concerted effort by the Rwandan state to influence local consumers to buy Rwandan clothing with the ‘Made in Rwanda’ campaign. Relating this specifically to the literature, scant attention has been placed to on the role of the domestic market in GPNs (Yang et al, 2012, Horner, 2014). Specifically capturing the domestic market can be the ‘first step’ towards economic development enabling domestic firms to successfully export within GPNs and prior to coupling with global lead firms (Horner, 2014). As such, with Rwanda and the Domestic Market Recapturing Strategy (2015), this dimension is necessary to explore the influence of the Rwandan state on the

(18)

*Adapted from Horner, 2017. Yellow – key dimensions to be operationalised. Green – not currently applicable to research context.

This adapted typology (table 1) will used to explore and analyse the role of the Rwandan state in its development of the T&G production networks at the expense of the SHC trading. At once both creating the dual conditions for the ‘embedding’ of domestic and foreign T&G firms while ‘dis-embedding’ SHC importers and traders (see below). Much of the literature places the emphasis on the role of the state as a facilitator and regulator of economic coordination, rather than examining the broader political-economic context at international, regional and domestic scales that a state is embedded within that influence its role in GPNs (Alford & Phillips, 2018, Alford, 2016, Smith, 2015). As such, this thesis will examine both the type of the roles the Rwandan state is taking and question the motivations behind them due to international, regional/bilateral and domestic pressures. This will lead to a discussion around the tensions which rise as a result of these differing roles.

2.3. Embeddedness and value creation in global production networks

One key aspect of global production networks as noted above is the category of ‘embeddedness’. The concept was first theorised by Karl Polanyi in 1944 and further developed by the economic sociologist Mark Granovetter in 1985 to understand the evolution and economic success of regions built by locally clustered firms. It looked at local and regional economic system and showed how local embeddedness of actors led to institutional ‘thickness’ (the influence of institutions in a regional economy) that was seen a critical factor for regions in the globalising economy (Hess, 2004, Weller, 2006).

In recent conceptualisations of embeddedness, it has come to define the social relationships between economic and non-economic actors with economic action grounded in these structures (Hess, 2004). The concept of ‘territorial’ and ‘network’ embeddedness provides a useful

framework in which to understand how the varying roles of the Rwandan state outlined above Table 1: Typology of the role of the state in GPNs*

(19)

are manipulating and changing the ‘embeddedness’ of firms in both the development of T&G production and the active ‘dis-embedding’ of the SHC production network. Hess conceptualises ‘network’ embeddedness as the network of actors a firm or organisation is involved in and the

relationships between these actors.The durability and stability of these relations affects the

actor’s embeddedness as well as the structure of the network as whole. For ‘territorial’

embeddedness Hess considers the extent to which an actor is ‘anchored’ in a particular territory. Certain strategic actors in a GPN may take advantage of certain regional assets that pre-date their arrival.5

Looking specifically at how this is relevant to this research project, embeddedness analysis is typically focused on lead firms, specifically how territorial and network embeddedness affects and informs lead firms actions and activities (Hess, 2004, Weller, 2006). However, analysis does not traditionally focus on how the state can manipulate the embeddedness of firms through a number of differing roles in pursuit of the development of certain GPNs (Alford, 2016) and what effect this has on the creation or disintegration of value. This is the focus of this thesis. As outlined above, ‘value’ in GPNs is the way firms can create, enhance and capture surplus and economic rent through a number of differing activities, such as investing in skills training of employees to increase productivity. However, by analysing the role the state has played in the (dis)embedding process of firms in the T&G and SHC GPNs, we can look at how this leads to a shift in value from the dis-embedded SHC importers and traders to new value and rent creation in the nascent T&G production network. To examine how the state’s role is creating value in the T&G GPN, the dimensions of domestic production, creation of backward and forward linkages,

trade and employment will be operationalised (see appendix III).

2.4. Worker agency in relation to dis-embedding and disintegration of value

The Rwandan state’s role in its dual effect of embedding firms and creating value in the T&G GPN is contrasted by the dis-embedding effect on the used clothing and shoes GPN that is disintegrating value that affects worker agency. We can define the ‘dis-embedding’ process as the way in which state policy has reduced space for firms and workers to ‘anchor’ themselves in SHC production networks in Rwanda. In effect, through tariff increases, importers are no longer as easily able to access international exporters as prices for the commodity are too high. This affects the sustainability of their business models and thus their territorial embeddedness within

Rwanda. Regarding ‘network’ dis-embeddedness, the complex production and consumption networks built up around the used clothing industry have been reduced or destroyed due to a collapse of the industry and disintegration of value. The last question will examine in what way the process of dis-embedding and disintegration is occurring in Rwanda and how it is affecting workers.

Linked to the process of dis-embedding, is the concept of worker agency. Worker agency here is defined as ‘how workers seek and terminate particular types of employment, make incremental improvements in their work lives, and are embedded in broader social and community relations’ (Lund-Thomsen, 2013, p.74). Worker agency has typically been associated through collective action in unions (Nielson & Pritchard, 2009, 2010), however here it is conceptualised as the efforts of individual workers at the micro-level to ‘get by’ or to cope with challenges (Carswell and De Neve, 2013). This is because there are no unions or formal groups defending the rights of SHC traders in Rwanda. In regard to this study, this concept is integral in understanding how

5_{This concept is call ‘Strategic Coupling’, a useful concept for understanding how firms take advantage of regional assets.}

However, this thesis is looks specifically at state actions instead of firm activities. For further reading on this concept see: (MacKinnon, 2012, Yeung, 2009, Yeung, 2016)

(20)

the role of the Rwandan state and the powerful dis-embedding process is affecting workers in the SHC GPN. In effect, how are actors in this disintegrating GPN coping and responding to this process of dis-embedding? Three dimensions will be used that are taken from a

conceptualisation of labour agency by Katz (2004):

- Resilience: This is to do with the way’s individuals or groups cope with everyday realities that do not necessarily challenge the status quo. Looking specifically at Rwanda SHC importers and traders, resilience looks at how importers and traders are finding ways to continue trade despite the increase in import tariffs such as by trading in neighbouring countries.

- Resistance: This dimension is used to assess they ways in which workers challenge capitalist relations through attempts to develop alternative systems of production that provide workers with greater control of their labour time and use. Looking at Rwanda, resistance refers to the way SHC importers and traders may actively be ignoring the government’s wishes to curb all SHC trade by either finding illegal or informal routes to continue trade in Rwanda.

- Reworking: This is dimension looks at how workers seek to improve their situation by either seeking reform or finding other ways to improve their conditions or employment situation. For Rwandan SHC traders and importers, reworking looks specifically at re-employment strategies to move away from SHC trade and start new trading in new products.

This thesis also draws upon the analytical framework of Neilson and Pritchard who recognise that worker agency should not be seen just in terms of the ‘vertical’ dimensions of GPN structure but also the ‘horizontal’ embeddedness of local places and institutions (2009). In effect, worker agency can be analysed as the conflict or tension arising from the intersection of the ‘vertical’ dimension (normally associated with lead global firms) and the ‘horizontal’

dimension (the local institutional context such governments and local workers). However, in the case of the Rwanda, the intersection of the ‘vertical’ and ‘horizontal’ is blurred as the state is influencing the ‘vertical’ dimension (through its governance role) that is directly affecting the ‘horizontal’ dimension (the local embeddedness of workers). As such, this thesis is addressing the issue of worker agency from a new angle.

2.5. Conceptual framework and operationalisation

Figure 4 below visually depicts the conceptual framework that shows the relationship between the main concepts of this study: the role of the state, (dis)embeddedness, value creation and work agency. The blocks in grey on the left are the multi-scalar political influences that inform the role of the Rwandan state and the development of the textiles and garment strategy. These

influences are explored in detail in the chapter 4 when analysing the variegated role of the state.

The four roles the state is taking are depicted in the middle column. Specifically, chapter 4 looks at how the variegated roles have created tensions between its differing roles. Through these roles, the state is trying to create the optimal conditions for firms to ‘embed’ in Rwanda as

represented by the value creation block (see chapter 5). The circular-looped arrows between the

role of the state and the value create T&G block show that there is a constant feedback and communication loop between the small number of garment firms and the state as the

government tries to catalyse growth. However, this does not necessarily mean the government is creating the ideal conditions or listening to the needs of all firms, as will be explored in chapter 5. The five purple blocks at the top represent the dimensions this thesis uses to measure value creation. Finally, the work agency block represents the response of SHC importers and traders to

(21)

the severe process of dis-embedding they experience as a result of the role of the state,

specifically as a regulator in increasing import tariffs on SHC. This is explored in chapter 6

through the dimensions of resistance, resilience and reworking. The line between the roles of the state and the agency of workers represents the lack of support and communication between the state and SHC importers and traders during the implementation of the T&G strategy.

The major concepts underlying this research have been operationalised to show their application in appendix III. Each concept has been split into dimensions, variables and indicators that informed the approach to primary data collection.

(22)

(23)

3. Methodology

3.1. Philosophical position

The philosophical stance of this research follows a pragmatist framework insofar as the

worldview ‘arises out of actions, situations and consequences rather than antecedent conditions’ (Creswell, 2013, p.10). As such, instead of focusing on methods, this research focuses on the problem (increased import tariffs on SHC), and the necessary research approaches to understand the problem. The ontological belief of this thesis adopts a ‘what works’ mixed methods approach to collect and analyse data. This is particularly relevant to this study, as this study seeks to

understand the effects of the tariff increases from multiple levels. Firstly, to understand the overall trade picture and state policy decisions (through policy and economic analysis of trade data), secondly through the small but growing number of domestic T&G firms (through semi-structured interviews) and finally through SHC importers and the number traders (through the distribution of a survey). The epistemology of this study is practicality, allowing for use of multiple research tools to get a comprehensive understanding of the role of the state and effects of the SHC tariff increases on the full range of relevant actors (Creswell, 2013, p.11). According to pragmatists, research also occurs in social, historical and political contexts that directly affects the research participants involved (2013, p.11). For this research, this is of paramount

importance as the import tariff increases on SHC in Rwanda cannot be explained without reference to the political development trajectory of Rwanda and the divisive issues of SHC between the global north and south (as explained in the introduction).

3.2. Research design and methodology

This research follows a convergent parallel mixed method design as defined by Creswell (2013). A convergent mixed methods design uses both qualitative and quantitative data which are collected and analysed separately before they are compared with one another in a number of different combinations. Looking at figure 3, this research collected four different types of data including trade data on the import/export of SHC and new apparel and footwear, government policy documentation, semi-structured interviews (with government officials, T&G firms and SHC importers) and a survey of SHC traders. This was because the research aimed to gain a holistic understanding of the state’s far reaching and comprehensive strategy to industrialise and catalyse growth in the T&G GPN. Data analysis then took place in two phases where different mixtures of the qualitative and quantitative data were ‘converged’ and analysed to triangulate the results, this can be summarised as follows:

- Data analysis phase 1: To answer the first sub-question on the role of the state in its pursuit of import substitution industrialisation and development of the T&G GPN, policy documentation was analysed alongside semi-structured interviews with a number of governmental officials to get a comprehensive picture of the state’s role and motives for their approach.

- Data analysis phase 2: Phase 2 can be broken into two parts that answer sub-question two and three consecutively.

o Phase 2.1: To understand how the role of the state and its policy interventions have affected the development and embeddedness of domestic T&G firms in Rwanda, semi-structured interviews with firms were triangulated with trade data.

(24)

In this way, if firms suggested there had been an uptick in exports of garments, the trade data could be used to verify these claims. It was also used to estimate whether production had increased by looking at the import of intermediate inputs (imports of cotton, fabrics etc).

o Phase 2.2: To answer sub-question 3 on how the role of the state dis-embedded SHC actors from Rwanda and how their agency has been affected, three sets of data were triangulated. Specifically, semi-structured interviews with former SHC importers (of which there are only a few) and data from a survey of SHC traders in three markets (of which there are many) were analysed to see whether the SHC tariff increases had affected these two key actors in the same way or differently. This was triangulated with trade data to see if imports had drastically reduced as a result of government intervention.

(25)

(26)

3.2.1. Units of Analysis and Sampling

The key unit of analysis for this study was the Rwandan state and the effects of state intervention on key T&G players and SHC importers and traders. In order to understand the policy

mechanisms employed by the state, policy documentation and trade data obtained from the Rwandan Revenue Authority (RRA) were analysed to understand the Rwandan state’s role and influence on the development of the T&G GPN. Secondly, the semi-structured interviews were conducted with actors involved in the development of the T&G GPN including government officials and T&G firms to understand if embedding and value creation had taken place. Finally, a survey and semi-structured interviews were conducted with actors involved in the SHC GPN to understand the dis-embedding effect and response to the increase of SHC tariffs. Taken together these three units of analysis were used to understand the powerful role of the state in the development of the T&G GPN and the (dis)embedding effect it has on both T&G firms as well as SHC actors.

Purposive sampling was used for a large portion of the study as it allowed for targeted selection of non-random participants most relevant to this research (Kemper et al. 2003). I obtained an internship with the Ministry of Trade and Industry (MINCOM) via my local supervisor (a PhD student from the University of Amsterdam) who acted as my main source of contact and connections while in Kigali. MINICOM provided me with very useful contact lists of all the key T&G firms in the country and former SHC importers. These lists formed the basis of my semi-structured interviews. As the T&G sector is relatively new, there are not many key firms and thus I was able to interview almost all of them. MINICOM also provided me the key policy

documentation around the economic development goals of the state and the 2016 T&G strategy and implementation road map. These documents were used during phase 1 of the data analysis to understand the role of the state. The Rwandan Revenue Authority, via MINICOM, provided me with the trade data (HS codes 50 to 64) from 2014 to 2018 that was used to analyse the trade

flows of SHC and new apparel and footwear (see appendix VIII for a breakdown of how the

trade data was organised) .

However due to the informal nature of the SHC industry and with no official representation by a trade association, it was more difficult to conduct purposive sampling with SHC traders. As a result, I developed a survey that was distributed in three markets in Kigali to understand the effects of the tariff increases on their ability to trade and how it has affected their agency. Snowball sampling was used to find traders in the three markets who were comfortable in filling out the survey (Becker et al, 2012). As there are no statistics on the number of SHC traders in the country, it was not possible to estimate a sample size with a 95% confidence interval that could have been used to generalise the results to all Rwandan SHC traders. As such, the results of the survey are only representative of the respondents in Nyabugogo, Kimironko and Gisozi-Duhahuriane markets. However, this was triangulated with trade data that showed that imports of SHC drastically reduced in the country, suggesting the negative effects felt in these three markets by traders are most likely generalisable to other markets.

3.2.2. Semi-structured interviews

25 people were interviewed in Rwanda ranging from governmental officials, managers or owners

of T&G firms, SHC importers and a number of associations (see appendix I). These interviews

were carried out in person. For T&G firms, a script was followed and similarly for SHC

importers (see appendix IV and V) most of which were conducted by myself in French or

English. For some of the interviews, participants did not want to be recorded due to fears of exposure in the face of criticising government policy and so extensive notes were taken. All

(27)

interviews were anonymised, and recorded interviews were transcribed. The transcribed

interviews were then transferred into Atlas.ti where I went through a process of open coding on each interview (Creswell, 2013). The codes of all the interviews were grouped into key themes that were used to answer sub-question two and three. Each theme and assorted quotes were then analysed in Atlas.ti to check for specific patterns. For example with sub-question two, T&G firm responses to the government strategy were analysed according to their size (mass

production vs boutique), role in the value chain (textiles vs garment production), market orientation (export vs domestic market) and customer base (expats and rich Rwandans vs poor Rwandans).

Regarding external reliability, I believe there were no threats to this as everyone interviewed worked at established firms in Rwanda and another researcher could take my script to replicate the same interviews. However, internal reliability was at risk during interviews in Kinyarwanda which I am not able to speak (Bryman 2012, p.120). I relied on my research assistant who was present during the interviews to translate my questions to the participant and then retranslate into English for me. These were recorded, then we exchanged observations and went over the translation after the interview to ensure there was consistency in our observation.

Regarding external validity, this research is not generalisable out of the specific-Rwandan context. Internal reliability was assured by reviewing the theoretical framework once the interviews had been coded and analysed.

3.2.3. Development and distribution of SHC trader survey

Due to the informal nature of the SHC industry, it was difficult to get a sense of how many SHC traders are still working today even with the tariff increases. However, I identified three markets in Kigali which still sold SHC. To understand how the tariffs have affected trade (profits,

number of customers and clothes sold) and how traders have responded to it (different sourcing strategies or looking for new employment), I developed a survey that was distributed in

Nyabugogo, Kimironko and Gisozi-Duhahriane markets (see appendix II). Before distributing

the survey, I interviewed each market leader to obtain permission to distribute the survey in their

market. 40 respondents completed the survey (see appendix I). There were more respondents in

Kimironko because traders seemed to be more willing to fill it out. There was more hesitancy and reservation observed at the other two markets. I cannot account for why this was the case.

Regarding reliability, the survey was administered in three different markets in Kigali to see if there was a high correlation between results, which there was, suggesting the measure was stable (Bryman, 2012). The survey was designed to make sure one question flowed logically to the next in order to ensure internal reliability in the respondent’s answers. As there were two observers, myself and my research assistant who were distributing the survey separately, there was a risk to inter-observer consistency, however after piloting the survey we decided to distribute the survey together and compare observation notes at the end (2012, p.22). As such, the survey was more of a ‘structured interview’ than a ‘self-completion questionnaire’ as described by Bryman (2012, p.224). The main constraint was language as a majority of the surveys were conducted in

(28)

Kinyarwanda and not English, so I was unable to participate. However, in these circumstances, I took observational notes of body language and was able to follow the line of questioning with my English version of the survey.

Regarding validity, there are legitimate questions regarding whether my measurement of (dis)embeddedness is really related to the policy decision or to other economic factors such as personal financial difficulties. In order to counter this issue and ensure face validity, the questions were extremely explicit in linking financial burden to the change in policy, most

questions were divided into ‘before’ and ‘after’ segments. For a question such as average monthly profits, the question asked for results both before and after the phase out to give a comparison. Regarding concurrent validity, as all the traders came from the same socio-economic strata in Rwanda and were all still traders in SHC, the group was largely cohesive in their response even across different markets. Furthermore, by treating the survey as a structured interview, the questions were asked in exactly the same way to all respondents in order to remove any doubt of the traders not understanding the question and thus leading to inconsistent results.

All survey participants were anonymised and assigned a respondent number before the data was inputted and analysed in SPSS. SPSS was used to generate descriptive statistics, graphs and cross-tabulation between different variables which are presented in chapter 6. The survey provides a powerful view into the effects of dis-embedding and unemployment as a result of the tariff increases on SHC traders.

3.2.4. Trade data analysis

To see the overall impact of the state’s intervention to catalyse growth of the T&G GPN and curb imports of SHC, I requested trade data from the Rwandan Revenue Authority for the last five years. The time period was chosen as it includes two years before the tariff increases on SHC (2014 and 2015), the implementation year (2016) and the two subsequent years (2017 and 2018). This provides an early indication of what was happening to trade as a result of the T&G strategy and phase out of SHC (for a detailed list of assumptions made around the data see

appendix VIII). I requested access to trade data for the HS codes 50 to 64 which covers imports,

exports and re-exports of all products related to textiles, apparel and footwear (see appendix I

for a detailed list of the data analysed). I adjusted all the trade data to 2014 constant prices in

order to account for inflation (see appendix IX for a detailed breakdown). I created a financial

model in Excel to aggregate the data by HS code to show import and export figures plus a breakdown of country of origin (for imports) and country of destination (for exports).

I was initially concerned about the reliability of the data because the trade data did not seem to show consistent measurement. The 2018 data seemed to come from a different dataset from those of 2014 to 2017. However, I finally obtained the same data for 2018 as the other years that ensured the reliability of data. For validity, the only concern was that data does not allow me to separate second-hand clothing from shoes. The HS code 630900 categorises shoes and clothing as the same, while not essential it would have been nice to show the trends in both clothing and shoes separately.

3.2.5. Policy analysis

20 policy and strategy documents were analysed to understand the industrialisation trajectory of

Rwanda and how this has informed the Textiles and Garment Strategy (see appendix I). The

documents were either given to me by MINICOM or were found on the GoR website. I also attended a two-day conference on the EAC regional strategy for leather and T&G. The

(29)

presentations in these conferences as well as the notes taken were also analysed. All documents were coded in Atlas.ti using the same technique as described in 3.1.2. As these were policy documents, they provided fascinating insight into the scale and ambition of the Rwandan state and how it conceptualises and situates itself within the T&G GPN. The very biases presented within the documentation were interesting in their own right and are explored in sub-question one (Bryman, 2012). The ambitions and self-articulated positioning of the government in view of industrialisation provides the basis of the analysis for sub-question one. The reliability was assured by verifying that the policy documentation as the final version (one document was a zero draft; this is clearly indicated in appendix I). The validity of the documentation is context specific to Rwanda and thus is not generalisable outside of this institutional context.

3.4. Methodological reflections and limitations

I have addressed the reliability and validity quality concerns of the quantitative aspects and qualitative aspects of my research in the four preceding sub-sections (3.1.2., 3.1.3., 3.1.4., 3.1.5.). I used the criteria to assess the quality of my research outlined by Bryman (2012) for both quantitative and qualitative research based on validity and reliability.

Reflecting on my research approach, I realise that I was extremely ambitious in what I was trying to achieve in the short amount of time we had to conduct primary research and to analyse the data. My mixed methods approach made my time in Rwanda extremely intense and it was only with the help of a research assistant, Sonia, and an amazing local supervisor, Edward, that I was able to complete it. The government were also extremely helpful in giving me all the information I needed. The ambition has also made my analysis phase in Amsterdam stressful as I had to develop multiple analytical approaches (financial model, coding and statistics) across multiple platforms (Excel, SPSS and Atlas.ti).

The limitations to this research are the short time period, some data limitations and language barrier:

- Short time period: If this was a longer study, I would have surveyed more SHC traders across Rwanda in many different markets to get a countrywide picture. This is a

potential avenue for further research. - Data limitations:

o Garment production data: as the garment industry is relatively new, there was no country-wide garment production data so it was difficult to assess whether there has been an increase since the implementation of the government strategy. I have tried to estimate this by looking at the import of intermediate inputs, the export of finished garments and a domestic sales figure given to me by the Bank of Rwanda. There is an ongoing study by UK Department for International Development to survey producers that will provide an overall estimate of production figures. However, I have not been given access to early drafts of the research.

o Trade data: The implementation of the T&G strategy is only three years old. While the empirical chapters show a vast decline in SHC imports, the data does not as yet show a large increase in exports of domestic clothing. As I have been told by a number of economists, to see real change in trade trends as a result of an industrialisation policy, it will take longer than just three years. This study should be replicated in five to ten years to test the longer-term effects.

Out with the old, in with the new : the role of the Rwandan state in pursuit of the development of the domestic textiles and garment (T&G) industry and its effects on the embeddedness of T&G firms and second-hand clothi

OUT WITH THE OLD,

IN WITH THE NEW

The role of the Rwandan state in pursuit of the development

of the domestic textiles and garment (T&G) industry and its

effects on the embeddedness of T&G firms and second-hand

clothing importers and traders

Charlie Robertson

Out with the old, in with the new: The role of the Rwandan

state in pursuit of the development of the domestic textiles and

garment (T&G) industry and its effects on the embeddedness of

T&G firms and second-hand clothing importers and traders

MSc International Development Studies 2018-19

Graduate School of Social Sciences

Charlie Robertson

Abstract

Acknowledgements

Table of Contents

Acronyms and Abbreviations

List of Figures and Tables

1. Introduction and context

2. Theoretical Framework

3. Methodology