The influence of tariff regimes on global production networks (GPNs)

Abstract

Despite the recognition that trade policy—in particular, tariff regimes and rules of origin—can affect the geography of production, much GPN analyses pay scant attention to the tariff context of the sector studied. This paper proposes an analytical framework to more effectively integrate these regimes into applied GPN research. We test the framework, drawing on analysis of exports to the EU market in four sectors: textiles and apparel, floriculture, fisheries and leather goods. The analysis confirms that, in the presence of high tariffs, preferences do seem to impact on sourcing for the EU market.

1. Introduction

Global production networks (GPNs) and the cognate global value chains (GVCs) frameworks have attracted much attention among academics and policymakers, seeking to understand the dynamics and impacts of global production and trade linkages (Henderson et al., 2002; Gereffi et al., 2005; UNCTAD, 2013; AfDB, 2014; Neilson et al., 2014; Coe and Yeung, 2015). The focus in much GPN and GVC research has been on how ‘lead firms’ organise, co-ordinate and govern production networks and the consequences that arise for the ability of local producers to upgrade or bring about ‘strategic coupling’ (Gereffi et al., 2005; Ponte and Sturgeon, 2014; Coe and Yeung, 2015; Yeung, 2016). Despite the importance of trade relations in these networks, the role of trade policy and its attendant tariff regimes has been subject to relatively limited study in much GPN and GVC analyses. This paper addresses this gap, firstly by bringing together research which identifies how tariff regimes can impact the structure, geography and organisation of GPNs; secondly, by providing a framework to enable future research to integrate tariff regime dynamics into GPN analysis and thirdly, by testing that framework on recent trade flows into the European Union across selected key sectors.

A country’s trade policy includes a wide range of government regulations which affect the capacity of suppliers to access the market, including standards for product safety, sanitary and phytosanitary measures, intellectual property rights and technical barriers to trade. For reasons of space (and complexity) we do not cover these aspects here. Rather, we focus on the tariff schedules applied to imports, the preferential market access provided to certain countries and the rules of origin (RoO) which regulate their application.

The relative lack of attention to tariff regimes in much existing GPN and GVC research is surprising, given that there is extensive evidence that trade restrictions (Gereffi, 1999; Curran, 2015), preferential access (Lall, 2005; Staritz and Morris, 2013; Campling, 2016) and the terms of such access (Azmeh and Nadvi, 2013; Azmeh, 2015; Curran and Nadvi, 2015; Frederick et al., 2015; Smith, 2015a; Campling, 2016) influence where the structures of production ‘land’ in the world economy. The role of the institutional context, and by extension trade policy, was noted as critical in early seminal work on GVCs (e.g. Gereffi, 1994; Ponte, 2002; Gibbon and Ponte, 2005; Bair, 2006), yet most sector-specific empirical studies in both the GPN and the GVC traditions pay little attention to tariff regimes.

We recognise the commonalities, but also key distinctions between the GPN and GVC fields (Bair, 2008; Parrilli et al., 2013). Within the GVC tradition, the primary focus is on inter-firm relationships between lead firms and their disparate and often distant suppliers, and the implications for governance and upgrading (Gereffi et al., 2005; Ponte and Sturgeon, 2014). GPN scholars, while sharing similar concerns around the political economy of power in global production, have taken a more nuanced view that recognises territoriality and spatial embeddedness, while highlighting the equal importance of non-firm actors for GPN outcomes. For example, Yeung and Coe’s (2014, 32) recent definition of a GPN as ‘an organizational arrangement comprising interconnected economic and non-economic actors coordinated by a global lead firm and producing goods or services across multiple geographic locations for worldwide markets’ specifically recognises the role of non-economic actors. Still the role of the state as a framing factor, while noted, remains under-conceptualised and, in the context of this paper, few GPN studies have sought to interrogate the influence of trade policy or tariff regimes in depth (for an exception, see Smith, 2015b).

In the most recent GPN 2.0 formulation, extra-firm networks are acknowledged as important, but the primary attention remains focused on firm characteristics—particularly cost-capability ratios and financialisation (Coe and Yeung, 2015; Yeung and Coe, 2014). We agree that firm characteristics are important. However, in concentrating on inter-firm linkages within the GPN, there is a danger that the macro context within which such linkages emerge is under-conceptualised. As Smith (2015a) and Horner (2017) have pointed out, this is partly a reflection of the limited attention paid to the role of the state in applied GPN research. In our view, trade policy is a major part of the state’s potential role in GPNs. Consequently, we highlight one key aspect of the macro context in which GPNs operate, and formulate a framework for its integration into future research, by asking the question: how do tariff regimes influence the configuration and dynamics of global production networks?

We identify two factors that might help explain the relative lack of attention to tariff regimes in recent research. First, there is a tendency in much mainstream GPN and GVC analysis to assume that liberalisation has reached a stage where trade is broadly free. Yet, the World Trade Organisation’s (WTO) most favoured nations (MFN) status still implies a level of tariff protection. In some sectors this can be significant. Second, public and private standards have, rightly, received growing attention because of their importance as ‘gate keepers’ to developed country markets (Gibbon and Ponte, 2005; Henson and Reardon, 2005; Nadvi, 2008). Indeed, Yeung and Coe (2014, 41) view standards as critical to the management of risks for lead firms in GPNs. In this context, Tran et al.’s (2013, 325) argument is indicative: ‘The relative importance of food safety standards affecting production and trade has increased because of tariff reductions resulting from World Trade Organisation (WTO) negotiations…’. While there is no question that standards are relatively more important than in the past, multilateral tariff reductions at WTO level have not advanced since the end of the Uruguay Round in 1994. This has left important tariff peaks in most countries. These peaks often remain in place in subsequent bilateral free trade agreements (FTAs) and non-reciprocal preferential trade arrangements. They reflect commercial and political interests that benefit from their maintenance, support foreign investment by businesses from preference giving states and/or prioritise particular developing countries over others. Thus, to assume that tariff regimes are no longer relevant risks obscuring important parts of the picture.

This failure to systematically include tariff regimes in GPN analysis is particularly problematic in the context of the current anti-trade political rhetoric. Trade policy debates cannot ignore the potential negative effects of the retraction of unilateral, bilateral and even multilateral trade preferences proposed by some governments (Rashish, 2017). For example, the trade diversionary effects of Brexit have been identified as having potentially profound impacts on developing country producers, including in the three sectors explored here. Systematic GPN analysis incorporating tariff regimes has an important role to play in this context (Stevens and Kennan, 2018).

In the following sections, we set out our methodology and approach, summarise existing research on a selection of sectors (drawing on both the GPN and GVC traditions) and leverage this work to propose a framework enabling researchers undertaking GPN (or GVC) analyses to rapidly identify whether the trade policy context might be pertinent to their research. We then explore EU trade flows in these sectors, highlighting the interactions between tariff regimes and trade flows. Trade patterns are found to be consistent with the assumptions within our proposed framework. In our concluding remarks, we go on to provide some proposals for future research.

2. Methodology and approach

The sector-specific nature of tariffs—and indeed RoO—makes it difficult to derive generic conclusions about their effects on the geography of GPNs. Therefore, in order to identify the influence of tariff regimes more clearly, we focus on the EU context and three well-researched sectors (textiles and clothing, cut flowers and fish) while including leather goods (bags, wallets, etc.)¹ as a control sector. In choosing these sectors we build on the work of several researchers who have argued that high MFN tariffs, combined with preferential access make it much more likely that the tariff regime will influence GPNs (Stevens, 2001; Curran, 2015; Campling, 2016). The EU charges relatively high MFN tariffs in the first three sectors and preferential access has been shown to be important to trade (Curran et al., 2007). Leather goods provide a useful contrast in that EU imports benefit from low MFN tariffs. Hence, tariff preferences are unlikely to impact leather goods trade flows (Curran et al., 2007). The sectors chosen are all relatively labour intensive, where developing country producers are highly integrated into GPNs coordinated by lead firms. They are also representative of the wider context in labour intensive manufacturing, horticulture and food. Several developing countries have leveraged preferential access to increase exports in these sectors. This has stimulated research on the structure and functioning of these production networks, thereby providing us with an existing research base for our analysis. Crucially, these are also key sectors for many low-income developing countries seeking to industrialise, generate sustainable employment and alleviate poverty through integration with GPNs. Hence, the policy implications of our comparative analysis are substantial.

The first sector—textiles and clothing—has been extensively studied and we have clear evidence that trade policy, both historically under the quota restrictions of the Multi Fibre Arrangement (MFA) and currently in a post-MFA world, impacts the geography of production (Gereffi, 1999; Lall, 2005; Azmeh and Nadvi, 2013; Staritz and Morris, 2013; Azmeh, 2015; Curran, 2015; Curran and Nadvi, 2015; Frederick et al. 2015; Smith 2015c; Pickles et al., 2016). The other two sectors—cut flowers and fish—have been subject to less extensive analysis and, critically, with a few notable exceptions (Ponte et al., 2007; Zeigler, 2007; Patel-Campillo, 2010; Campling, 2016), the existing work often ignores the trade policy context. Our comparison sector (leather goods) is little studied, but the research that does exist suggests crucial similarities with the other three sectors: it is a ‘buyer-driven’ value chain marked by labour intensive manufacturing especially in the developing world (Memedovic and Mattila, 2008; UNCTAD, 2012).

In our comparative trade analysis across the four sectors, we use a combination of trade and tariff data, interviews with key actors and prior research. We draw on extensive existing fieldwork on these sectors (including by the authors) and re-interpret it in view of our focus on tariff regimes to build our proposed framework. To test it, we analyse EU import data extracted from the International Trade Centre (ITC) database.² As trade flows can be very volatile, especially for smaller developing country suppliers, in calculating the shares of trade covered by each level of market access we use 3-year averages of EU imports between 2014 and 2016. Tariff data come from the WTO’s import tariff database.³

Our analysis also draws on interviews undertaken with trade associations and policy makers, both specifically for this project and for past research over the last 10 years by the authors. The key recent interviews (in 2016 and 2017) drawn upon here were with the Foreign Trade Association (recently renamed Amfori) representing EU importers across the sectors covered here (two interviews in December 2016 and November 2017), Euratex representing the EU textiles and clothing sector (December 2016) and a fisheries trade policy officer in the European Commission (December 2016). The interviews followed a semi-structured format, focusing on the perceived importance of tariff regimes and RoO to EU firms involved in international trade.

This paper is intended as an iterative building block in the ongoing development of GPN analysis. It is not designed as a critique. This, combined with the confines of space, means that we do not elaborate on other crucial components of the GPNs we explore. For example, a key concern of both GVC and GPN scholars is the role of inter-firm power relations (chain and network governance, respectively), questions of ‘upgrading’ or the importance of territorial embeddedness. Given that we are comparing tariff regimes across sectors, we cannot meaningfully summarise these—often highly differentiated—aspects of the GPNs studied here. It is worth noting, however, that all of the GPNs are characterised by a considerable degree of captive governance by the lead-firms that coordinate the networks (‘buyer-drivenness’ in Gereffi’s 1994 formulation), where suppliers are ‘clear price taker[s] with no or little bargaining power’ (Yeung and Coe, 2014; Table 1). In each of our cases, the highly competitive conditions of the ‘retail revolution’ dominate and lead firms are major retailers and branded-firms (Hughes 1996; Reardon et al., 2003; Coe and Hess, 2005).

Table 1

Simple average tariffs applied to imports under key EU regimes (2016)

	MFN	GSP	GSP+	EBA	EPA
Leather goods (42)	4, 5	1, 2	0	0	0
Cut flowers (0603)	8, 7	4, 6	0	0	0
Textiles and clothing
Fibres	4, 25	3, 5	0	0	0
Textiles	7, 6	6, 2	0	0	0
Clothing	11, 5	9, 2	0	0	0
Fish
Raw fish (03)	11, 1	6, 8	0	0	0
Processed fish (1604 + 5)	19, 3	11, 7	0	0	0

Source: Authors’ calculations from tariffdata.wto.org.

It is also worth highlighting that the materiality of the commodities produced in the four GPNs and their socio-political contexts are very different, with important implications for territorial embeddedness (Henderson et al., 2002). For example, on the one hand, the processing of marine fisheries is often highly embedded in particular territories, as it is shaped by access to natural resources, such that major branded-firms are often backward-integrated into processing (Campling and Havice, 2014; Havice and Campling, 2017). For clothing, on the other hand, the rapid shifts in market share that have followed prior trade liberalisation attest to the ease with which lead firms switch between suppliers and the low degree of territorial embeddedness of such firms (Azmeh and Nadvi, 2013; Azmeh, 2015; Curran, 2016). This latter dynamic is largely mirrored in the cut flower industry which, despite the dependence on a natural resource (soil), can be easily switched between locations by using modern agriculture technologies and logistics, with the result that lead firms rarely own the means of production (Zeigler, 2007; Patel-Campillo, 2011).

3. Trade policy in GPN and GVC research

There is now an extensive literature investigating the structure and evolution of networks of global production. These are largely empirical studies and Yeung and Coe (2014) rightly question the inadequate theorisation within GVC and what they term ‘GPN 1.0’ research. A key conceptual contribution to the literature was Gereffi’s identification of the four dimensions in GVC analysis: input–output structure, territoriality, chain governance and institutional context (Gereffi, 1995). The latter dimension can be defined as the context ‘that shapes the inter-firm networks that connect the various links in the chain and mediate the outcomes associated with the operation of the chain in different environments’ (Bair and Gereffi, 2003, 145). Institutional context is the dimension most closely associated with the issues covered in this paper and was a key factor in the analysis which underpins Gereffi’s early work (Gereffi, 1994, 1995, 1999), not least because his research was often focused on an industry—clothing—which, until 2005, had a complex and constraining trade regime in place—the MFA.

Notwithstanding these important contributions, it is the chain governance and upgrading dimensions that have been most extensively studied in the GVC literature. This is akin to the high levels of engagement with the concepts of network governance, embeddedness and strategic coupling in the GPN approach (Henderson et al., 2002; Hess, 2004; Yeung, 2005; Coe et al., 2008). Through their primary concentration on governance, upgrading and embeddedness, both the GPN and GVC traditions have tended to downplay the significance of trade policy. For example, Gereffi et al.’s (2005) influential paper on transaction-cost oriented re-conceptualisation of chain governance notes only in passing the ‘important impact’ of ‘trade rules’ on clothing GVCs (92). Similarly, a recent paper by Ponte and Sturgeon (2014) proposes a model of ‘multipolar’ GVC governance where trade policy is simply subsumed into ‘regulatory factors’ within a group of ‘other macro-level determinants’ (214). Recently, Smith (2015a) and Horner (2017) highlight the need for research in both the GVC and GPN traditions to look beyond the firm and better integrate the role of the state, including as a regulator of market access. Yet the most recent, and relatively firm-centric, theorisation—GPN 2.0—still underplays the state, which is folded into ‘extra-firm bargaining’ processes. However, states are not non-firm actors like any other. They are, as Yeung and Coe (2014, 51) themselves note ‘… the key regulators of uneven market access even in an interconnected world economy.’ Despite this recognition of the importance of the state’s regulatory role in structuring market access, studies of its influence in GPN literature remain rare.

An early approach to identifying the potential role of trade policy and market access regimes within the GVC framework was proposed by Stevens (2001). He defined three factors as being necessary for tariff regimes to have an impact on sourcing decisions in production networks: high tariffs on imports (which he defined as over 10%); preferential access which gives full or partial relief from these; and limitations on this access (competitors should not have similar advantages). Steven’s work was primarily focused on agriculture, which has a very specific trade policy context in the EU, with quota limitations on many imports and high tariffs. In this sector, it is clear that tariff regimes can have a strong impact on sourcing patterns and indeed most analyses of such global production structures highlight this fact (Stevens, 2001; Gibbon and Ponte, 2005; Tozanli and El Hadad-Gauthier, 2010). We build on this work below, but firstly undertake a more detailed discussion of the three sectors at the core of this paper.

3.1. Textiles and clothing

Textiles and, especially, clothing are sectors which have been heavily studied in the GVC and GPN literature (Gereffi, 1999; Nadvi and Thoburn, 2004; Palpacuer et al., 2005; Bair and Dussel Peters, 2006; Frederick et al., 2015; Pickles et al., 2015, 2016). As discussed above, much of the early work highlighted how the MFA quota system, which governed trade until 2005, helped spread production across the developing world (Gereffi, 1999; Ramaswamy and Gereffi, 2000). Specific trading arrangements for neighbouring countries, which had quota and tariff free access, also attracted research attention in the EU (Begg et al., 2003; Bair, 2006) and the USA (Bair and Gereffi, 2001; Bair, 2006). As the MFA quota system lapsed, there was much concern about potential negative impacts on poor developing countries integrated into the clothing GPN (Bair and Dussel Peters, 2006; Pickles 2006), although some researchers argued that trade policy would continue to have a strong impact on the geography of production (Abernathy et al., 2006; Bair, 2006).

Since the quota system was dismantled other issues have drawn the attention of researchers. This is unsurprising as, in a more liberal trade policy environment, new factors inevitably increase their relative influence on sourcing. The importance of rapid production turnaround to fit the requirements of ‘fast fashion’ has attracted much interest (Abernathy et al., 2006; Rossi, 2013; Zhu and Pickles, 2015). A major concern in this regard is the potential for simultaneously positive (upgrading) and negative (downgrading) impacts on workers depending on their skill level (Rossi, 2013). In common with the other sectors explored here, the issues of labour standards, compliance with private codes and international public norms in the industry has also stimulated a lot of research (Nadvi and Thoburn 2004; Palpacuer et al., 2005; Barrientos et al., 2011; Bair and Palpacuer, 2012; Rossi, 2015). Given the scandalous history of industrial accidents, especially in the clothing sector, this focus is unsurprising (Rossi, 2015). Only recently has research begun to delve into the link between tariff regimes and labour standards within clothing GPNs (Smith et al., 2018).

More importantly, for our purposes, there are now several studies that underline the importance of RoO associated with preferential trade agreements in shaping textiles and clothing GPNs (Pickles et al., 2015). RoO are complex in the sector. In the USA, preferential access is generally only accorded to clothing made up from textiles which have been spun and woven in the preference receiving country, or in the USA. In the EU, RoO generally require that the textile inputs should be woven domestically, or in the EU. As highlighted by Bair and Dussel Peters (2006) such regulations define the production structures required to secure market access. Research has found that more ‘flexible’ RoO in certain trade agreements, which allow third country fabric to be used, have stimulated trade and investment in Jordan (Azmeh and Nadvi, 2013; Azmeh, 2015), Nicaragua (Frederick et al., 2015), Lesotho (Lall, 2005), and Bangladesh (Curran and Nadvi, 2015). This research has generally concluded that countries accorded such flexible RoO have increased their clothing exports, although the extent to which this has strong developmental effects on the domestic economy and employment varies widely. Specifically, certain regimes have been found to result in very limited indigenous industrial development, with most of the rents captured by foreign investors (Azmeh and Nadvi, 2013; Azmeh, 2015), while others seem to have more positive effects, including increasing unit prices (Curran and Nadvi, 2015). In summary, despite the end of the quota system, there is clear evidence that preferential tariff regimes, and their associated RoO, continue to be an important influence on both the geography of textiles and clothing production and the potential for developing countries to upgrade.

3.2. Fish and fish products

Fish exports from developing countries generate higher export earnings than coffee, bananas, cocoa, tea, sugar and tobacco combined (FAO, 2010). Although several major import markets apply high MFN tariffs to fish products (OECD, 2003), the sector attracts surprisingly little analyses focusing explicitly on the impacts of tariff regimes. Much existing GVC and GPN analyses of the sector have investigated how private and public standards affect upgrading possibilities within the chain (Ponte, 2008; Tencati et al., 2008; Tran et al., 2013; Adhuri et al., 2016). Although tariffs may have fallen, they remain crucial determinants of trade in certain fish products. Sourcing in a sector with MFN tariffs of up to 24% and 36% (for the EU and USA, respectively, both for canned tuna) is inevitably impacted by the differential application of that tariff. Quantitative analyses of the global fish trade confirm both the importance of trade regimes to flows and the positive relationship between low labour costs and fish processing trade (Natale et al., 2015).

An early contribution to understanding the importance of trade policy to fish trade was the work of Ponte et al. (2007). This focused on the potential negative impacts on African fish exports of multi-lateral trade negotiations. As the likelihood of successful negotiations receded, researchers shifted their attention to the impacts of exacting public and private standards (Ponte, 2008; Tencati et al., 2008; Tran et al., 2013; Adhuri et al., 2016). A recent special issue of Marine Policy exploring fish value chains underlined the important impacts of regulatory intervention, especially for managing fish stocks, which can sometimes impede upgrading and value capture (Hamilton-Hart and Stringer, 2016). However, none of the papers in the issue addresses trade policy issues in any depth. Yet we know that differential tariffs influence fisheries GPNs. In earlier interviews, EU fish processors highlighted the importance of the tariff regime to the geography of production: ‘We have canneries around the world and if customs protections were removed and the 25 percent [sic: 24 percent] disappeared, the canneries would also disappear.’ (Author Interview, 2006, cited in Campling, 2012).

Recent research has highlighted how the persistent impact of tariff regimes, particularly preferential access to the EU and US markets, has shaped the geography of processed fish production (Ponte et al., 2007; Campling, 2015), especially tuna (Havice and Campling, 2013; Campling, 2016). Campling (2016) details how preferential access to the EU market for certain countries, especially in the Africa, Caribbean and Pacific (ACP) group, combined with complex RoO and ad hoc derogations, shape global manufacturing, trade and investment patterns in the sector. He concludes: ‘This is not to claim that the preference was the sole determinant in this process, but it was an integral aspect of a set of necessary conditions.’ (op. cit., 225).

However, he questions the extent to which the trade regime was designed to provide developmental benefits to the recipient countries, indicating that the key beneficiaries of much preferential access are the owners of the EU’s distant water fishing fleet (Campling, 2015). Countries with preferential access, which do not have an adequate domestic fleet, represent a captive market for these companies because of the EU’s RoO for fish products. Generally, fish must be ‘wholly obtained’ from within a given country’s 12 mile territorial seas (with a tolerance of 15%) to qualify for preferential access. Given that many species (such as tuna) live outside of these waters, different rules apply, in terms of ownership, flag and registration of the fishing boat. Developing countries have long criticised the restrictiveness of these requirements, especially on ownership, which tend to benefit EU fishing fleets (Campling, 2015). The UK government’s 2005 Commission for Africa reiterated these criticisms stating that EU RoO can be ‘… applied in a deliberately obstructive manner …’ and are ‘… taken to ludicrous extremes—to the extent that fish are ruled ineligible if the boat they are caught from is Ghanaian but the master of the vessel is South African.’ (Commission for Africa, 2005, 55–56).

The EU provides some derogations to these RoO, which have had clear impacts in certain countries. In the Cotonou Agreement (ACP-EC, 1995), derogations are provided in situations where: ‘the development of existing industries or the creation of new industries justifies them’.⁴ The most important was an ‘automatic derogation’, which allocated a total annual quota to the ACP of 8000 mt of canned tuna and 2000 mt of tuna loins. This was then distributed among the beneficiaries through negotiations.

Such agreements secure market access, but restrict new entrants. When eligible countries (or new firms within existing countries) started exporting processed tuna, the re-distribution of the quota was often conflictual (Campling, 2012). In recent negotiations with the EU on Economic Partnership Agreements (EPAs), discussions on the quota of fish covered by derogations have been among the most contentious (SmartFish, 2012). These intense efforts to secure greater market access reflect its importance to the countries in question. Research confirms that the tuna processing industry in several key exporting countries is largely a product of preferential market access, especially in certain small island economies, such as Mauritius, Papua New Guinea (PNG), Seychelles and the Solomon Islands (Ponte et al., 2007; Havice and Campling, 2013).

The case of PNG is illustrative. In the EPA negotiations between the Pacific ACP region and the EU, the Pacific secured global sourcing RoO for their processed fish exports.⁵ PNG was the first to actively benefit from this derogation in 2008 and at least two studies were undertaken on the impact (Hamilton et al., 2011; Sullivan et al., 2011). Both concluded that it was too early to draw firm conclusions but highlighted the potential for major impacts, with several new processing facilities planned in PNG. Hamilton et al. (2011, 2) concluded: ‘while not the primary driver for attracting onshore investment, the derogation will play a critical role in industry expansion in the future and its survival’. However, as with clothing, the long-term developmental impacts of this preferential access have been questioned. Sullivan et al. (2011) note poor employment conditions in the PNG canning industry, with low wages, high labour turnover and corresponding social problems.

In summary, there is significant evidence that tariff regimes influence the geography of production in the fish sector, especially in processing and in products subject to high tariffs, like tuna. Nevertheless, market access is only ever an enabling, not a sufficient condition, for the development of a local fish and processing industry. Several researchers in the sector underline the importance of local political-economic context to the interaction between trade policy interventions and trade and investment flows (Havice and Campling, 2013; Campling, 2016; Hamilton-Hart and Stringer, 2016).

3.3. Cut flowers

In comparison to clothing and fisheries, the cut flower trade takes place in a relatively straightforward context. Flowers have simple RoO—the goods must be ‘wholly obtained’. However, the industry is labour intensive and attracts relatively high applied tariffs. These are two key factors which, as discussed above, favour the geographic shift of a sector to low-cost locations with preferential market access. Although it has attracted less research than either fish or clothing, the floriculture GPN has been the subject of some analysis which informs our research.

As in the fish sector, studies on cut flowers investigate the impact of standards, although in this case labour, rather than product, standards, have tended to be the focus. Early work on African horticulture, including cut flowers, highlighted the gender bias in employment in the industry and the vulnerability of the largely female part-time employees, despite the industry being subject to a plethora of codes of conduct (Barrientos et al., 2003).

Researchers have explored the impact of the industry on socio-economic development, particularly in relation to the quality of employment provided and the extent to which basic labour standards are respected in production (Riisgaard, 2009; Riisgaard and Hammer, 2011). Gebreeyesus and Sonobe (2012) focus on the evolution of firm capabilities and the role of foreign expertise and government efforts for local capacity building. Others have explored changes in the governance of the chain, particularly the globalisation of production and the decline in the importance of the Dutch auction system (Patel-Campillo, 2011) and more recently, the impact of the global financial crisis and subsequent economic slump (Keane, 2012).

However, analyses of the trade policy context for sourcing decisions is limited. For example, Riisgaard (2009) and Gebreeyesus and Sonobe (2012) note the large increase in EU imports from Africa without referring, even in passing, to the tariff preferences that stimulated that growth. In our literature search, four studies were identified that explore the impact of tariff regimes on the cut flower trade. The first estimates the importance of EU trade preferences for cut flower exports from Colombia and Ecuador (Muhammad et al., 2010). This confirmed that a return to the standard MFN tariff of 8.5% would reduce exports—by up to 7.3% in the case of carnations from Colombia.

In the context of this paper, the three most pertinent studies are Patel-Campillo (2010), Ziegler (2007) and Keane (2017). Their analyses of Latin American and African cut flower production structures identify the importance of tariff regimes. Patel-Campillo (2010) shows how the growth of the sector in Latin America was strongly linked to preferential access to the US market. Her analysis of trade flows clearly demonstrates the impact on sourcing of reductions in the applied tariff for Colombia and Ecuador.

Ziegler (2007) explores the various US preferential trade regimes introduced in the early 1990s to benefit cut flower exports from Latin America and the Caribbean. These schemes were only rolled out after the US cut flower industry spent many years lobbying against imported flowers. However, by 2001 several preferential regimes were in place and 85% of flowers entered the USA duty free (DF) (Ziegler, 2007, 67). Patel-Campillo and Ziegler both highlight that the cut flower sector requires quite complex logistics infrastructure—not least rapid and refrigerated transportation. Given the cost of such infrastructure, one would expect high MFN tariffs to be a necessary condition for tariff regimes to impact on the competitiveness of the landed product.

Finally, Keane (2017) highlighted the importance of preferential access to the EU market to the development of the cut flower sector in Kenya (a non-LDC) and Ethiopia (an LDC). She also underlined how the differing status of the two countries impacts on the security of that access. ‘The trade preference rent made available to Ethiopia was perceived as more secure compared to Kenya over the period 2007 and 2014 (because Ethiopia is an LDC, whereas Kenya is not).’ (2017, 105). Her research found that the fact that Kenya risked moving to a less advantageous EU tariff regime had encouraged some Kenyan cut flower producers to relocate some of their production to Ethiopia.

4. Proposing a structured approach to identifying tariff regime effects

Below, we propose a framework incorporating the key factors which emerge from the literature as important to defining whether tariff regimes should be included in a GPN analysis of a given sector targeting a given market. The objective is to provide a basis for researchers to identify those sectors where tariff regimes are likely to be a vital element of the macro-institutional environment impacting firms and thus require systematic inclusion in their research. In the schema in Figure 1, these would be the sectors in white boxes. The product level at which such analyses should be undertaken depends on the sector. In trade data, products are specified using Harmonised System (HS) codes. These vary widely across sectors in their level of detail, while the variation in tariffs across codes is also very heterogeneous.⁶ In clothing, for example, HS2 is adequate (HS61 and 62), as the tariffs and RoO are essentially the same across that level. In fish, HS4 level is usually adequate, although there are some important variations in tariffs at HS6 level. Cut flowers have very similar tariffs at the HS6 level, so HS4 is adequate. In order to decide at what level their analysis should be focused, researchers need to identify the level of variability in the tariffs applied on a given market across their sector of interest through analysis of the WTO tariff schedules, or from the trading partners’ own data sources.

Figure 1

Schematic framework for identifying tariff regime effects in GPNs.

Having defined the product level of the analysis, the first issue of importance is whether MFN tariffs are applied by the importing country. Clearly if they are not, then there is no preferential access to provide and tariffs should not impact on procurement and investment decisions in GPNs. However, the absence of MFN tariffs does not mean that trade policy will not impact on a given sector. Recent analysis in the solar panel sector highlighted that, even when a sector has low, or zero, applied tariff protection, if it is sensitive in terms of a country’s industrial or employment strategy, it may be vulnerable to ad hoc trade policy interventions like anti-dumping duties (Curran, 2015). These have the same effect as differential tariffs, deflecting trade toward unaffected suppliers. Thus, the evolving political sensitivity of the sector and its strategic nature should also be taken into account.

In cases where MFN tariffs are applied, their effect will obviously depend on their level. Low, so-called ‘nuisance’ tariffs will have less impact than the tariffs of 5–24% applied in the key sectors we explore here. The higher the tariff, the greater the ‘preference margin’ provided to preferential suppliers, the greater their cost advantage compared with non-preferential suppliers and the higher the expected impact on the geography of production. There is no consensus on the level at which a tariff is significant enough to affect sourcing choices. For example, ACP Parliamentarians have argued that tariff preferences which are greater than 3% provide a competitive advantage (Campling, 2012). Davenport et al. (1995, 67) defined a ‘significant’ margin of preference as over 5%, while others consider this margin to be ‘trivial’ (McQueen et al., 1998, 40). Stevens (2001, 50) states that: ‘tariffs of 10 per cent or more … offer the possibility of a potent protectionism preference combination’. Work in cut flowers supports the view that even ‘low’ (less than 10%) tariff preference margins have real and measurable impacts on sourcing decisions (Ziegler, 2007; Patel-Campillo, 2010; Keane, 2017). In reality, the commercial value of a preference can only be assessed on a product-by-product basis and even then, quality differentials between products on the same tariff line may impact the relative benefit of a tariff advantage (with a tariff having proportionately greater impacts on more expensive products).

Where MFN tariffs are considered significant, their actual influence on sourcing and/or investment decisions will depend on whether there are exceptions to the applied tariff, what countries are subject to these preferences and their industrial capacities. For preferences to translate into investments, orders and trade flows, the supplier needs to have the capacity to produce the good in question competitively. All of these points speak to the particular type of analysis and method associated with the GPN approach: i.e. a tariff preference on paper does not translate directly into commercial advantage, it must be assessed in articulation with real-world GPN dynamics and practices. It is therefore necessary to integrate such trade policy factors with the micro level determinants highlighted in GPN 2.0 (Yeung and Coe, 2014), especially the cost-capability ratios of different suppliers in the light of buyer requirements.

To benefit from preferences, a supplier firm needs the capacity to produce in conformity with the RoO required. If the RoO are straightforward their potential impacts on the geography of production are likely to be limited, or at least easily understood. Sectors with more complex RoO, like clothing and processed fish, often require very specific production processes and/or ownership structures to benefit from preferences. These RoO tend to be more onerous for companies to fulfill. Furthermore, in cases where the preference giving country provides derogations from these RoO for certain countries, like Bangladesh in clothing and PNG in canned tuna for the EU market, this provides important competitive advantages and we observe clear and rapid impacts on the geography of these GPNs (Curran and Nadvi, 2015; Campling, 2016). It is vital that the tariff regime is factored into analysis of such contexts.

5. Exploring trade flows across regimes in the EU

The above schema proposes that certain aspects of trade regimes are likely to affect the geography of production. This section investigates the situation in the EU market in our chosen sectors in terms of the varied tariffs applied to imports and explores how these tariffs influence trade. The objective is to test the extent to which the relationship between tariffs and trade flows proposed in our framework above is supported. Trade flows are by definition, a macro indicator. We cannot ascribe these flows to individual firms or explore the decision-making processes behind them. However, we would argue that these macro-level flows are the result of many individual micro-level decisions between and within firms, which are partly driven by outside context, including trade policy.

5.1. EU tariff regimes and differential access

In seeking to investigate the interaction between EU trade and tariff structures, it is important to explain some basic aspects of EU trade policy. The tariffs applied to EU imports vary depending on many factors, not least the domestic political economy of the sector in the various member states and the extent to which importers and domestic producers have lobbied for or against tariff reductions. As in many markets, higher tariffs tend to be applied to more processed goods—so called tariff escalation—a practice which has long been criticised by NGOs, international organisations and beneficiary countries (Oxfam, 2005; ICO, 2011). However, tariff escalation can work in favour of those developing countries which benefit from preferential access, by making processed goods from third country competitors significantly more expensive in the market.

Preferential access to the EU market varies widely. We focus here on the key unilateral preferential tariff regime provided by the EU—the Generalised System of Preferences (GSP). The GSP is divided into three schemes—the standard GSP, GSP+ and the Everything But Arms (EBA) initiative for Least Developed Countries (LDCs).⁷ The standard GSP provides a reduction in tariffs for poorer developing countries (those classed by the UN as, at best, lower-middle income), for products where the country is not considered so competitive that they no longer need preferences to access the EU market.⁸ GSP+ is available to ‘vulnerable’ countries which have implemented international conventions on human and labour rights, the environment and good governance. EBA is accorded to all LDCs, the list of which is established annually by the UN.

Thus, eligibility for both GSP and EBA is dependent on a certain (low) level of development, while access to GSP+, although related to development level, is also contingent on adoption of international conventions. In addition to the GSP, many countries have bilateral FTAs with the EU. EPAs are a specific group of FTAs negotiated to replace the previous unilateral preferences for ACP countries. In most cases, the sectors discussed here are covered by the EU’s FTAs and thus imports from FTA partners enter the EU DF. The exception is fish, which is subject to complex tariff rate quota systems. This is the case even within the Economic Area Agreements with Norway and Iceland, described by one EU official as ‘a nightmare’ in their complexity (Author interview, December 2016). For this reason, FTAs are not included in the table below.

The result of these different market access regimes and differential sensitivity between sectors is a variable EU tariff profile within and across sectors, with subsequent differential impacts on the sector’s geography of production. Table 1 provides overall EU applied tariff data for the three key sectors explored here and our control sector—leather goods. The tariff profiles of the textiles and clothing, leather goods and cut flower sectors are fairly straightforward. The latter is subject to limited processing, thus tariff escalation is virtually nonexistent. Tariff levels in leatherware vary. All products in the sector are processed, however there are variations across the different products. Most products attract low ‘nuisance’ tariffs in the 1.7–4% range, although a few are subject to more significant tariff peaks of 9.7%. All products are DF under the EBA and GSP+ schemes, although a few still attract (low) tariffs under GSP. In the textiles sector, fibres and yarn attract lower tariffs than made up textiles, which are lower than clothing, but within each level of processing, tariffs are very similar.

Fish is far more complicated, with some unprocessed fish of interest to EU fishing fleets—like mackerel—attracting very high tariffs (20%). Tuna imported for processing is DF, while the canned product attracts a tariff of 24% (a classic policy of tariff escalation to protect EU-based processors). Thus, while Table 1 provides the simple average tariffs for fresh or frozen raw fish and fillets (HS 03) and processed fish (HS1604+5), within each grouping there is a lot more variation—and tariff peaks—than in the other three sectors, complicating analysis of the sector.

From the table, it is immediately obvious that the highest tariff preferences are provided to GSP+, EPA and EBA suppliers and the sectors where these regimes provide the greatest advantage are clothing and fish, especially processed fish. It is in these sectors that we would expect the greatest tariff regime effects. It is also clear that the reduction in duty provided under GSP is not very significant in sensitive sectors (all our sectors except leather goods).

6. Trade trends by trade regime

We now explore actual trade flows across different EU trade regimes over 2014–16. Detailed tariff line and country level analysis are beyond the scope of this paper. Our objective is simply to gauge the extent to which DF or preferential access influences trade flows across the different sectors. We extracted EU import figures by exporting country and then defined each as trading under either MFN, reduced duty (RD)—usually GSP—or DF. On this basis, we calculated the percentage of EU imports in each sector subject to each regime. We also include figures on the structure of total goods trade for comparison.⁹ The results are shown in Table 2, where the average MFN tariffs are also recalled. To indicate the geography of each GPN, we detail the top 10 suppliers for the period in Table 3.

Table 2

Trade profile across sectors and trade regimes in the EU

Share of EU imports under DF, reduced tariff and MFN regimes—average trade in $ (2014–2016)
	Total trade	Leather goods (4.5%)	Fibres and yarns (4.3%)	Textiles (7.6%)	Clothing (11.5%)	Cut flowers (8.7%)	Raw fish (11.1%)	Processed fish (19.3%)
Duty free	27.9	12.3	36.5	47.0	46.1	97.8	21.9	54.2
Reduced duty	8.8	19.5	5.3	2.1	13.9	0.5	53.7	35.0
MFN	63.3	68.2	58.2	50.8	39.9	1.7	24.4	10.7

Source: Authors’ calculations from ITC trademap.

Table 3

Top 10 EU suppliers by trade value in $M (average 2014–2016)

^aThailand’s fish did not benefit from GSP over this period, it is however, consistently the leading beneficiary of an EU tariff-free quota for processed tuna.

Defining the categorisation of each supplier was straightforward in all sectors, except fish. The EU provides essentially tariff-free access in the other sectors to FTA partners. However in fish several such partners, and even some overseas territories, still pay reduced tariffs on a number of tariff lines.¹⁰ Although a certain percentage of their trade is dutyfree, the political economy of the sector would suggest that the most protected sectors are most probably those of greatest export interest to their suppliers. These flows are therefore included in the RD category.

As Table 3 indicates, the importance of suppliers who enjoy DF market access varies extensively across sectors. It is most important in cut flowers, where 9 of the top 10 suppliers to the EU operate under a duty-free regime. Moreover, the top four suppliers (Kenya, Ecuador, Ethiopia and Columbia) account for 84.5% of total EU imports. In the highly perishable flowers sector, proximity to market also matters (key for Kenya and Ethiopia), although effective logistics (especially air freight) is also critical, allowing more distant Ecuador and Columbia to account for 31.5% of total EU imports. The next highest DF levels are seen in processed fish where MFN tariffs are relatively high and again African and South American duty-free suppliers are important. The relatively low levels of duty-free trade in raw fish are surprising given the high tariffs applied, however, Norway and Iceland, whose complex market access agreements with the EU defy easy classification, account for over 40% of these imports.

The textiles and clothing sector is more varied. MFN tariffs for fibres and yarns are relatively low (under 5%) and only 2 of the top 10 suppliers have DF access. In textiles, where average MFN tariffs are higher (at 7.6%) 4 of the top 10 suppliers (most notably Turkey and Pakistan) operate DF, accounting for 44% of imports. Yet, 5 of the top 10 (48% of total trade) operate under MFN. In clothing, however, where MFN duties are highest, Asian and North African duty-free suppliers dominate and only 1 of the top 10 suppliers (China) operates under MFN. In leather goods, where MFN tariffs are low, China, an MFN supplier accounts for 63% of imports.¹¹ This suggests that tariff preferences of below 5% may not be significant enough to influence sourcing decisions by lead firms and traders especially if low-cost non-preferential competitors are available.

Overall, this brief overview of the headline figures from our chosen sectors indicate that those which have the highest MFN tariffs also have higher levels of duty-free trade than seen in total trade and far higher than for leather goods. This observation supports the assumptions within our framework on the likelihood of tariff regimes influencing certain GPNs. While correlation does not always imply causation, we would argue that this data, together with the extensive evidence from sectoral studies discussed above, provide strong support to our central point that the higher the MFN tariffs applied, the more likely it is that macro level tariff regimes will impact on the micro- and meso-level relations within and among firms which decide the geography of GPNs. The relationship between tariffs and share of duty-free trade is not linear. Cut flowers attract lower average tariffs than preserved fish, but preferential access seems to be as (if not more) important a determinant of trade in the former sector. Thus, other factors are clearly at play. As our discussion on Table 3 underlines, tariffs need to be looked at in the context of other exogenous explanatory factors, like geography and logistics (which clearly affect trade in perishable products like flowers) and resource endowments (which clearly affect trade in fish and fish products). These interactions are complex, underlining the necessity of comprehensive GPN analyses. In addition, there are large differences in MFN tariffs within one of our sectors—fish—which call for more detailed study to unpick the relative importance of product and species-specific tariff regimes across territories accorded different levels of access.

7. Conclusion

Despite the recognition that trade policy can impact on global production, recent studies on GPNs (and GVCs) rarely systematically interrogate the role of trade policy in general, and tariff regimes in particular. We address this critical gap by providing a framework for researchers embarking on sectoral analyses to ensure that those contexts where tariff regimes are likely to impact on GPNs are identified. In the light of our analysis above, and drawing further on our interview data, in this section we provide some conclusions on the impact of tariff regimes on GPNs and suggestions for further work.

We acknowledge that tariff regimes are never the sole defining factor in the decision making of lead firms and other GPN actors. In most of the sectors we explore, there are major suppliers to the EU market whose exports pay MFN tariffs. Thus, even in sectors where we expect a high level of sensitivity to tariff preferences, some EU buyers still prefer to use non-preferential suppliers. This is clearly a reflection of the importance of cost–capacity ratios, highlighted in GPN 2.0. In addition, those suppliers that are subject to preferences do not always exploit them (Curran et al., 2007). This happens for a variety of reasons, including the constraints of RoO identified above. Furthermore, other exogenous factors, like national regulation, political economy and geography clearly interact with tariffs in all of the sectors we explore. Their relative importance will vary across sectors and end markets, underlining the importance of comprehensive GPN analyses. Therefore, our conclusions on the importance of tariff regimes need to be nuanced by the evidence that they are not always a defining factor, even where theoretically they could be a key motivator.

With this caveat in mind, we argue that existing analyses and our own data support the view that preferential market access which gives substantial reductions in applied tariffs provides an important cost advantage to certain countries’ suppliers. This can have an important impact on the geography of production and sourcing. Of course, much depends on the level of the MFN tariff applied. We propose that a minimum of 5% could be used as a rough benchmark, although the price sensitivity of the sector and the relative competitiveness of key non-preferential suppliers will also intervene.

We argue, therefore, that before embarking on the firm-level analyses which is a core element of a comprehensive GPN study, researchers need to pay greater attention to the macro context framing their target GPN and, in particular, consider the ‘fine print’ of the tariff regimes applied by the importing country. In so doing, they can correctly identify both those sectors subject to tariff regime affects and the likely nature of these effects. Our hope is that this approach will enable tariff regimes to emerge as a vector of research in sectors beyond those discussed here, where they have not yet been extensively explored. Of course, our proposed approach will also enable researchers to eliminate them from their analysis, where they have little impact, but on a sound and systematic basis.

Even in today’s relatively open world economy, trade liberalisation is not universal. This has always been the case and is likely to be even more so if political demands for greater protectionism gain traction. Trade liberalisation and preferential access schemes are increasingly controversial in the developed world. A decade ago, the concerns of researchers on trade policy change focused on the negative impact of multilateral liberalisation on preference dependent countries—so called preference erosion (Curran et al., 2007; Ponte et al., 2007). Future concerns are more likely to focus on the impact of unilateral reductions in preferential access and the threat of new tariffs in the US market (VanGrasstek, 2016; Rashish, 2017), post-Brexit trade policy changes in the UK (Razzaque and Vickers, 2016; Curran, 2018) and trade preferences in Southern markets with the rise of ‘polycentric’ trade (Horner and Nadvi, 2018).

GPN research has an important role to play in identifying sectors and countries vulnerable to negative effects from such change (Stevens and Kennan, 2018). Our framework can contribute to this work by helping to orient researchers toward those sectors where tariff regime impacts are most significant. There is substantial evidence, explored above, that preferential market access can stimulate industrial development in certain circumstances, although as we also highlight, increased trade flows do not always contribute significantly to sustainable local development and this link certainly requires further analysis. Nevertheless, the removal of preferential access could have major negative effects, potentially undermining poverty reduction and industrial upgrading. More micro-level research is needed to deconstruct the effect of trade preferences at worker, labour regime, firm and country levels. Whichever direction trade policy takes in the coming years, it is certain that impacts will be felt along many GPNs and thus careful sector-level research on differential trade preferences and their impacts are needed to inform decision-making. This also has a clear bearing on the ways in which the policy community, at the international and national scales, integrates tariff regimes into GPN level industrial policy and sector strategies.

Acknowledgements

We would like to thank the editors of this Special Issue, as well as the three anonymous reviewers of this manuscript, for their helpful and insightful inputs, which certainly improved the paper. We would also like to thank Adrian Smith, Rory Horner and Martin Hess, and members of the GPN, Labour and Trade research group at the University of Manchester, who provided comments on an earlier version of this paper, as well as all of the participants at the GPN conference in Singapore in December 2017 for their helpful feedback. Any remaining errors are our own.

Funding

Khalid Nadvi would like to acknowledge funding received from the ESRC through the project on 'Rising Powers, Labour Standards and the Governance of Global Production Networks' (grant number ES/J013234/1).

Footnotes

1 Leather goods (HS42) does not include footwear.

2 http://www.trademap.org/Index.aspx.

3 http://tariffdata.wto.org.

4 CPA, Annex X, Protocol 1, Title V, Art. 38:1.

5 Technically ‘global sourcing’ involves the acceptance that a change in tariff heading—in this case from HS03 to HS16—is adequate to confer origin.

6 All traded goods need to be clearly identified when entering or leaving a customs territory. The HS is an international system of codes that ensures that all countries define goods in the same way. Goods may be defined by up to 8 (HS8) or even 12 (HS12) digits. These codes are also used to define the tariffs applied to each good and variations across a given level vary substantially by sector.

7 For full details on these schemes and how they differ see the European Commission website http://ec.europa.eu/trade/policy/countries-and-regions/development/generalised-scheme-of-preferences/ and Beretta (2013).

8 Countries whose exporters represent more than a given threshold of the EU’s imports are ‘graduated’ from preferences in that sector.

9 We excluded the oil and gas sector (HS 27) from total trade for two reasons. Firstly, oil prices are very volatile which tends to distort overall trade flows. Secondly, the sector faces very low, or zero tariffs in the EU, thus tariff regimes have no significant impact on trade.

10 For simplicity, overseas territories like Greenland and the Faroe Islands were excluded from the import figures and all other partner countries where some fish exports pay tariffs (Norway, Serbia, Chile, etc.) were included in the ‘reduced duty’ category.

11 Although, in theory, China is eligible for GSP, in practice, most of its sectors are ‘graduated’ because of its high share of sectoral trade. Thus, Chinese exporters to the EU tend to face MFN tariffs, including in raw fish, textiles and leatherware.