Supplier Strategies and Routines for Capability Development: Implications for Upgrading

This paper examines the strategies and routines adopted by small and medium-sized suppliers to develop capabilities that enable them to engage in upgrading, despite a precarious relational and institutional context. To this end, we investigate the strategic behaviour of two Bangladeshi garment manufacturers. Both started out as small suppliers for multinational enterprises (MNEs) and have eventually grown into micro-multinationals. The ﬁ rms are involved in ‘ tacit promissory contracting ’ with their buyers, a speci ﬁ c form of international outsourcing relationship. The study adopts a multiple case study design that involves interviews with managers/owners of the ﬁ rms. The analysis yields two key ﬁ ndings. Both ﬁ rms have devised strategies and taken coherent routines involving actions to develop skills and motivation needed to perform appropriate functional activities (i.e. pre-production, production and post-production) as they embarked on di ﬀ erent stages of upgrading. Furthermore, ﬁ rms have designed routines to internalise the challenges originating from their relationships with their buyers and the institutional environment at the time that had the potential to a ﬀ ect their upgrading goals. The paper contributes to IB studies by highlighting how suppliers, even in a precarious context, can control their own strategies and routines, so as to develop capabilities that allow them to gradually redress the power imbalance between themselves and their buyers.


Introduction
The present study sets out to examine how two Bangladeshi garment manufacturers built capabilities to progress from one stage of upgrading to another despite a challenging institutional environment. We define capabilities as the "resources and ability needed to generate and manage technical change, including skills, knowledge and experience, and institutional structure and linkages" (Bell and Pavitt, 1997:89). The concept of upgrading is defined as "a process of improving the ability of a firm to move to a more profitable and/or technologically sophisticated capital and skill-intensive economic niche" (Gereffi, 1999:38). The findings contribute to the literature on micro foundations of routines and capabilities by focusing on the question of what organisations can do to create and change capabilities (cf. Felin et al., 2012;Felin et al., 2015;Foss, 2009). Our study also contributes to research on the absorptive capacity of emerging market firms. More specifically, to the emerging theme of barriers to absorptive capacity and what firms can do to overcome these (Cuervo-Cazurra and Rui, 2017). Last but not least, our study contributes to integration efforts between the international business (IB) and global value chain (GVC) analysis literatures .
While the term 'global value chain' is increasingly utilised in the IB literature, there is a steadily growing field dedicated to the analysis of GVCs based on the pioneering work of a multidisciplinary group of scholars (Sturgeon, 2009). The GVC analysis literature is mostly focused on externalisation and inter-firm relationships through the lenses of governance and upgrading (Lee and Gereffi, 2015). It provides a very useful entry point for the investigation of knowledge flows and innovation dynamics that take place outside of the hierarchical boundaries of multinational enterprises (MNE) (De Marchi et al., 2014). To this end, we draw on the economic upgrading concept from the GVC analysis literature (Humphrey and Schmitz, 2002) to investigate the development and modification of capabilities in emerging market supplier firms over time. Bilgili et al. (2016) divide emerging markets into three categories based on their degree of institutional and factor market development. Bangladesh belongs to the category of traditional emerging economies because of its low levels on both dimensions (Bilgili et al., 2016). While studies on mid-range emerging economies and newly developed economies are steadily increasing, indepth case studies in traditional emerging economies are rare due to access difficulties. The present paper provides detailed case studies on two large garment manufacturing companies that successfully passed through all stages of functional upgrading and eventually developed into emerging market multinationals (EMNE). Generally, firms in traditional emerging markets have been found to adopt a short term focus and avoid more risky strategies requiring considerable investment (cf. Awate et al., 2012). The volatile and unpredictable institutional environment and underdeveloped factor markets tend to hamper firms' attempts to improve their capabilities (cf. Hitt et al., 2005). As a consequence, Bilgili et al. (2016) propose that firms from traditional emerging markets are more likely to engage in duplicative imitation learning as opposed to creative imitation and innovation. They furthermore propose that the main focus of firms from this type of emerging markets is on knowledge from external sources and on objective knowledge rather than on internal sources of knowledge resulting in only small increases in absorptive capacity.
However, our case firms succeeded in building innovative capabilities over time despite the precarious institutional environment and underdeveloped factor markets. This makes them a valuable study object to examine what barriers to absorptive capacity they faced and how they addressed these barriers. Lastly, most studies on absorptive capacity and/or catch-up strategies have been conducted in the context of high-technology and knowledge intensive industries (e.g. Awate et al., 2012;Chuang and Hobday, 2013;Cuervo-Cazurra and Rui, 2017;Saranga et al., 2018). In contrast, the Bangladeshi garment industry is labour intensive, characterised by high power-asymmetries and a low willingness of MNE buyers to transfer knowledge (cf. Hoque et al., 2016). The high power asymmetry stems from buyers' preferred choice of tacit promissory contracting defined as "a form of outsourcing relationship whereby supplier firms are involved in recurrent discrete transactions with the same buyers since their inception or at least for a longer period of time, but without the existence of any original legally binding written agreement" (Hoque et al., 2016:258).
The paper is structured as follows. The next section provides an overview of the literature against which the study is framed. This is followed by the methodology section, including a discussion of the conceptual building blocks that make up our analytical framework. The results section gives an in-depth account of our analysis, and outlines the main findings. Finally, the paper concludes with a discussion of the results and highlights of the contributions and limitations of the study.

Literature review
As set out in the introduction, this paper aims to investigate the process of capability development, by examining the functional strategies and routines deployed at each stage of upgrading. Routines and capabilities form one of the key constructs in strategic management research. A central argument in strategic management is that routines and capabilities should be seen as fundamental units of analysis, and that organisations should be conceptualised as repositories of routines and capabilities (e.g. Kogut and Zander, 1992;Nelson and Winter, 2002). In line with this thinking, routines are often analysed in the strategic management literature at an organisational level. However, recent interest in this area of research suggests new insights into routines and capabilities could be generated by the combination of analytical approaches and methods that focus on breaking the organisational entities into parts and mapping their interrelationships across time and levels of analysis (Salvato and Rerup, 2011). Nevertheless, only a few studies have examined the functional routines that constitute organisational capabilities (e.g. Peng et al., 2008;Sako, 2004).
Within the IB landscape, however, the process of capability development has remained at the periphery of key theoretical perspectives such as Hymer (1972), Buckley and Casson (1976), Williamson (1973), Dunning (1979), and Johanson and Vahlne (1977). Even though it can be argued that strategies, routines, and capabilities are key to creating firm-specific advantages in Dunning's eclectic framework, and also for building co-ordination and control skills in Buckley's global factory, the behavioural processes underlying the development of firm-level capabilities are underemphasised (Teece, 2014). Buckley contends that behavioural perspectives are hard to integrate into IB research as they do not follow the "rational choice axioms" (Buckley, 2015:5). As a result, the concept has mostly been applied in IB studies to examine the relationship between capabilities and the performance of MNEs in their internationalisation activities (e.g. Michailova and Zhan, 2015;Pinho and Prange, 2016;Prange and Verdier, 2011).
Thus, the focus has predominantly been on the benefits of possessing capabilities, rather than on the process of how they are developed (Foss, 2009;Zhan and Luo, 2008) or on how firms acquire, deploy, and upgrade capabilities at different functional levels (cf. Luo, 2000;Teece, 2014). Therefore, while the concept of capabilities is not new in IB, there is still need for explorations in terms of how they are developed and/or augmented through micro-processes (cf. Prashantham and Floyd, 2012). Furthermore, there are calls for the integration of more micro-foundations research into IB (Verbeke and Calma, 2017). While a few recent IB studies have examined this process in the case of MNEs (e.g. Michailova and Zhan, 2015;Prashantham and Floyd, 2012), the micro-processes involved in the development of capabilities are even less systematically studied from the perspective of smaller suppliers from developing countries, which connect to the global economy through GVCs (Liu and Zhang, 2014). The relationship between developing country suppliers and their multinational buyers within a GVC is often characterised by an extensive power and knowledge asymmetry that affects the strategies and routines which these firms adopt to develop capabilities (cf. Bradley et al., 2006;Hoque et al., 2016).
Firm capabilities have received much more attention in the GVC literature. However, this body of literature is often criticised for its limited focus on firm-level strategic behaviour (Starosta, 2010). GVC analysis embraces the capabilities of the supplier base as one of the key factors shaping the mode of governance that MNEs adopt (Gereffi et al., 2005). However, while GVC studies consider the role of supplier capabilities within the context of governance, there is less emphasis on the ways capabilities are developed (Mahutga, 2012). The existence of a certain degree of supplier capability is more or less taken for granted, and firm-level efforts to develop new and/or improve capabilities are largely underexplored (Kawakami, 2011). Overall, the firm tends to be treated as a black box in GVC analysis (Starosta, 2010), along with its intentional, strategic actions. To this end, Kawakami (2011) argues that suppliers' strategies, and the governance mode of the lead firm, jointly mould the development of suppliers' capabilities. In line with this thinking, Fujita (2011) finds that Vietnamese motorcycle suppliers have developed capabilities in different value chain functions by adopting firmlevel strategies that were also influenced by their relationships with Japanese or local assemblers. Therefore, a gradually increasing number of GVC studies conclude that the development of supplier capabilities may be a function of their firm-level strategic behaviour and their relational setting with their buyers (e.g. Fujita, 2011;Kawakami, 2011;Kawakami and Sturgeon, 2011). Nevertheless, further research is needed to achieve a deeper understanding of how suppliers in different relational settings strategise and set up appropriate routines for building capabilities within GVCs, and how those capabilities feed into the process of step-by-step upgrading.
Furthermore, previous GVC studies infer that suppliers build capabilities in order to move from simpler to higher-value-added functions (e.g. Kadarusman and Nadvi, 2012;Morrison et al., 2008). Thus, capability dynamics in the GVC literature are generally considered in relation to upgrading (De Marchi et al., 2014). However, the interplay between these two concepts is under-researched (Kawakami and Sturgeon, 2011). Capabilities and upgrading are discrete concepts, originating from different disciplinary areas, yet the terms are often used interchangeably within GVC analyses (e.g. Isakseen and Kalsaas, 2009;Ivarsson and Alvstam, 2011). Therefore, while there is a clear indication of the link between capabilities and upgrading in GVC studies, the strategic route from the formation of capabilities to the achievement of different forms of upgrading is still not entirely clear.
In sum, the literature on GVC upgrading points to the dynamic nature of the fragmentation of economic activities in which supplier capabilities play a key role. Based on the above arguments, the present paper adopts the view that supplier strategies and routine micro-processes are geared towards the development of a specific set of capabilities that can influence their upgrading options within the GVC. The suppliers' strategies and routines are in turn affected by the nature of the relationships they have with their buyers (Hoque et al., 2016), as well as the local context (Kawakami and Sturgeon, 2011). The rest of the paper is devoted to the empirical investigation of these relationships.

Methodology
We adopted an exploratory multiple case study approach, with nested template analysis Sinkovics et al., 2014), to carry out the research. Table 1 provides an overview of the two case companies. We conducted interviews in two rounds. In 2014, we interviewed eight managers from different divisions. Interviewees were asked about the history of their firm, critical incidents, the opportunities and challenges their business faced at the time, and the future plans of the organisation. The interview guide also included questions about their relationships with their buyers (such as the existence of a contract, length of relationship, process of ordering, materialising and finishing a transaction, exchange of knowledge and information, modes of contacting buyers, and critical shapers of repetitive relationships). In 2017, follow-up interviews were conducted for the purpose of gaining more specific insights into the processes they used to develop the capabilities they needed to support upgrading, and the functional routines they deployed to do so. The interviews were complemented with corporate presentations by divisional managers, factory visits, and participant observation aimed at better understanding the functional routines. Finally, we collected organisational documents to gain insights into their upgrading history. We analysed the data using a flexible pattern-matching technique (Table 2 and Table 3). The initial template builds on the four conceptual dimensions, and we will elaborate on it in the following sections.

Conceptual building blocks
Our exploratory case study design with nested template analysis required the a priori definition of the main theoretical concepts. Our analytical framework for the data analysis rests upon the following five concepts: i) GVC upgrading; ii) supplier strategy; iii) capability building; iv) supporting routines; and v) barriers to absorptive capacity. The definitions and frames of these concepts are discussed below:

Strategy
In this paper, we adopt Rumelt's (2011:6) definition of strategy as "a coherent set of analysis, concepts, policies, arguments and actions that respond to a high-stakes challenge". Three functions related to strategy formulation and implementation are considered in our initial template: prescient diagnosis, a guiding policy, and coherent action. A good strategy must include these three functions, which, when combined, are called the "kernel of strategy" (Rumelt, 2011). The diagnosis refers to the identification and assessment of challenges or opportunities. The guiding policy denotes the plans undertaken at the corporate level in order to implement processes and organise resources. Actions stand for activities to be undertaken at the functional level in order to achieve an overall policy/plan. This involves the implementation of changes in R&D, marketing, production, procurement, and human resource activities (Rumelt, 2011).

Capabilities
Fujita (2011) proposes a capability matrix for analysing the depth and width of suppliers' capabilities, and the change in capability development over time. We adopt these dimensions for our analysis. Capabilities are defined as the "resources needed to generate and manage technical change, including skills, knowledge and experience, and institutional structure and linkages" (Bell and Pavitt, 1997:89). The framework considers two dimensions: functions and levels. The functional dimension consists of four categories of capabilities (Fujita, 2011: 71): • Planning capability (pre-production): This category includes market research, product concept development, new product design, and development capabilities.
• Equipment/raw-material-related capability (pre-production): This group encompasses capabilities related to the operation, maintenance, design, and manufacture of equipment, dies, moulds, jigs, and tools.
• Production management capability (production): This set refers to the organisation and management of the production process. • Sales and marketing capability (post-production): This category embraces marketing, branding, sales, and customer-relationshipbuilding capabilities.        Fujita (2011) argues that suppliers typically rely on mature, standardised technology imported from advanced economies, while adopting more advanced technology over time. Thus, firm-level efforts are necessary for developing the capabilities needed to master the imported technology in the first place. Over time, suppliers may develop capabilities that allow them to adapt the technology to their needs in the domestic market, and finally create new technologies of their own. Subsequently, the framework looks at four levels of depth: • The operational level refers to the extent of suppliers' ability to work with the existing technology.
• At the assimilative level, suppliers have mastered the existing technology and developed the ability to maintain operations over time.
• At the adaptive level, suppliers have developed the ability to make modifications and amendments to the existing technology.
• At the innovative level, suppliers are able to develop original and novel technologies.
The four functional and four depth-related categories above form the capability matrix. The adoption of this framework allows the in-depth analysis of supplier strategies at different functional levels. Its main advantage is that it considers the dynamism involved in the improvement of functional capabilities over time. This makes possible the tracing of how firms create new and/or refine existing capabilities throughout the upgrading process.
A further relevant concept for the acquisition/development and deepening of each functional capability is absorptive capacity. It can be defined as "the ability of a firm to recognise the value of new, external information, assimilate it, and apply it to commercial ends" (Cohen and Levinthal, 1990: 218). Cuervo-Cazurra and Rui (2017) suggest to further divide the recognition of external knowledge into the recognition of the need for and the value of external knowledge. Such a separation is important, because firms in an emerging market context may well recognise the value of external knowledge but given certain institutional conditions decide that they don't immediately need it or are simply not able to acquire it (Bilgili et al., 2016;Cuervo-Cazurra and Rui, 2017).

Barriers to absorptive capacity
Cuervo-Cazurra and Rui (2017) investigate the barriers to absorptive capacity based on two cases in the Chinese automotive industry. They distinguish between internal and external barriers to absorptive capacity as well as certain emerging market characteristics moderating the effect on these barriers on firms' absorptive capacity. As internal barriers they identify: • managerial biases: this dimension encompasses the prejudices of decision makers regarding specific sources of knowledge • weak social integration mechanisms: refers to the quality of processes and procedures that facilitate the coordination of employee action External barriers include: • muted activation triggers: these are limitations in managerial ability to recognise the need for external knowledge or its value • conflicting source relationships: these are factors limiting the willingness of the source of external knowledge to willingly share it with the firm • feeble appropriability regimes: this dimension refers to the difficulty of acquiring external knowledge due to weaknesses in intellectual property rights protection The emerging market barriers identified by Cuervo-Cazurra and Rui (2017) include • higher restraints on incentives: government interventions discouraging firms from pursuing profit maximisation or capability upgrading • higher information asymmetries: factors limiting firms in gathering information about the full range of potential choices and possibilities due to the under-developed innovation system in their home country (also cf. Pietrobelli and Rabellotti, 2011) • weaker contract protection: this dimension refers to weaknesses in the judicial system In order to understand what managers can do to facilitate the creation and deepening of capabilities, it is also important to consider whether they recognise barriers to absorptive capacity and if yes how they address these barriers.

Routines
Routines are defined as "complex and analytic processes that extensively rely on existing knowledge, linear execution, and repetition to produce predictable outcomes at different organisational levels" (Eisenhardt andMartin, 2000:1106). The link between routines and capability is widely discussed in the strategic management literature (Salvato and Rerup, 2011;Teece, 2012). However, the causal links from routines to capabilities are not direct; rather, they involve skills, actions, and motivations (Nelson and Winter, 2002). The analysis of routines thus starts from the notion of the skills of an organisation, and a repetitive pattern of activity throughout the entire organisation (Abell et al., 2008). In addition, Abell et al. (2008) argue that a firm can develop capability from routinised micro-processes by internalising the stochastic factors (i.e. contextual challenges). Therefore, developing capabilities from routines involves organisational skill, motivations, and actions that are put in place based on consideration of the contextual challenges. Routines are typically analysed at the collective level in strategic management, given that capabilities are built not just on individual skills but also on collective learning derived from how employees have worked together within a specific time and place (Teece, 2012).
Given the unit of analysis in the current paper is the firm, routines are analysed at the collective level here. However, beyond looking at collective entities, a recent line of interest in strategic management suggests breaking organisations into parts and mapping the cross-functional team routines and their interactions, in order to examine the micro-foundations of capabilities (Eisenhardt and Martin, 2000;Felin et al., 2012;Salvato and Rerup, 2011). In line with this view, the current paper investigates the routines in different functional divisions. This involves exploring what skills, motivations, and repetitive actions are undertaken by the employees in the planning, production, raw material, and marketing divisions, and how these divisional routines internalise the contextual challenges.

Upgrading
In the GVC literature, upgrading is defined as "a process of improving the ability of a firm to move to a more profitable and/or technologically sophisticated capital and skill-intensive economic niche" (Gereffi, 1999:38). We utilised Gereffi and Frederick's (2010) typology to determine our case firms' level of upgrading. The typology consists of four functional upgrading trajectories: • Cut, make and trim (CMT) producers: The focus of the supplier is on the production and assembly of imported inputs, following buyers' specifications.
• Package contractor or OEM: The supplier takes on a broader range of tangible manufacturing-related functions, such as sourcing inputs and inbound logistics, in addition to production.
• Full package provider or original design manufacturer (ODM): The supplier carries out some of the pre-production processes, including design and R&D.
• OBM: The supplier acquires post-production capabilities and is able to fully develop products under its own brand names.

Sample
We selected two Bangladeshi garment-manufacturing firms based on theoretical considerations. Both firms started operating in the early years of the Bangladeshi garment industry (i.e. during the 1980s-1990s). Their current employee numbers range from 10,000-12,000 and their production capacity from 220 to 230 lines. Their turnover is between £380 million and £450 million. Both firms have vertically integrated units for spinning, weaving, knitting, dyeing, and washing, and even accessory production plants. These firms produce most raw materials in-house and have their own design and lab-testing units. Each has its own brand, sold in Bangladesh and neighbouring countries. These two firms have recently engaged in foreign direct investment (FDI) to produce finished garments as well as other product lines. Furthermore, they are both among the very few garment manufacturers in Bangladesh that started as simple assemblers and have gone through an incremental transition to high-value-added activities -OEM, ODM and now OBM, along with off-shoring activities. For this reason, they offer a suitable context in which to explore how such firms have developed capabilities for progressive upgrading and what routines and strategies they have adopted in the process.
Furthermore, both firms cater mainly to European and US based retailers. They produce low-to medium-value-added products for middle-class customers. The product lines mostly comprise fashion items and office wear. Both businesses have recruited a number of overseas managers, industrial engineers, designers, and trainers from Sri Lanka, India, Malaysia, the UK, and the US. Some of the dimensions used for constructing the company profiles were adapted from Locke (2013); others emerged from the data (see Table 1).
Both firms have been involved in repeat transactions with the same buyers for at least 10 years, although there are no legally binding original agreements for repeat purchases. As a result, each transaction has been carried out as a discrete transaction. The firms receive tentative planning and projections of orders for a one-year period (often at the beginning of the year). The projection is not a legally binding contract or a formal purchase order. Rather, it is a mere planning tool. Although this provides a certain degree of assurance of future repeat purchases, it would not allow the suppliers to take any legal action should the buyers change their plans or terminate the relationship. Instead, there is an implicit (or at most orally expressed) promise of future orders that is subject to the suppliers' ability to maintain the expected level of performance within the current exchange in terms of price and quality specifications, on-time delivery, social compliance, and protection of intellectual property (Hoque et al., 2016). They receive orders, specifications, and planning and projection details, if any, directly from the buyers. In order to be able to meet individual buyers' expectations, they have had to make a moderate degree of relationship-specific investment, such as the purchase of customised raw materials, the specialisation of the labour force, and the implementation of corrective action plans imposed by compliance auditors. This investment has been necessary, not only for the completion of the current transactions, but also to attain the buyers' promises that they will make repeat purchases.

The context surrounding the Bangladeshi garment industry
The garment industry in Bangladesh emerged in response to the Multi Fibre Agreement (MFA) of the World Trade Organisation (WTO) that was in force between 1974 and 2004. The MFA, combined with favourable policy initiatives made by the Bangladeshi government, contributed to the rapid growth of the industry (Kabeer and Mahmud, 2004). As a result, the number of garment factories increased from 50 in the early 1980s to 6000 by 2016 (BGMEA, 2016). In contrast to the predictions of numerous researchers (e.g. Yang and Mlachila, 2007), the Bangladeshi garment sector has exhibited a 5% increase in its growth rate, even since the phasing out of the MFA quota in 2005 (BGMEA, 2016). This surge in growth has resulted from Bangladeshi garment manufacturers' continued focus on lean manufacturing systems and low-cost, low-value-added, and large-volume supply with short lead times (McKinsey, 2011).
In addition, Bangladeshi garment manufacturers enjoy the benefits of the Generalised System of Preference (GSP) in 28 countries in the European Union. GSP facilities allow duty-free quota-free access to garments wholly manufactured or substantially transformed in Bangladesh. This facility has eased European market access to Bangladeshi garment manufacturers, especially those that are vertically integrated or source the majority of their raw materials from local suppliers (Export Promotion Bureau, 2016). The government of Bangladesh has also supported the growth of the industry by designing favourable policies for 100% export-oriented garment companies, including providing discounts on corporate tax, duty-free imports of machinery and raw materials, reduced interest on loans, and bonded warehouse and back-to-back letter-of-credit facilities (McKinsey, 2011).
Despite these positive initiatives, the industry has been suffering from a number of challenges that have hindered the growth of the sector. One of the most significant is aggravated production costs and delays due to an inconsistent electricity supply, poor transport networks, and administrative red tape. The second challenge relates to access to capital finance and high interest rates on bank loans (Chowdhury et al., 2014). Given that the bank loans are made available based on firms' order quantity and export volume, large firms enjoy easy access to capital, while small firms have limited access (Labowitz and Baumann-Pauly, 2014). In their study, Chowdhury et al. (2014) find that 52% of the respondent firms complain about high interest rates when seeking loans from private or government banks. Further to this, the commercial and social activities of both public and private banks are highly driven by their need to demonstrate patronage to the ruling party and parliamentarians (cf. Uddin and Siddiqui, 2016). As a result, access to bank loans is also shaped by the political patronage and family influence of firm owners. With 10% of parliamentarians involved in the garment manufacturing business, their firms enjoy fairly easy access to credit facilities (cf. Labowitz and Baumann-Pauly, 2014).
The third challenge relates to buyers' indirect sourcing practices. With most foreign buyers placing orders through purchasing agents, indirect sourcing has become a routine practice in the industry. The Bangladesh Garment Manufacturers' and Exporters' Association (BGMEA) members' guide lists about 1000 agents in operation throughout the industry. These agents identify and select suppliers based on buyers' preferences (i.e. compliant or non-compliant supplier), negotiate terms of orders, and ensure delivery to the freight forwarder (Uddin, 2017). By exploiting their direct linkages with buyers, these agents utilise bargaining power over suppliers to bring the price down (Labowitz and Baumann-Pauly, 2014). The absence of direct interaction with the buyers heightens the asymmetry of power and knowledge between Bangladeshi suppliers and their buyers (Hoque et al., 2016).
Another factor that fuels this asymmetry is the buyers' preference for tacit promissory contracting, as the dominant form of GVC governance in the sector (Hoque et al., 2016). Finally, recent terror attacks by Islamic extremists across the country (The Guardian, 2017) have created security concerns for buyers and resulted in fewer factory visits in Bangladesh, or even in the relocation of their outsourcing activities to neighbouring countries (The Economic Times, 2016).

Results
The analysis yielded two key findings. First, both firms have devised strategies and followed coherent routines involving patterned actions aimed at developing the skills and motivation needed to perform appropriate functional activities (i.e. pre-production, production and post-production) as they have embarked on the different stages of upgrading (i.e. CMT, OEM, ODM, and OBM). The strategies and routines adopted at each stage of upgrading have aided the development of appropriate functional capabilities. At the same time, the depth of functional capabilities has also increased stage-by-stage, from the operational level to the assimilative, the adaptive, and finally the innovative level, as the firms have progressed through each stage of upgrading.
Second, the studied firms have designed these routines in such a way that they allow the firm to address the challenges stemming from their relationship with their buyers, or from the institutional environment, at any given time. These routines have been adjusted at each stage of upgrading.

The CMT stage
Both Firm A and Firm B started their operations with a small garment-producing factory in a rented, shared building, aiming to utilise the growing opportunities within the Bangladeshi garment industry resulting from the MFA. Initially, both firms offered CMT services to foreign buyers based on imported raw materials. The key capability required at this stage was knowledge of basic garment production and the skills needed to execute technology and design specifications. In the beginning, Firm A sold products through buying agents (i.e. independent third parties who would receive orders from foreign buyers and subcontract them to local factories). In contrast, Firm B exported directly to buyers, in addition to selling via purchasing agents. As a result, the marketing capabilities required by Firm B involved the ability to contact foreign buyers directly, promote products, and develop and maintain relationships with buyers. In the case of Firm A, however, the development of marketing capabilities was less important at the initial stage. As both firms relied on design and technology specified by their buyers, and on imported raw materials, they required no special capabilities to manage pre-production functional activities. In sum, in both cases it was sufficient for them to possess operational-level capabilities, allowing them to produce and export garments by using technology, design, and raw materials imported from abroad.
At the CMT stage, both firms designed their routines in such a way as to facilitate the co-ordination of functional divisions. Given they mostly had to produce small order quantities, usually for multiple buyers, internal interaction was critical to meeting buyers' specifications and ensuring on-time delivery. With both firms having few employees (i.e. one or two managing each functional activity), internal co-ordination was easily achieved through the devising of routines for cross-functional team work. For instance, in both cases, the routine in the marketing division involved contacting existing and potential buyers, either via agents or by approaching them directly and then presenting samples and company profiles to them in order to secure orders. Once orders were received, the planning division drew up schedules for processing each order in a systematic manner. The schedules were then coordinated with the production division, which devised daily targets for each production line in order to implement the planned schedule. Understanding the buyers' specifications (related to design, raw materials, and technology) was also a key skill required by the planning, production, and raw material divisions so as to attract repeat purchases from buyers. Each functional division achieved this skill by designing its own skill development routines. For instance, the planning teams in both firms attended regular training provided by the BGMEA to gain knowledge about decoding buyers' designs and scheduling/processing orders. The managers and supervisors in the production division implemented the tasks related to each order by providing specific guidelines and on-the-job training to the workers on a daily/weekly basis.
In the CMT stage, both firms had to manage challenges originating from their relationships with buyers. The owner of Firm A said that selling via purchasing agents restricted their direct interaction with buyers, making them more vulnerable to abuses of power by the agents. The chief operating officer (COO) of Firm B pointed out that the absence of any legal commitment from their buyers meant there was no certainty regarding repeat purchases. Thus, one of the routine activities of the marketing divisions of both firms involved making new contacts with agents on a continuous basis, meeting them in person and proving their credibility by presenting samples and work records, and creating goodwill. Such marketing routines enabled the suppliers to create links with large numbers of agents/buyers and thus reduce the adverse impact of being over-dependent on only a few of them.

The OEM stage
In the second stage, both firms sought to become OEM service providers by establishing vertically integrated textile and accessory production plants. The upgrading initiative was primarily driven by the urge to utilise the GSP facilities that had been made available to vertically integrated Bangladeshi garment manufacturers. However, a potential challenge that could be experienced by any Bangladeshi garment firm at this stage of upgrading is access to the bank loans needed to finance the new investment. Given that both firm owners studied here were politically affiliated and successfully running other business units, they enjoyed relatively easy access to bank loans. Thus, using their political networks and reputations from previous businesses, both firms secured cheap capital from banks to finance their upgrading projects. Further to this, interviewees from both firms mentioned factors that had affected their upgrading projects, such as the absence of roads leading to the chosen location for the project, and limited availability of electricity. Thus, they had invested in the construction of roads linking their factory to the main highway, and also in the establishment of their own electrical substations. Seeking permission for these initiatives from the government had been relatively easy, given their political linkages and powerful industrial positions.
The head of the textile division of Firm A indicated that, in addition to accessing the GSP benefits, they wanted to build in-house raw material production units so as to gain more control over price, quality, and delivery times, and also to demonstrate a strongly differentiated position to their buyers. Subsequently, during the period of 1994-2006, Firm A established a number of vertically integrated textile units, including spinning, fabric, weaving, washing and dyeing, processing, and finishing plants. Six years after its inception, Firm B also decided to gradually develop vertically integrated raw material production units. It established a fabric production plant and then gradually also set up denim (2003), washing (2003), dyeing (2004) and accessories production (2004) plants. Thus, it took four years for the firm to become a full OEM service provider. The COO of Firm B stated: "Our first motivation underlying this upgrading was getting the incentives under the Generalised System of Preferences (GSP) in the European market. At that time, GSP facilities were pertinent for garments imported from Bangladeshi export-oriented firms that had in-house raw material production units. Further to this, C&A, one of our long-term buyers, promised that they would increase orders by 40%, should we be able to develop in-house raw material production units." At the OEM stage, both firms needed to develop a range of pre-production capabilities for the in-house production of raw materials. They had to learn how to set up and operate the machinery used for raw material production. In addition, they were pushed to strengthen their marketing capabilities, in order to promote their new, more differentiated position to high-fashion buyers. Both firms increased their capacity to produce larger volumes of output. Subsequently, Firm A underwent a major process upgrading when it established a new garment unit on a larger scale (with 220 production lines) on its own 200-acre campus in 2005. It installed high-tech machines and integrated advanced facilities into the garment factory in order to increase its production volumes. In the words of the firm's managing director: "Before we established the factory, I personally talked to many CEOs or top managers of the reputed brands to find out what they expect from us. I met some of them in the USA at a conference, while others I knew through my social contacts…After having discussions with them, it seemed to me that having an extensive capacity proves the strength and stability of a factory and thus provides better access to reputed high-fashion buyers." Similarly, Firm B built a new plant with 10 production lines in the year 2000. Between 2007 and 2009, it established two green factories in the industrial area. The two factories combined had 130 production lines compared to the 23-49 production lines the firm had in previous stages. In 2010, it established another green factory, with 90 production lines, advanced high-tech machinery, and energy-saving facilities. Both firms not only enhanced their production capacity but also their efficiency, through systematic planning with the use of advanced software, the recruitment of industrial engineers, and the establishment of alliances with industrial consultants. Therefore, at this stage, both firms were seeking to develop assimilative-level capabilities by building and enhancing their expertise in raw material and garment production, as well as in the area of marketing.
As both firms proceeded towards the OEM stage, they grew bigger and the need for internal co-ordination increased. Hence, both firms had to accommodate regular cross-functional meetings and team-based activities in their daily routines. The in-house manufacturing of raw materials saved time previously lost in sourcing from outside. This extra time was put into their quality control routines, which helped to reduce their defect rates. This change in routine made it easier for them to approach high-fashion buyers, and indeed laid the foundation for progressing to the ODM level. Selling to high-fashion buyers was also necessary for deriving optimum value from their OEM status. Hence, their marketing divisions devised routines including actions aimed at attracting such buyers, for instance regularly attending trade fairs and industrial events, proactively setting up meetings with potential buyers to present the company profile and samples, and various others means of maintaining good relationships with buyers (e.g. gifts, invitations to events). These marketing division routines helped establish longer-term relationships with their buyers, which in turn helped counter the uncertainty caused by the absence of legal contracts. Given that tacit promissory contracting relationships impose high levels of power and knowledge asymmetry between buyers and suppliers, both firms integrated various actions into their daily/ monthly routines so as to develop their absorptive capacity and enhance their skill base. These actions involved on-the-job training provided by internal experts (mostly expatriate engineers or managers) and off-the-job training provided by the BGMEA, overseas organisations, and external consultants. Over this period, these knowledge/skill development routines reduced the knowledge and power asymmetry experienced by these firms.

The ODM stage
In the third stage, both Firm A and Firm B became ODM service providers, after establishing their own in-house design studios. The interviewees from both firms indicated that firms with such facilities were better able to attract high-fashion buyers. Nevertheless, at this stage, both firms were challenged with limited design knowledge. This challenge was further aggravated by the shortage of designers in the local market. Thus, both firms had to overcome this barrier by engaging in a collaborative design arrangement with design consultancy firms and also by recruiting designers from foreign countries. For instance, in the year 2007, Firm A became an ODM service provider with the establishment of their in-house design studio in collaboration with a design consultancy firm based in Spain. Two years later, after they had accumulated a sufficient level of knowledge of fashion design through this collaboration, they established their own design studio. Recently, they have opened two international design offices, in Barcelona and New York, in order to facilitate faster communication with European and US-based buyers. This has also enabled Firm A to bypass the challenge created by the hesitancy on the part of a number of buyers and their designers to travel to Bangladesh due to recent terror threats, with the international offices having made it easier for the buyers and their designers to co-ordinate their efforts. The Managing Director further added that having an international R&D team allowed the firm to keep track of fashion trends and constantly update its knowledge. Such an arrangement is particularly important because: "We primarily work with high-fashion buyers who have now gone season-less. Previously we had to provide designs for four key seasons: Autumn, Winter, Spring and Summer. But now we have to continually innovate designs and present them to buyers in order to get their feedback and also make changes if they demand them. These actions are now easier by having our R&D and marketing teams present in the market." One of their recent R&D activities has been an upcycling project involving the production and sale of garments from discarded, cut pieces on a mass scale. Further to this, they are currently developing designs for active formal wear (clothes that can be used for riding bicycles to work but are also good enough to serve as office wear) on behalf of a potential buyer from Canada. The Managing Director explained that a number of the firm's buyers were now completely dependent on their design service: "Buyers demand our design service as they can save substantially on costs from such design outsourcing. A number of our buyers have even shut down their design units. Thus, they are entirely reliant on us for design." Similarly, almost 17 years after its inception, Firm B decided to become an ODM service provider. It established an in-house design studio in collaboration with a US-based textile design firm. Within two years, it had developed an understanding of fashion trends, using advanced design software, and incorporating complex embroidery and appliqué into garments. The CFO explained: "We were already informally contributing to buyers' designs by making suggestions about the suitability of fabrics, accessories to be used for a particular design of garment. We wanted to formalise things by starting our own design studio so that they [the buyers] would also take us seriously… we did not have the knowledge to do it ourselves. Therefore, we connected with our US partner to learn the art of design, how to use software tools and more importantly how Western fashion trends function." The firm currently has its own independent design studio where it researches fashion trends, raw materials, and software, and creates new designs. An international office is set to start operations in 2018 in Milan. As the CFO puts it: "All the large firms like ours are now starting up their own international offices in Europe and America to make it easier for the buyers to place orders and co-ordinate with them regarding samples and design. Moreover, given the recent turbulences in Dhaka due to terrorist attacks, it is actually wise to have international offices. If buyers worry about coming to Dhaka, they can still coordinate with us via our international offices." In the ODM stage, therefore, both firms had to develop pre-production capabilities related to garment design and the use of advanced software, which they pursued through overseas training and mainly through the recruitment of experienced local and international designers. Therefore, both firms continued to develop assimilative-level capabilities by advancing their skills and knowledge in garment design and production.
During the early years of the ODM stage, therefore, both firms needed to co-ordinate with external firms in design collaborations, which required further adjustments to their functional routines. For instance, Firm A initially set up its design unit in collaboration with a consultancy firm. The new routine for its planning/design division involved constant interaction with the consultancy firm via online channels, creating initial design layouts using advanced design software, and then changing/customising the designs to accommodate buyers' specific requirements. Firm A planned actions that were necessary in order to develop the skills required to implement the changed routine. Examples included the recruitment of design graduates and internees from reputed foreign and local universities, regular training provided to the junior designers by the unit head (an ex-Zara designer), the attendance of fashion shows and fairs, and participation in research collaborations with foreign design institutes. The designers at the international offices proactively develop new designs and make recommendations to buyers. However, sometimes they go shopping along with the buyers to better understand their needs and accommodate their demands into their designs. A routine for incentives (e.g. a quarterly bonus offered to the design team that creates the most innovative and appreciated design) was also formulated to encourage a culture of innovation.
Similarly, Firm B worked in collaboration with a US-based design consultancy firm to provide an in-house design service to its buyers. The collaboration routine was set up by both firms and involved the provision of regular training to Firm B employees by the design consultant, and also daily interaction via email and the use of state-of-the-art design software. Given that the employees within the planning division also needed to use this software, they also had to plan a learning routine in collaboration with the software vendor (an Egyptian firm) involving an initial ten-day training course, followed by regular short trainings on updates and consultations on problem solving. During this stage of upgrading, the production and raw material teams of both firms needed to improve their skills in producing high-quality goods and reducing the defect/waste rate. This goal was achieved through the development of a more stringent quality check routine, and the provision of performance bonuses to the quality teams in both teams, if they hit quality and defect rate targets.
In the initial stage of the ODM stage, both firms experienced challenges in terms of lacking full control of their design units due to the collaborative arrangements with external firms. Both lacked sufficient knowledge and skill to operate the design units independently. Given the lack of quality designers in Bangladesh, they had to rely on overseas designers. However, both firms also recruited local design graduates and upgraded their capabilities to the international standard. They upgraded their skills by devising more rigorous training and knowledge-sharing routines. For instance, over a period of two years from 2007 to 2009, Firm A set the foundation for developing the capacity to absorb knowledge from external sources through these routines, which ultimately enabled it to operate a design unit independently. Furthermore, these skill development routines enabled Firm A to further reduce the challenges emerging from the knowledge and power asymmetry in its GVC.

The OBM stage
In the fourth stage, both Firm A and Firm B gained OBM status with the launch of their own brands, Brand A and Brand B, sold in the local market through their own retail stores. Both firms made this decision based on the growing fashion consciousness among younger Bangladeshi consumers, and the absence of local brands catering to this segment. The main challenge experienced at this stage was the lack of knowledge on marketing and brand management. Both businesses have overcome this barrier by leveraging their long-term experience of garment design, production, and marketing, developed through interactions with buyers, and adapting this to the needs of the local market. As the divisional manager of Firm A put it: "Our brand is trendy, casual and targeted at the younger generation, mainly from the upper and middle class… we have 25 years of experience of producing buyers' brands. Now we want to utilise that knowledge for our own brand, adding Bangladeshi flavour to it…While there is ample opportunity in the local market, there is no local fashion brand that can cater to middle or upper middle class consumers in Bangladesh. Therefore, we realised that Brand A could have great potential." In its recent research, Firm A has further found that there are prospects for a local brand in the lower income segment of the market. Accordingly, Firm A is soon to launch another brand, targeted at that segment. As the Managing Director of Firm A said: "While Brand A remains popular in the upscale market, it is not affordable for lower-income people. For instance, a garment worker wants to look the same as an upper-class girl. Therefore, our new brand will cater to these customers. We will keep the design as trendy as Brand A. However, we will use less expensive fabric, which will allow us to charge a lower price." With respect to overcoming the challenges in the OBM stage, the Chief Financial Officer of Firm B said: "Most local firms fear starting their own brand due to their lack of knowledge on marketing and brand management. We have tackled that fear successfully by leveraging our existing knowledge. We have done marketing activities even before launching our brand, which involved researching the fashion trends in the European and US markets, promoting new design ideas to the buyers and, of course, constantly searching for and adding new buyers. We have just adapted that experience for the local market, which was made even easier by the recruitment of a number of personnel who are experienced in managing other local brands. For instance, our marketing division for Brand B is currently headed by an ex-brand-manager of Firm X." These functional divisions had their own R&D, training, production, and quality control routines that were adapted to the opportunities and challenges of the local market. Given that both firms had only produced for exporting purposes since their inception, one of the challenges experienced during this stage was a lack of knowledge about the local market. They overcame this challenge by setting up new research routines for all the functional divisions. For instance, the marketing team of Firm A collects customer feedback from all the retail outlets, which is then communicated to the planning/design division to integrate the feedback into upcoming designs. Similarly, the raw material division in Firm B conducts research on local fabrics (such as Jamdani, khadi, toshor) and accessories, and provides information to the design division. Both firms have formulated routines for a cross-functional team working to produce designs for regular sale as well as for special occasions (such as Bengali New Year, Independence Day etc.).
After gaining OBM status, both firms began to strategize regarding embarking on the journey towards becoming an MNE. For instance, Firm A's first steps in this regard were the establishment of a joint venture in the US, with PVH, for the production of a wrinkle-free fabric (Cortex-2000; Cotra DP 3.5+) and the introduction of flagship retail stores in Pakistan. At this stage, Firm A was focused on the development of innovative-level capabilities, as it had to instigate the R&D of this wrinkle-free fabric.
The challenges experienced by the firms at this stage involved their limited knowledge of operating in international markets and limited access to different foreign institutions. In order to overcome these barriers, they used their existing knowledge and social and political networks in South Asian markets, while in other markets they adopted strategic alliances and acquisition as entry modes. For instance, Firm A chose to sell its Brand A garments via its own retail outlets in Pakistan, given its business knowledge and institutional access in this market due to its owners being of Pakistani origin. In contrast, it found that, in other markets such as Africa and South America, the use of wholesalers worked better, compensating for its insufficient knowledge of and access to these markets. In order to increase its international reach, the firm is planning to introduce an online shopping experience for global customers. In line with its entry strategy, therefore, Firm A has had to adapt its marketing capabilities, in terms of retail stores, distribution networks, and brand management, to different international markets.
Similarly, Firm B started selling its Brand B in the Indian market through its own retail stores. Afterwards, it established its first joint venture in Myanmar, with a Sri Lankan garment manufacturer, to design and produce lingerie for the South Asian market. It has adapted its marketing capabilities to promote, distribute, and sell lingerie in South Asian countries under its own brand name. In this respect the Chief Financial Officer said: "We know how to manage a brand in Bangladesh. We just had to use that experience for launching our lingerie brand in the South Asian market. The South Asian market is not very different from Bangladesh. Therefore, with some adaptation in branding and marketing strategies, we have achieved our goal." As both firms progressed to the stage of multinationality, they experienced an increased need for collaboration with overseas agents and joint venture partners. As a result, both firms set up routines for joint training and team work to facilitate knowledge transfer from their alliance partners. Regular research actions were also planned to gain more knowledge on international fashion trends. The outcome of the research was then integrated in the design and production of finished garments and raw materials. For instance, in the case of Firm B, the joint venture project for lingerie production and sale was supported by routines facilitating knowledge sharing (e.g. the Sri Lankan partner provided training on the design and production of lingerie, while Firm B shared its knowledge and experience of managing a brand in an international market). In addition, actions were undertaken on collaborative research activities in order to understand the lingerie specifications and fashions in South Asia.

Barriers to absorptive capacity
The main barriers to our two case firms' absorptive capacity directly derive from Bangladesh's market characteristics. More specifically, we observed a high initial information asymmetry between buyers and suppliers stemming from the country's weak innovation system (cf. Pietrobelli and Rabellotti, 2011). Another barrier derives from the absence of formally binding written contracts between buyers and suppliers. This form of buyer-supplier relationship is also termed tacit promissory contracting (Hoque et al., 2016). Given the heightened uncertainty regarding future orders, our case firms were initially hesitant to undertake the investments necessary to upgrade their capabilities before they secured a verbal agreement/tentative promise from buyers to continue the relationship and increase orders. As a result of the government's efforts to support exports, there are sufficient incentives for entrepreneurs to enter the garment manufacturing business and to attempt to capture higher gains. As a result, the observation made by Cuervo-Cazurra and Rui (2017) in the Chinese automotive context regarding the impact of higher restraints on incentives which in turn affect activation triggers does not hold in the Bangladeshi context.

Discussion, contributions and limitations
As outlined in the introduction section, one of the contributions of this research stems from the study context. Hoskisson et al. (2013), and later Bilgili et al. (2016) cluster emerging markets based on the development of their institutional and their factor market development. According to this categorization, Bangladesh is in the category of traditional emerging markets. What makes it an interesting environment to study, among other things, is that the country has remained in this category since its independence in 1971, whereas a number of other countries have moved to the mid-range or newly developed market categories over time. Most studies on emerging market firms cluster in the categories of mid-range emerging markets and newly developed economies. Studies on businesses from such traditional emerging markets are comparatively under-represented in the IB and strategic management literatures. To this end, we contribute to the literature on traditional emerging market firms by investigating how two garment manufacturers build capabilities that allowed them to progress through the different stages of economic upgrading and eventually engage in outward foreign direct investment. We do so by not only looking at the range of functional capabilities but also how these have been deepened over time.
This also answers the call by Bell and Figueiredo (2012) who identify two main patterns in empirical studies focusing on the technological capability development of emerging market firms. Historically, such studies either focus on firms that very rapidly grew into large Chandlerian-type MNEs, or on dyadic relationships between suppliers and Western MNEs, where the latter is a significant source of knowledge and learning. Bell and Figueiredo (2012), furthermore, emphasise that despite the depth of research undertaken in the area of learning, most studies have been conducted in the context of world-leading businesses in developed countries. As a result, they urge the more in-depth analysis of the learning mechanisms of emerging market firms. They also stress the importance of investigating these mechanisms over time, as well as the inclusion of a wider range of capabilities going beyond technological innovation capabilities, such as marketing, logistics, administrative capabilities, etc. (Bell and Figueiredo, 2012). Our analytical framework accounts for the temporal dimension as well as the scope and depth of capabilities. By focusing on how capabilities can be developed and changed through individual action, we also contribute to micro-foundation research. We broke down each functional capability into the skills that are necessary to build them, the actions managers utilise to ensure that employees master, improve, and correctly apply these skills, and the mechanisms managers use to keep employees motivated.
Building on Hoskisson et al.'s (2013) categorization, Bilgili et al., 2016offer a number of theoretical propositions about the absorptive capacity of emerging market firms. They propose that firms from traditional emerging economies are more likely to engage in a duplicative imitation strategy, rely more on external sources of knowledge and thus realise only a small increase in their absorptive capacity. While our case firms started out with duplicative imitation, over time they also engaged in creative imitation and later innovative learning. Also, although they started out with mainly relying on external objective and experiential knowledge, over time they managed to put internal knowledge creation mechanisms in place. As a consequence, they succeeded in increasing their absorptive capacity significantly. In contrast to previous findings from technology-and knowledge-intensive contexts, our case companies have only benefitted from their dyadic relationships with their MNE buyers to a very limited extent. From the start, they have mainly relied on partnerships with local and foreign universities, research institutes, consultancy firms, and suppliers of machines and raw materials, for knowledge and capability acquisition and development.
Furthermore, as tacit promissory contracting can be regarded as a hybrid form of governance, situated between the modular and captive GVC governance archetypes, Pietrobelli and Rabellotti's (2011) framework would either predict supplier learning through "deliberate knowledge transfer from lead firms confined to a narrow range of tasks" or "learning through pressure to accomplish international standards", whereby knowledge transfer is contained in standards, codes, and technical definitions. However, our case companies upgraded from simple CMT to OBM without input or explicit pressure from lead firms. MNE buyers would have been content with the initial CMT services, given that more sophisticated ODMs and OBMs existed in other developing countries. In other words, the learning that took place was not GVC driven; that is, it was not required by the lead firms. Rather, it was proactively initiated by the suppliers. Also, our case companies have not formed any joint ventures until 2013-2014. By that time both companies achieved OBM status.
The absence of an active lead firm's contribution to supplier learning can be explained by the historical development of the global garment industry. In the earlier stages of globalisation, the East Asian apparel industry did encounter a certain degree of knowledge transfer from lead firms (e.g. Gereffi, 1999). However, emerging economies that entered the industry at a later stage were confronted with a different experience. Given that strong OEM and OBM clusters already existed across newly industrialised Asian countries, and a number of emerging economies were competing for CMT orders, there was no need for lead firms to transfer knowledge or deliver inputs to countries like Bangladesh. For example, even small printing factories only receive small pictures of the designs they are required to print, without further specifications. They are expected to learn the skills needed to reproduce the designs and determine colour compositions on their own, if they wish to stay in business. Suppliers of varying sizes, and at varying stages of upgrading, have faced similar situations across a range of activities, from inception.
As outlined in Section 4.5, the main barriers to absorptive capacity directly derived from the institutional/economic environment. Admittedly, our case firms belong to the largest Bangladeshi garment manufacturers with prior business experience and political ties that allowed them to alleviate or at least bypass a number of significant constraints such as poor infrastructure (roads, electricity shortages, etc.) and a shortage of funds (easy access to bank loans). In contrast, Hoque et al. (2016) investigate the learning strategies of three smaller garment manufacturers. Indeed, due to resource constraints as well as lack of significant political ties and extensive social networks these companies could only achieve a relatively small improvement in their absorptive capacity. However, similarly to the two large firms in this study the main barriers to absorptive capacity did not stem from muted activation triggers, managerial biases, or weak social integration mechanisms (cf. Cuervo-Cazurra and Rui, 2017). Instead, they derived directly from the institutional/economic environment. More specifically, these firms did not have the means to hire consultants and employees from other countries to help them build more advanced capabilities and/or deepen existing ones (cf. Hoque et al., 2016).
The main limitations of this research are the following. The case firms in this sample are large, 'privileged' firms (Choksy et al., 2017). As a consequence, they can be expected to face fewer barriers to absorptive capacity than smaller firms. Although we could draw some comparisons in this discussion section between our large case firms and three smaller firms from a different study, future research will need to do more extensive research to create a more comprehensive catalogue of barriers to absorptive capacity in traditional emerging markets. Furthermore, while our findings are of relevance to micro-foundations research, future studies might consider going deeper into this area by considering interactions, contestations, and negotiations between buyers and suppliers and between different levels of employees, functional areas, etc. (Choksy et al., 2017). The practice framework by Jones and Murphy (2011) from the domain of relational economic geography may offer an interesting starting point. Future research may wish to integrate this framework with ongoing developments within strategic management and IB.