A multi-tier study on supply chain flexibility in the automotive industry

Int. J. Production Economics 158 (2014) 91–105

Contents lists available at ScienceDirect

Int. J. Production Economics journal homepage: www.elsevier.com/locate/ijpe

A multi-tier study on supply chain flexibility in the automotive industry Antonio Márcio T. Thomé a, Luiz Felipe Scavarda a, Sílvio R.I. Pires b, Paula Ceryno a,c,n, Katja Klingebiel d,e a

Industrial Engineering Department, Pontifícia Universidade Católica do Rio de Janeiro, Rio de Janeiro, Brazil Methodist University of Piracicaba, São Paulo, Brazil c Production Engineering Department, Universidade Federal do Estado do Rio de Janeiro, Brazil d Fachhochschule Dortmund University of Applied Sciences and Arts, Germany e Fraunhofer IML, Dortmund, Germany b

art ic l e i nf o

a b s t r a c t

Article history: Received 8 December 2013 Accepted 22 July 2014 Available online 1 August 2014

Over the past two decades, the scope of some key subjects in operations management has extended beyond the single company to include supply chain (SC) partners and their interactions. One example of such an extension is the contemporary concept of supply chain flexibility (SCF). Although there has been considerable academic progress on SCF, most of the previous studies on this topic have been confined to a single firm, thereby neglecting other important aspects of a supply chain. Therefore, the development of empirical multi-tier studies capable of investigating the inter-organisational components of SCF is required. Within this context, this paper has the purpose of exploring the main effects of flexible SC capabilities or their lack at various tiers that limit the SC's ability to provide products to end-customers. The research design was a multiple case study with internal and external validity checks, within-case analysis and cross-case comparisons, based on a research framework that scrutinises the relationships between SC contextual constraints and flexibility types. The study took place in three representative SCs of the Brazilian automotive industry and sought mainly to identify and compare SC contextual constraints that hinder product delivery to end-customers. Constraints such as suppliers´ capacity, diversity of suppliers, suppliers´ cooperation, trust and commitment, tariffs, exchange rates and inventory were identified in different supplier tiers of the OEMs as the main factors influencing the observed volume and mix flexibilities. Additionally, SCF types such as sourcing, relational, delivery, postponement, new product and responsiveness influenced the SC's flexibility provided to the endcustomers. & 2014 Elsevier B.V. All rights reserved.

Keywords: Supply chain management Supply chain constraints Automotive industry Empirical study Brazil

1. Introduction Over the past two decades, the scope of some key subjects in operations management has extended beyond the single company to include supply chain (SC) partners and their interactions. One example of such an extension is the current concept of supply chain flexibility (SCF) (Vickery et al., 1999; Lummus et al., 2003; Sawhney, 2006; Winkler, 2008; Stevenson and Spring, 2009; Esmaeilikia et al., 2014). SCF broadens the existing debate on manufacturing flexibility

n Correspondence to: Rua Marquês de São Vicente, 225 sala: 950L, 22453-900 Gávea, Rio de Janeiro (RJ), Brazil. Tel.: þ 55 21 3527 1285x1286; fax: þ55 21 3527 1288. E-mail addresses: [email protected] (A.M.T. Thomé), [email protected] (L.F. Scavarda), [email protected] (S.R.I. Pires), [email protected] (P. Ceryno), [email protected] (K. Klingebiel).

http://dx.doi.org/10.1016/j.ijpe.2014.07.024 0925-5273/& 2014 Elsevier B.V. All rights reserved.

(Slack, 1987; Sethi and Sethi, 1990; Gerwin, 1993) by “looking at those components that make an organisation flexible and [extending] them beyond the organisation's boundaries to other nodes in the supply chain” (Lummus et al., 2003, p. 3). Given the importance of flexibility for achieving a competitive advantage (Stalk, 1988; De Meyer et al., 1989; Kekre and Srinivasan, 1990), it is unsurprising that researchers increasingly study how entire SCs can deliver flexibility to their customers. Although there has been considerable progress on this topic, the current debate is still in its infancy. One major limitation of the literature is that most of the previous studies in SCF are confined to a single firm, thereby neglecting other important aspects of a SC (Moon et al., 2012). Various avenues exist for future research in SCF, and the development of empirical multi-tier studies capable of investigating the inter-organisational components of SCF is one of them (Stevenson and Spring, 2007; Chandra and Grabis, 2009; Fatemi, 2010).

92

A.M.T. Thomé et al. / Int. J. Production Economics 158 (2014) 91–105

During the XX century, the automotive industry developed and became a reference for practically all other industrial sectors in an increasingly global market. The accelerated global process of productive restructuring is at the origin of new factories with innovative productive arrangements in emerging countries, especially in terms of supply chain management (SCM). Therefore, interest in investigating supply chains in the automotive industry in these countries is growing (Pires and Sacomano Neto, 2008; Scavarda et al., 2010). Within this context, this paper's purpose is to explore the effects of flexible SC capabilities or the lack thereof at various tiers in three SCs of the Brazilian automotive industry, which limit the SC's ability to provide products to end-customers. The research design was a multiple case study with internal and external validity checks, within-case analysis and cross-case comparisons, based on a research framework that scrutinises the relationships between SC contextual constraints and flexibility types. In an effort to study connected systems (rather than individual companies and dyads), this research employs an approach similar to previous multi-tier case studies (e.g., Holweg and Pil, 2008), aiming for representation of several tiers rather than a comprehensive picture of any particular tier. The structure of the paper follows the purpose of the study. Section 2 is a review of the literature and serves as the basis to elaborate the conceptual framework for the analysis of SCF. The framework guides the research design and the choice of methods described in Section 3. The presentation of results is found in Section 4, followed by a discussion of the main findings in Section 5. Final remarks, conclusions and suggestion for future research close the paper.

2. Theoretical background Flexibility is a multi-dimensional concept with various facets (Slack, 1987; De Toni and Tonchia, 1998; Sánchez and Pérez, 2005; Stevenson and Spring, 2007; Gosling et al., 2010; Fatemi, 2010; Christopher and Holweg, 2011; Esmaeilikia et al., 2014). To understand which aspects of flexibility are within the scope of the current investigation, this section begins by reviewing the key contributions of the concept of manufacturing and SC flexibilities. Subsequently, previous contributions on empirical SCF research relevant to our case study are discussed. A discussion of contextual SC constraints that might hinder SC flexibility follows. The resultant conceptual framework used for the case analysis is introduced at the end of the section. 2.1. Manufacturing flexibility concept The subject of flexibility is the focus of several contributions in recent decades, including important contributions to the concept itself. Five key aspects of manufacturing flexibility have been suggested: (1) its types (e.g., Slack, 1987; Upton, 1994), (2) its dimensions (e.g., Slack, 1987; Upton, 1994; Koste et al., 2004), (3) its timeframe (e.g., Zelenović, 1982; Carlsson, 1989; Upton, 1994), (4) its hierarchy of capability (e.g., Koste and Malhotra, 1999) and (5) its uses (e.g., Gerwin, 1993; Sawhney, 2006; Hallgren and Olhager, 2009). Because flexibility is commonly associated with the ability to change or react (Upton, 1994; De Toni and Tonchia, 1998), a central aspect of flexibility is the object of change (e.g., “What is it that changes?”). This aspect is commonly referred to as the flexibility type (Slack, 1987; Suarez et al., 1996). According to Slack (1987), there are four types of flexibility in a manufacturing system: product, mix, volume and delivery. Product flexibility refers to the system's ability to introduce new products or make

modifications to existing ones. Mix flexibility denotes the ability of a system to alter its product mix (keeping overall output stable), while volume flexibility refers to a system's ability to change its overall production volume. Finally, delivery flexibility denotes a system's ability to change planned delivery times (or sequences) for existing orders. Upton (1994) suggested a distinction between “external” and “internal” flexibility. External flexibility refers to the flexibility types that matter to the system's customers (e.g., “What the customer sees”, Upton, 1994, p. 75), while internal flexibility comprises all types that are internal to the system and are used to deliver external flexibility (e.g., “What can we do”, Upton, 1994, p. 75). Internal flexibility is also referred as competency (machine, labour, material handling, and routing flexibilities) and external flexibility as capability (volume flexibility and mix flexibility) (Upton, 1995). External flexibility is commonly understood to include the four types originally suggested by Slack (e.g., Suarez et al., 1996; Pagell and Krause, 2004; Sawhney, 2006), while the number of internal flexibility types appears to depend on the specific operational setting. Zhang et al. (2003) found strong, positive, and direct relationships between flexible manufacturing competence and flexible capability, and customer satisfaction. In addition to different types, flexibility has three different meanings (or dimensions): range, response (cf. Slack, 1987) and uniformity (cf. Upton, 1994). Flexibility range denotes the spectrum of states that a system can take, such as the total number of products a manufacturing system can produce (e.g., its product variety) or the range of output volumes at which it can operate. Flexibility response refers to the ease (in terms of time and/or cost) with which the system can adjust within its given range (e.g., how long it takes to switch production from product A to product B). Finally, flexibility uniformity measures the performance distribution of the system over its range (e.g., the ability of the system to produce similar unit costs at different output levels). Koste et al. (2004) have suggested dividing the range dimension into rangenumber and range-heterogeneity to account for the degree of differences between the states a system can take. While rangenumber continues to be a measure for the absolute number of states (e.g., the absolute product variants a system can produce), range-heterogeneity denotes the differences between these states (e.g., a system that produces 100 different variants of the same product may be less flexible than a system that produces five completely different single-variant products). Another key characteristic is the timeframe (Zelenović, 1982; Carlsson, 1989; Upton, 1994). For example, a system may be very flexible in the long term but shows almost no flexibility within an operational time horizon of a few days. Carlsson (1989) provides a good discussion of three key timeframes, operational, tactical and strategic flexibility. Koste and Malhotra (1999) classify flexibility dimensions in a hierarchy of flexibility capability comprised of the levels of strategy (business unit), functional (R&D, marketing, manufacturing, organisational and system), plant (new product, volume, mix, modification and expansion), shop floor (operations and routing) and individual resource (labour, machine and material handling). The levels of flexibility in the hierarchy are viewed as a cone from individual resource to strategy, suggesting that firms gain in flexibility scope as they progress from bottom-up in the cone. Furthermore, some authors have noted the difference between a proactive and reactive use of flexibility (Gerwin, 1993; Upton, 1994; Sawhney, 2006; Stevenson and Spring, 2007). 2.2. Supply chain flexibility concept Research into the flexibility of SCs is relatively new, with the first contributions in this field appearing in the late 1990s (Fisher,

A.M.T. Thomé et al. / Int. J. Production Economics 158 (2014) 91–105

1997; Lowson et al., 1999; Vickery et al., 1999). Fatemi (2010) describes the basic idea behind the extension of the flexibility into SCs: “Now-a-days researchers emphasise that it is important to look beyond the flexible factory to the flexible SC(…). As the SC extends beyond the enterprise, SCF must also extend beyond one firm's internal flexibility (p. 140)”. According to Moon et al. (2012), SCF involves the application of SC resources according to market dynamics, and requires firms to develop cross-functional and cross-company strategies that eliminate bottlenecks and create a level of performance that allows firms to strengthen their competitive advantage in an uncertain market. The current literature on SCF covers a range of foci (for comprehensive literature reviews in SCF, see Stevenson and Spring (2007) and Fatemi (2010); and for SCF planning models, see Esmaeilikia et al. (2014)). For example, Fisher (1997), Randall and Ulrich (2001), Lee (2002) and Qi et al. (2009) study the relationship between SC structure, product structure and external environment. Similar to the literature on manufacturing flexibility (e.g., Swamidass and Newell, 1987; Berry and Cooper, 1999), these authors conclude that flexibility does not always lead to higher profitability; the level of flexibility needs to be aligned or fit with the requirements placed upon the SC. Another focus of SCF is the design of SCs. Chandra and Grabis (2009) present potential tools and techniques for designing and modelling flexibility in SCs, Graves and Tomlin (2003) study how SCs can cost efficiently deliver mix flexibility based on a mathematical model and subsequent simulation. Tsay and Lovejoy (1999) and Liao et al. (2010) have contributed to the topic of SCF measurement by studying how to quantify flexibility and its impact in the SC performance. Kumar and Deshmukh (2006) present a model of manufacturing with attempts to enhance the SCF through a volume multiple assembly line to fulfil the objectives of customisation along with timely delivery. Gosling et al. (2010) have examined how buying organisations can configure their supply networks to achieve SCF, while Lee et al. (2009) have studied supplier alliances in environmental uncertainty, showing their impact on flexibility and suggesting that firms should avoid close supplier relationships in uncertain environments to gain flexibility in switching suppliers. A range of inter-organisational flexibility types complements the flexibility concept. According to Lummus et al. (2003) and Stevenson and Spring (2007), these include re-configuration flexibility as the potential to re-align or re-invent the SC, relationship flexibility as the ability to build collaborative relationships both up and downstream, and logistics flexibility as the potential to rapidly send and receive products cost efficiently. For Duclos et al. (2003) logistics flexibility is the ability to cost effectively receive and deliver products as sources of supply and customers change. Fatemi (2010) proposes postponement flexibility as the ability to keep products in their generic form as long as possible downstream in the SC to incorporate the customer's product requirements in later stages and sourcing flexibility as the ability to find another supplier for each specific component or raw material. Esmaeilikia et al. (2014) offers a three-dimensional framework quantifying SCF measures that can be used for tactical SC planning and optimisation based on supply flexibility (e.g., flexibility in procurement and sourcing processes), manufacturing flexibility (e.g., flexibility in manufacturing and assembly processes), and distribution/logistics flexibility (e.g., flexibility in transportation and warehousing processes). While these new flexibility types are relevant for the interorganisational discussion within the SC, we argue that these types constitute internal types of flexibility because the end-customer does not care how the SC manages to be flexible (e.g., by adding new partners when required or by linking existing partners by flexible logistics). The customer's concern is with the four external

93

types, such as whether the SC can introduce novel products or change its product mix, volume or delivery arrangements according to the customer's requirements. As previously highlighted, these external types have been considered in the manufacturing flexibility context (Slack, 1987). However, the debate should encompass the SC beyond a single manufacturing system. From this perspective, volume flexibility requires close coordination between a firm and its suppliers, especially in the face of increasing demand. This external flexibility type directly affects supply chains' performance by preventing out-of-stock conditions for products that are suddenly in high demand or by preventing high inventory levels (Fatemi, 2010). The ability to rapidly introduce new products and product variety and to adapt lead times to customers' requirements (delivery) are external flexibility types that require the integration of numerous value activities across the SC (Fatemi, 2010).

2.3. Supply chain flexibility empirical studies Despite the increasing importance of the topic, few empirical contributions investigate the structure of SCs from a flexibility viewpoint. Chang et al. (2006) extend the knowledge of manufacturing flexibility regarding its integration with SC activities, focusing on supplier involvement in the motherboard industry in Taiwan. They find that supplier involvement plays a major role in the development and performance of a firm's manufacturing flexibility. However, although they consider part of the supply chain, their focus is restricted to manufacturing flexibility. Sawhney (2006) investigates the interplay between uncertainty and flexibility in the SC and discusses how the processes in a SC interact to deliver flexibility. This author proposes a transformation framework that articulates how managers can configure flexibility simultaneously between the proactive and the reactive uses that coexist in a firm's day-to-day operations. However, a key shortcoming of Swahney's study is also shared by other authors (see Vickery et al., 1999; Swafford et al., 2006; Hua et al., 2009; Merschmann and Thonemann, 2011; and the conclusions reached by Stevenson and Spring (2007) on their literature review on SCF): it does not study the companies (tiers) in an interconnected way. It treats the flexibility of SC partners as a secondary input and not as it actually occurs. Sánchez and Pérez (2005) also do not consider connected firms, but they treat the flexibility of SC partners as a primary input to explore the relationship between the dimensions of SCF and firm performance in Spanish automotive suppliers. They show that companies enhance more the basic flexibility capabilities (at the shop floor level) than aggregate flexibility capabilities (at the customer–supplier level). More recently, Moon et al. (2012) develop a multifaceted scale for SCF through an empirical study among firms. These authors determine how an instrument with a set of multi-item measurement scales representing the SCF construct could be developed and validated. Although they have included many firms within the textile and clothing industry in China, they also do not focus on connected firms within a given SC. The empirical study of Avittathur and Swamidass (2007) embrace connected firms within the SC to investigate the effect of the match between buyer and supplier flexibilities on the performance of U.S. manufacturing plants located in India. Nevertheless, their study is limited to multi-tier pairs (manufacturer/supplier). Stevenson and Spring (2009) investigate the specific inter-firm practices that are used to achieve increased flexibility in multi-tier pairs and in the wider SC and how these practices and effects interact. To date, their study is the closest to the objectives of this current investigation. However, a key shortcoming of their study, also shared by Avittathur and Swamidass (2007) and the other empirical studies on SCF, is that they did not examine the constraints in a SC that limit its ability to deliver products to end-customers.

94

A.M.T. Thomé et al. / Int. J. Production Economics 158 (2014) 91–105

2.4. Supply chain constraints At this point, it is important to define the use of the term constraint in this paper. Goldratt and Cox (1984) initially used the term bottleneck to describe a physically limiting element, such as a machine or a worker. Later, these authors acknowledged that production is a complex system that involves elements such as resources and capabilities, making the identification of real bottlenecks difficult. Hence, Goldratt and Cox (1992) used the term constraint to describe the elements that restrict the throughput and goal. Thus, the term constraint is used in this paper for everything that restricts the SC throughput and goal. When applied to the SC, constraints can be internal (e.g., raw material, capacity and distribution) or external (e.g., market and supply). In addition, the main sources of SC constraints can be physical (raw material shortages, limited capacity resources, limited distribution capacity, and lack of customer demand) and non-physical (obsolete rules, procedures, measures, training, and operating policies) (Simatupang et al., 2004). The SC constraints can be classified in network, institutional and infrastructure contextual dimensions and their impact on SCM may be felt on an operational or strategic level (Lorentz et al., 2012). Table 1 exhibits different SC constraints offered in the literature classified within the contextual dimensions. The use of the terms “capacity” (e.g., output) and capability (e.g., quality or technological state of the system) in this study is consistent with Sethi and Sethi (1990). The lack of manufacturing capacity at different tiers of the SC hinders sourcing and volume flexibility. It might hinder the ability of a system to expand, in the sense of the expansion flexibility in Koste and Malhotra's (1999) framework. It can equally hinder the ability to expand or contract production without laying off workers, as in the original meaning of Bylinsky's_(1983) classic example of a volume flexible automated machine tool plant of 215 workers, which was equivalent to a conventional factory employing 2500 people (Sethi and Sethi, 1990; Koste and Malhotra, 1999). From a Theory of Constraints (TOC) standing point, the flexibility as well as speed and punctuality are strictly related to the presence of capacity slack (excess) or “protective capacity”. In this sense, the points of supply chains that do not have capacity slack are considered constraints. Ideally, the constraints should be closer to customers and market demand. Hence, all resources (of manufacturing or not) throughout the SC, should have a capacity slack, which is seeing as a necessity to obtain a quick, balanced and predictable material flow. In this sense, flexibility ends up being a common feature in companies adopting TOC (Cox and Schleier, 2010). The distinction among the three dimensions of SC constraints of infrastructure, institutions and network (Lorentz, 2009; Lorentz et al., 2012) is of paramount importance for the research framework, as it regroups constraints in a logical set. Barriers to flexibility and actions to overcome them will vary according to the locus of the SC constraints dimensions and flexibility types. The cross-reference of contextual constraints to flexibility types guided field observation, as described further under Section 3.4 and Appendix Table A2. In closing this review, it is worth reinforcing that one major limitation of the literature is that most previous studies in SCF are confined to a single firm or dyads and they do not investigate SC constraints to flexibility. This opens important avenues for multitier studies investigating the inter-organisational components of SCF, in light of SC constraints.

2.5. A SCF conceptual framework A summary with the key flexibility concepts provided herein is offered in Table 2.

The conceptual framework of this study is presented in Fig. 1. It is based on the SCF literature review summarised in Table 2 and was adapted from Kumar et al. (2006), who reviewed and integrated 12 SCF frameworks, including those quoted in our literature review. The framework of Fig. 1 incorporates the SC constraints contextual dimensions from Lorentz et al. (2012) as well, and serves as the basis for the within and cross-case analysis. The framework of Fig. 1 embraces the strategic, tactic and operational levels of the Koste and Malhotra (1999) typology of manufacturing flexibility. From top-down, the framework depicts the overarching influence of the environment, business mission and manufacturing and marketing strategies over the types and dimensions of SCF. The environment, business mission and strategy directly interact with new product and modification/responsiveness flexibilities. Flexibility types such as sourcing, product, delivery, relationship and postponement, as well as SC constraints, are situated in this framework in the interplay of SC members, from the 2nd and 1st tier suppliers to manufacturers, distributors and end-customers. The interaction of SCF and SC constraints results in implemented and observed levels of flexibility types, which might or might not fit the SC structure and the environment. Flexibility implementation and fit will affect SC performance, which in-turn interacts with the environment, mission and strategies in the SC. The dotted rectangle in Fig. 1 is the focus of the research and analysis presented in this study: the types of flexibility implemented, observed and their fit or absence thereof with SC requirements. Particular attention was devoted to the extension of the types of manufacturing external flexibility of volume, mix and delivery to the SC through the analysis of inter-firm flexibility types.

3. Research design and methodology This section describes the methods used for data collection and analysis, chosen based on the conceptual framework of subSection 2.4. After a brief explanation of the method, sampling selection, research propositions, within-case and cross case validity checks, the protocol and observational guidelines adopted for data collection and analysis are introduced. 3.1. The method The research was conducted through a multiple case study approach (Yin, 2008). Case study is indicated especially for the investigation of contemporary phenomena in a real life context, when the boundaries between them are not clearly defined and where the researcher observes the facts and attempts to understand, systematise and analyse them. It is prone to answer questions about why and how, such as those addressed in the search for flexibility presence/absence in the SC studied. Furthermore, it is the method of choice for exploratory research in an emergent field such as SCF (Voss et al., 2002; Yin, 2008). 3.2. Sampling Conceptual representativeness governed case selection for sampling, as opposed to a statistical sampling searching for representativeness of the population at-large. Cases that represent the concept domain under investigation were selected (e.g., flexibility types in the automotive industry in Brazil). An additional consideration in selecting the SC was the possibility of investigating several tiers rather than a buyer–supplier dyad, in accordance with the study objectives. Therefore, this study took place in three

Table 1 SC constraints: a review. Contextual Holt Törnroos and dimensions (1999) Nieminen (1999)

Inventory (storage, backorders, risk pooling) BOM including spare parts Diversity of suppliers Suppliers´ capacity Capacity of local links Sales and operations constraints Flexibility rates a Distribution channels Budgets Financial incentives Customer demand Country's logistics performance (LPI) Country's logistical Infrastructure (e.g., roads, ports) Cooperation, trust and commitment Culture/Language Government policies Tariffs and duties Non-tariffs international trade barriers Exchange rates Workers skills/ availability/training Commercial legislation

N

Holt and Button (2000)

Tan Simatupang Meixell and (2001) et al. (2004) Gargeya (2005)

X

X

Arvis et al. (2007)

Lorentz (2009)

X

Kesen et al. (2010)

X

N N N N

Souza and Tynjälä (2011) Pires (2010)

Amrani et al. (2011)

X

X

X

X

X X

X

X X

Costantino et al. (2012)

Lorentz et al. (2013) X

X X

X X

X

X

X X

X

X

X

X X

X X X

X

X X X

I

X

I I I I

X X

I I

X X X

Lim et al. (2014)

X

X

In

I

Tomino et al. (2009)

X

N

N N N N N In

Melo et al. (2009)

X X

X

X

X

X

X X

X X X

X

A.M.T. Thomé et al. / Int. J. Production Economics 158 (2014) 91–105

Supply chain constraints

X X

Contextual dimensions – I: Institutional; In: Infrastructure; N: Network. a

Flexibility rate is a percentage of the allowed modifications to known orders.

95

96

A.M.T. Thomé et al. / Int. J. Production Economics 158 (2014) 91–105

Table 2 Manufacturing and SC flexibilities: a summary of key features. Categories Manufacturing flexibility Types Dimensions

Time-frame Hierarchy of flexibility capability Uses

Characteristics

References

External: product, mix, volume, delivery Internal: context/operation specific Range (number and heterogeneity) Response Uniformity Short-long term; strategic-tactic-operational Strategy, functional, plant, shop floor and individual resource levels Proactive/reactive

Slack (1987), Upton (1994), Upton (1995), Suarez et al. (1996), De Toni and Tonchia, (1998), Pagell and Krause (2004), Sawhney (2006) Slack (1987), Koste and Malhotra (1999), Koste et al. (2004) Slack (1987) Upton (1994) Zelenović (1982), Carlsson (1989), Upton (1994) Koste and Malhotra (1999) Gerwin (1993), Upton (1994), Sawhney (2006), Hallgren and Olhager (2009)

Supply chain flexibility (SCF) Foci Relationships between SC, product and the environment Fisher, (1997), Randall and Ulrich (2001), Lee (2002), Sawhney (2006), Qi et al. (2009) SCF design Graves and Tomlin (2003), Chandra and Grabis (2009) SCF measurement and effect on SC performance Tsay and Lovejoy (1999), Liao et al. (2010) Customisation Kumar and Deshmukh (2006) Purchasing/suppliers Chang et al. (2006), Gosling et al. (2010), Lee et al. (2009) Inter-firm flexibility types Re-configuration, relationship, Lummus et al. (2003), Stevenson and Spring (2007) Logistics Duclos et al. (2003), Lummus et al. (2003), Stevenson and Spring (2007), Esmaeilikia et al. (2014) Postponement flexibility, sourcing Fatemi (2010), Esmaeilikia et al. (2014) SCF empirical research Relationships

Context-structure Impact Measurement

Early supplier involvement SCF sourcing flexibility (buyers and suppliers) Inter-firms relations Uncertainty environment and flexibility in the SC SCF and firm performance Multi-item measurement scale development for SCF

Chang et al. (2006) Avittathur and Swamidass (2007) Stevenson and Spring (2009) Sawhney (2006), Lee et al. (2009) Sánchez and Pérez (2005) Moon et al. (2012)

Fig. 1. Supply chain flexibility conceptual framework (adapted from Kumar et al. (2006) and Lorentz et al. (2012)).

supply chains, two of which are among the five supply chains with higher sales volume of automobiles in Brazil in the last years. In an effort to study connected systems, as opposed to individual

companies and dyads, this research adopted an approach similar to previous multi-tier case studies (e.g., Choi and Hong, 2002; Holweg and Pil, 2008), aiming for representing many tiers and

A.M.T. Thomé et al. / Int. J. Production Economics 158 (2014) 91–105

their interactions rather than a comprehensive picture of any given tier. This is in line with the argument of Choi and Liker (2002, p. 202) that “it is certainly easier to get data on dyadic relationships, but the more challenging and perhaps more interesting questions involve longer supply chains. This is where the key system dynamics will be revealed.” The respondents of the OEMs' of each SC acted as key informants (Yin, 2008) and helped the researchers during the interviews to identify key first-tier suppliers that could participate in the research. This procedure was also conducted with these first tier suppliers towards identifying the second-tier suppliers to be included. The main goal was to assemble a representative picture of actors across the main tiers of the supply chains.

3.3. Validity checks Any finding or conclusion in a case study is likely to be much more convincing and accurate if they pass construct, internal, external and reliability checks (Yin, 2008). As a check for construct validity, this research used a multitude of data sources. Combining sources of evidence, while shifting between analysis and interpretation, usually denotes triangulation (Denzin, 1978; Yin, 2008), as an attempt to guard against researcher bias and to establish a line of evidence during within-case analysis (Taylor and Bogdan, 1984). Triangulation has been sought both within firms, by comparing the interview responses and field visit observations, and between firms, by comparing the responses of inter-related firms, similarly to the method used by Stevenson and Spring (2009) in their multi-tier study. Internal validity was checked against the conceptual framework, scrutinising for pre-established flexibility types and SC contextual constraints in each SC, through pattern matching by comparing flexibility types and constraints among SC and by checks for rival explanations when discrepancies were encountered (Yin, 2008). The conceptual framework of Section 2.4 and a replication logic were used for an external validity check and to guard against undue generalisations. The adherence to the case study protocol was reinforced during data collection to strengthen the reliability of findings. The field notes and meetings were guided by the observational guidelines of Appendix Table A2. They were scrutinised by at least four researchers, using the data analysis strategy of replication and rival theories proposed by Yin (2008). Internal and external validity checks guided the analysis of results.

97

3.4. Field work protocol The fieldwork protocol was guided by a single research question, based on theory and the conceptual framework of Section 2.4: how do flexibility types observed fit with SC requirements and constraints affect SC ability to provide products to end-customers? Data were gathered during the second half of 2011. Sources of evidence were literature search and documentary reviews, openended interviews with executives of different stakeholders and field visits for direct observation. The research team consisted of four members who jointly conducted the interviews and plant visits. Interviews were conducted with 18 senior executives from three different SCs in the Brazilian automotive industry, covering three or four of their tiers. These interviews were based on a semi-structured questionnaire designed for this purpose (Yin, 2008) respecting the interview protocol adapted from Manuj and Sahin (2011) and depicted in Appendix Table A1. Open-ended questions were sufficiently broad as to elicit non-induced answers from respondents at the focal point of retailers (dealers). A more focused set of questions aiming at corroboration of findings were addressed in the subsequent interviews in a bottom-up, customer-centric approach as the interviews moved up the focal point in the SC (for a distinction between open-ended and focused questions, see Yin (2008); for a similar type of general inquiring in exploratory case studies, see Voss et al. (2002), Table 2, p. 200). The more focused questions are regrouped in the observational guideline provided in Appendix Table A2. The guidelines are based on the conceptual framework of Fig. 1 and intend to assist interviewers during field visits to map the three dimensions of SC contextual constraints of institutions, infrastructure and network (Lorentz, 2009; Lorentz et al., 2012) for the flexibilities types of new product, responsiveness, sourcing, product, relational, delivery and postponement. It is worth noting that the tiers (companies) investigated are connected and maintain client–supplier relationships (or vice versa) in the supply chain. Furthermore, data collection began at the distribution channel (car dealerships and OEMs' regional sales offices) of each of the three SCs with the goal of identifying the main cases of operational volume, mix and delivery flexibility constraints to end-customers. Face-to-face interviews were conducted at the dealers, OEMs' regional sales offices, vehicle assembly plants and tier-1 suppliers. These interviews lasted between 2 and 4 h and were followed by field visits to the plants for direct observation. The interviews with tier-2 suppliers lasted 1 h and were conducted by telephone. A member checking process was also

Table 3 Number of interviews and position of interviewee in the studied companies. SC Interviews conducted with

A

Tier 2 supplier

Tier 1 supplier

OEM

Distributors

– Director/electronic injection supplier – Director/engine blocks and crankshafts supplier

– General manager/engine plant

– Logistics general manager – Production planning manager

– Sales manager/car dealer – Sales manager/OEMs' regional office

– Production planning manager – Procurement manager

– Sales manager/car dealer – Sales manager/OEMs' regional office

– Logistics general manager – Production planning manager

– Sales manager/Car Dealer – Sales manager/OEMs' regional office

– General manager/power steering supplier plant – General manager/alloy wheel supplier plant

B

C

–

Production manager/seat supplier plant

98

A.M.T. Thomé et al. / Int. J. Production Economics 158 (2014) 91–105

conducted with the same managers in order to validate the researchers' results findings and analyses (Bloor, 1997). This was accomplished by showing the data analysis to the participants to allow them to evaluate and provide feedback about the accuracy of the researchers' understanding. This validation process was performed face to face in the executives' offices. Table 3 summarises the number of interviews and position of interviewee in the studied companies.

4. Supply chains studied Hereafter, the three supply chains studied are described and analysed under the perspective and purpose of the research. The first sections cover the supply chains individually and the last offers the cross cases analysis. Fig. 2 presents an overview of the three supply chains highlighting the SC members included. 4.1. Supply chain A This SC is already well established and responsible for producing and distributing vehicle models from a multinational OEM, including one of the best-selling passenger cars in South America. Most of these models belong to the low-cost subcompact market segment, and most of their sales include a 1.0 l engine. In recent years, the Brazilian auto market has experienced significant and beyond expected demand growth. As a result, dealers and OEM's regional sales departments increased their orders to OEM plants. However, during the interviews at the distribution channel, it was evident that two external flexibility types regarding the most popular vehicle model limited end-customers' choice. The limitations were volume, as the market demand for this model was higher than the dealers' offerings; and the mix of the available offer, as the engine

requested by the market was the 1.0 l, and the engine offered by most dealers was the 1.6 l. One significant constraint in the SC was the engine plant capacity for both 1.0 and 1.6 l. Although this first-tier supplier could enhance its overall output from 454,000 engines to 555,000 (more than 20%) just by implementing flexible labour force journeys and contracts, this increase was not sufficient to meet the end-customers' new demand. From this perspective, this firsttier supplier capacity creates a supply-chain constraint by restricting volume flexibility, resulting in a loss of sales to end-customers. This supplier decided to invest US $50 million to increase its daily production from 1964 to 2300 engines. In 1 year, the investment should conclude, and the plant should be able to meet the demand. A potential constraint could then arise in second-tier suppliers, especially those responsible for the engine blocks and crankshafts. Both OEM and engine suppliers are aware of this possibility and are negotiating capacity increases with these second-tier suppliers, but this requires significant investment and time for negotiations. The second-tier supplier has two problems increasing its volume capacity. The first is the high investment for the tools for machining crankshafts and the time to start the new production (up to 2 years). Another problem is that second-tier suppliers are afraid of a future demand decrease that would make them idle again, which happened in the late 1990s after a boom in demand and high investment in capacity in Brazil's automotive industry. These facts create a SC constraint due to a lack of supplier capacity as well as cooperation, trust and commitment between SC members. The engine case offers another interesting perspective and highlights another SC constraint by a second-tier supplier that limits the chain's ability to provide flexibility to its end-customers. The aforementioned increase in demand was stronger for the models and versions with a 1.0-litre engine, but the volume

Fig. 2. An overview of the three studied supply chains.

A.M.T. Thomé et al. / Int. J. Production Economics 158 (2014) 91–105

production increase was limited at the vehicle assembly plants due to a lack of this specific engine. From this perspective, the SC constraint was not the engine plant capacity itself, but its dependency on a sole source supplier of electronic injection. This second-tier supplier could not meet the engine plant's demand for the 1.0-litre engine because its production was designated for another customer with priority (another OEM that was also the supplier's owner), resulting in a lack of cooperation, trust and commitment within SC A. Because this engine component was a “black box” developed under an early supplier involvement (ESI) approach and the development of a new supplier requires time and investment, the engine plant increased the production of another engine type (1.6 l) with another electronic injection supplier that could increase its volume sales. The availability of more 1.6-litre engines made the OEM produce more models using these engines. As a result, these models were pushed to dealers, resulting in a mix flexibility limitation to end-customers that absorbed the 1.6-litre engine vehicles with discounts. In the meantime, the electronic injection supplier for the 1.6 l engine was developed to supply the component for the 1.0 l engine. During the interviews with the managers of the OEMs and the engine plants, it was possible to identify another case in which constraints in the SC at the second-tier supplier level limited the chain's overall ability to provide flexibility to its end-customers. This case resulted from an unexpected supply process breakdown. One second-tier supplier responsible for two key components of the diesel engines had to stop production due to financial problems. Because it was a single sourcing operation, operating under the SC constraint of a lack of diversity of suppliers, the engine plant of SC A could not produce diesel engines to deliver to the vehicle assembly plant, which led to cancellations of sales by the OEM of vehicle models and versions with this fuel for several months. A new second-tier supplier was developed, but in the short term, there was a loss in sales and a high inventory level of diesel engine components that could not be used. 4.2. Supply chain B This SC produces different vehicle models in Brazil. One of these is a highly valued compact model that is considered sophisticated for the South American market. Therefore, its endcustomers are more demanding of quality and product variety. Another significant model produced is a low-cost subcompact vehicle that is largely exported to other South American markets. The SC B also had shortcomings with component suppliers. Air conditioning, power steering and alloy wheels capacities were identified as the main constraints for the two vehicle models. The first two components influenced the production mix of the vehicle assembly plants because these plants had to consider the supply capacity constraints of the component suppliers' production. Thus, the OEM sales department could not offer dealers more cars with these components, resulting in a forced “push” to dealers with a mix that was not ordered. This was a significant issue for the highly valued compact model, where only 60% of the dealers' orders containing air conditioning were delivered with the component by the OEM. Again, the supply chain's ability to provide flexibility to its end-customers was limited. In the case of these components, the limitation for the end-customer was the mix caused by volume capacity from the OEM plant's first-tier suppliers. End-customers of the low-cost subcompact model were not completely disappointed because dealers were able to configure some of the units at their points of sales with these missing components. However, this action was not successful for the highly valued compact model because the customers were more demanding of product quality and were concerned about the late configuration service performed at the dealerships rather than at the vehicle assembly plants. The SC

99

reaction took 2 months. A new contract was developed with the power steering supplier that involved a larger volume of procurement so that the first-tier supplier could contract and train new employees. Meanwhile, flexible labour force journeys were used to increase the production level. A second air conditioning supplier was selected because the current supplier could not cover the gap in demand for this component. The supplier capacity of alloy wheels offers another interesting constraint. Depending on the vehicle model, the consequence to end-customers was different. For the cheap subcompact model, there was a mix flexibility limitation to end-customers because the model was pushed downstream with steel wheels instead of alloy wheels. For the highly valued compact model, there was a volume flexibility limitation. Because this model could not be pushed downstream with steel wheels due to its basic configuration, the OEM had to reduce its production at the vehicle assembly plant. The potential end-customers for this model were disappointed because the units delivered to dealers were insufficient to meet the demand, resulting in lost sales. Within the next 6 months, the OEM resorted to its relational flexibility negotiating an increase in production capacity with the alloy wheel supplier because it did not succeed in finding available capacity from other suppliers due to a diversity of supplier-related constraint. 4.3. Supply chain C In contrast to the other supply chains, where the respective OEMs had been producing in plants in Brazil from the 1960s, the OEM of SC C is a newcomer to the Brazilian market that established its first local vehicle assembly plant in the 1990s. Like many newcomers, this OEM began producing in a greenfield area, attracted by government incentives. Far from important auto-part suppliers and with poor road and train connections, the new plant established a supplier park nearby to locate significant first-tier suppliers, mitigating the logistics problems associated with the greenfield location choice and increasing its logistics flexibility (the potential to receive products rapidly and cost effectively). Another common characteristic of many newcomers is the initial production of only one vehicle model based on many imported components but with an increasingly local content programme. Because this vehicle model belonged to the premium compact segment, the offer to end-customers of a high product variety level (high product mix) within each of its versions was an important issue for the OEM. The interviews at dealers and the OEM's regional office highlighted the mix flexibility limitation to end-customers during the first year of vehicle production in Brazil. During the launch of the new vehicle model, the cost/price of many vehicle components (e.g., engines and transmissions) increased significantly partially due to exchange rates fluctuations, resulting in a significant decrease in this vehicle model's sales from the planned forecast. As a result, the vehicle assembly efforts were concentrated in just two of the four versions originally planned for the vehicle model's offer on the Brazilian market, which reduced the mix range available to end-customers. These two versions had to use components that were already in the pipeline, further reducing the mix within these two versions. For instance, vehicles were produced during the first year with doors and seat trims that were in stock (inventory SC constraint) or with orders already placed (purchased). This prevented end-customers from ordering available trims in the catalogue, restricting their choices to the available trim in the SC pipeline. 4.4. Cross case analysis Herein, the cross case analysis is presented. A summary of the similarities and discrepancies among the three supply chains is presented in Table 4 and synthesised in Fig. 3.

100

A.M.T. Thomé et al. / Int. J. Production Economics 158 (2014) 91–105

Table 4 Cross case comparisons. Characteristics

Supply chains A

B

Time in the market Well established Market segment Low cost subcompact Market dynamics Expansion/growing Main constraints OEM 1st tier supplier

2nd tier supplier

C

Well established New comer Low cost subcompact and highly valued compact Premium compact Expansion/growing Retraction/decrease

None Supplier capacity (engine plant limited capacity production)

None Supplier capacity (limited capacity production of air conditioning, power steering, and alloy wheels) Diversity of suppliers (alloy wheels – single sourcing)

None Exchange rate (engine and transmission) Inventory (overstocked components of the doors and seat trims)

Suppliers capacity as well as cooperation trust and commitment (engine blocks, crankshafts and 1.0 litre electronic injection); Diversity of suppliers (1.0 litre electronic injection and diesel components supply – single sourcing)

Flexibility observed Dealers Postponement: order decisions of trivial configurations/ low heterogeneity OEM

Postponement: late configuration of air conditioning and power steering Air conditioning: sourcing (new supplier) Power steering: flexible labour Alloy wheels: relationships (negotiated an increase in production capacity)

1st tier supplier

Engine plant: labour force journeys and contracts

2nd tier supplier

Electronic injection and diesel components: sourcing (new supplier)

Order-to-delivery strategies

Forecast based; Built to stock; long order lead times

Forecast based; Built to stock; long order lead times

Flexibility use

Mainly reactive, dealers proactive (trivial postponement)

Limitation to end-customers

Volume and mix

Mainly reactive, dealers proactive (trivial postponement) Volume and mix

Two SCs were traditional carmakers in South America, and one was a new comer to the region, offering low-cost and mid-range priced vehicle models. While SC A and SC B were facing a market expansion, SC C was facing a market retraction. The main constraints observed referred to the suppliers´ capacity, diversity, cooperation, trust and commitment in a growing market for SC A and SC B, as well as a high level of inventory of components for a decreasing market for SC C. SC A and SC B reacted to constraints by postponing trivial configurations at the dealers' level and resorting to labour and sourcing flexibility upstream. For SC C, on the other hand, logistics flexibility was deployed at the OEM supplier park. All three adopted order-to-delivery strategies of building to stock, with long order lead times (months) from OEM to dealers. The SCs were mainly reactive regarding flexibility use, but some proactive examples were observed in each of them (the use of postponement and logistics flexibilities). The OEMs from SCs A and B adopt a make-to-stock policy based on sales forecast for their low-cost vehicle models, with few significant vehicle sales built to customer orders (less than 1%). Accordingly, dealers must fully specify their orders (in terms of volume and mix) to the OEM many weeks prior to production. This period is considerably longer than their parent companies' operations, where accommodations to minor changes take place up to a week prior to production. Because of this long order lead-time, dealers postpone some decisions about minor customisations of a vehicle (e.g., including air-conditioning or leather covers on the seats) until they have a specific end-customer order for them because they find it difficult to match actual consumer preferences accurately. The customised options are fitted at the dealers in vehicles already in stock at the manufacturer park or at the dealers

Logistics (supplier park nearby)

Forecast based; built to stock; long order lead times Mainly reactive, OEM proactive (supplier park) Mix

shop as soon as there is an end-customer order instead of being fitted weeks before at the OEMs plants. The chains' ability to provide flexibility to their end-customers is initially restricted because dealers only order mainstream combinations to OEMs to avoid high inventory costs with undesirable combinations at the distribution channel. However, late configuration is adopted among dealers (with a high mix response/short lead-time for end-customers because the inclusion of options is performed within 1 day at the dealer) in an attempt to increase the mix range number output to end-customers. However, the heterogeneity of individual product variants must be considered. This increase is mainly for peripheral variants that have a lower degree of difference compared to fundamental varieties (e.g., factors that lead to a different body-in-white manufacturing process). This leads to a low mix range heterogeneity offered to end-customers. A graphical cross-case synthesis is provided in Fig. 3, where the SC constraints were represented in ellipsis with an arrow pointing to the SC tier affected by the constraint. It is interesting to note that SCs well-stablished in the local market had similar constraints, in contrast with new entrants. SC A and B suffered mainly from 1st tier suppliers capacity constraint and the lack of local link capacities. In the case of SC A, first tier suppliers constraints were further aggravated by supplier diversity and a lack of cooperation, trust and commitment constraints at 2nd tier suppliers level. In both cases, there was a resultant lack of product flexibility (volume and mix) at the dealer and subsequently at the customer level. In SC C, a new entrant endured SC contextual constraints related to the country's lack of logistics infrastructure associated with institutional constraints of tariffs and duties for imported components and fluctuating exchange

A.M.T. Thomé et al. / Int. J. Production Economics 158 (2014) 91–105

CTC

SD

SCA

2º TierSupplier

1º TierSupplier

Engine blocks

SCa, LC

& crankshafts 1.0 litre electronic injection

1.0 and 1.6 litres engines

OEM

101

Distribuition

SCa

PMV

Vehicle Assembly Plant for low cost subcompact models with 1.0 and 1.6 litre engines

Car dealer

Diesel engine components

In

PMV SD

SCB

SCa, LC

Alloy Wheels Air Conditioning

Vehicle Assembly Plant for a low cost subcompact and a highly valued compact models

Car dealer In

Power Steering

CLI

Engines and Transmissions

SCC

TD, Er

Seats

PMV

Vehicle Assembly Plant for a premium compact model

Car dealer In

Doors

SCa: Supplier Capacity; LC: Local link capacity; CTC: Cooperation, trust & commitment; SD: Supplier diversity; PMV: Product (mix and volume); In: Inventory, storage, backorders; CLI: Country’s logistic infrastructure; Td: Tariffs and duties; Er: Exchange rates.

Fig. 3. Cross case synthesis.

rates. In terms of the SC constraints dimensions of Table 1, one can expect that infrastructure and institutional constraints would operate at the strategic and tactical levels (mostly) while network constraints would be situated more at the operational level (Lorentz et al., 2012). In this case, long stablished SC faced mainly operational network-related constraints, while new comers faced strategic-level and tactical level institutional and infrastructure constraints. Consequently, SC A and B resorted to relational and postponement flexibilities to counteract mainly suppliers and local link capacity constraints at 1st and 2nd tiers, while SC C had to endure lost sales and high inventories facing a recessive market with institutional and infrastructural constraints.

5. Discussions SC A was exposed to uncertainty both upstream and downstream from the vehicle assembly plant. From the perspective of the usage aspect of flexibility, it can be said to be reactive with a low level of responsiveness. Despite managing part of the higherthan-expected demand increase with internal labour flexibility at the engine plant, it lacked other internal flexibility types for success. The lack of a flexibility relationship increases the possibility of a constraint at the engine block and crankshaft supplier because there is no cooperation, trust and commitment between this second-tier supplier and other members of the SC (e.g., by sharing investments and/or risks related to future idle production

capacity), which is in keeping with the empirical findings of Stevenson and Spring (2009). The cases where there was a lack of sourcing flexibility that created a SC constraint (diversity of suppliers), regarding second-tier suppliers for the 1.0 engine's electronic injection and components for the diesel engine are also notable; it was impossible to find new suppliers in the short term. While the former was caused by the demand increase, the latter resulted from a supply process breakdown. Here, capacity constraints resulted in a lack of volume-mix flexibility to endcustomers. Both cases were exacerbated by a low level of diversity of suppliers, a consequence of the supplier base rationalisation trend within the automotive industry's SC. On the one hand, this rationalisation seeks the benefits of a close relationship that can provide flexibility with a greater willingness on the part of the supplier to cope with change, but on the other hand, such relationships can make SC re-configuration more difficult, as seen in both cases. This reinforces the concerns in the literature regarding this type of relationship in an environment of uncertainty (Lee et al., 2009), resulting in trade-offs with risk and flexibility in the SC (Stevenson and Spring, 2009). In turn, SC B was also exposed to uncertainty as the demand increased unexpectedly. From the perspective of the usage aspect of flexibility, it was also reactive, but its response varied depending on the limitation and on the analysed product (vehicle model). Considering the air conditioning and power steering components for the low-cost subcompact model, the chain's ability to provide flexibility to its end-customers can be considered satisfactory,

102

A.M.T. Thomé et al. / Int. J. Production Economics 158 (2014) 91–105

although there was a mix limitation delivered to end-customers. The SC managed the demand increase by using different internal flexibility types, such as labour, relationships, postponement and sourcing. Both postponement and sourcing flexibilities can be considered examples of allowing re-configuration flexibility. The same SC did not react as successfully with the highly valued compact model because the postponement flexibility was not successful, resulting in a different flexibility level provided by the chain to end-customers. Considering the supply of alloy wheels, the lack of sourcing flexibility caused a constraint of supplier diversity that consequently limited the supply chain's ability to provide flexibility to its end-customers and impacted them differently (mix for the low-cost subcompact model and volume for the highly valued compact model). Postponement flexibility is present in SCs A and B. This flexibility type is used to address uncertainties in customers' variety preference (mix) resulting from the long order lead-time of the vehicle assembly plants' delivery strategies. Dealers use late configuration as a means to accommodate specific customer preferences within considerably shorter lead-times. Although postponement increases the range number of the mix offered to end-customers, it also has limitations because dealers can add only peripheral variants to the vehicles, resulting in low range heterogeneity. Like SCs A and B, SC C was exposed to uncertainty regarding end-customers' demand. However, this chain had a decrease in its forecasted demand, and its lack of flexibility in dealing with existing inventories and purchased orders in the pipeline caused an inventory SC constraint that limited the mix range-number offered to end-customers. As the product variety offered was considered important in its vehicle's market segment, this mix constraint not only resulted in lost sales but also forced discounts at dealers and ended with some damage to the OEM's brand in the market. It is possible to examine the limited chain's ability to deal with this decrease in demand. In this case, however, the constraints in the SC had a strong influence on other SC members, especially those hosted in the OEM's supplier park near the plant. This supplier park was established to offer logistics flexibility to the OEM in a greenfield area. However, the vehicle model demand was lower than forecasted, resulting in the transference of the seat supplier assembly line from the OEM's supplier park to another existing facility and a huge overcapacity of the other suppliers' installations in the supplier park, as they could not use their capacity to meet demand in other plants. Here, forecast errors at the OEM level did not meet with logistics flexibility and induced a lack of mix flexibility for the end-customer. The flexibility capabilities observed in the three supply chains were at the level of individual resources, shop floor, and plant level in Koste and Malhotra (1999) cone's framework. The cone shape helps portray flexibility as a capability. In other words, observed flexibility is located at the bottom and middle of a cone of flexible capabilities, corresponding to more operational and tactical flexibility, with no strategic flexibility dimensions observed.

6. Conclusions From an academic perspective, this paper contributes to filling the gap in the literature regarding empirical multi-tier studies capable of investigating the inter-organisational components of SCF and SC constraints. Therefore, unlike other empirical studies that examine the theme of SCF under the one-sided view of the company being researched or from dyads, this paper analyses the constraints in three SCs at many different tiers that limit the supply chains' ability to provide products to their end-customers. A framework is proposed that scrutinises the relationships between SC contextual constraints (institutional, infrastructural and network)

and flexibility types. The research is intentionally restricted to focus on the flexibility types that matter to end-customers (and their dimensional aspects) as a reaction to uncertainty in operational and tactical time-frame perspectives. Constraints were identified in different SC members, including tier-1 and tier-2 suppliers, which reinforce the academic literature findings about the need to extend the flexibility debate to go beyond the boundaries of a single company'. The main constraints identified were suppliers´ capacity, diversity of suppliers, suppliers´ cooperation, trust and commitment, tariffs, exchange rates and inventory. These constraints jeopardised SCs flexibility and significantly reduced their ability to meet the requirements of the end customers. This study also provides useful guidance for SC practitioners. In this sense, the SCs presented different aspects of flexibility that highlighted or mitigated the effects of SC constraints. Supplier capacity constraints caused by supply process breakdowns and supply disruptions, highlight the growing need to manage risks in SCs and support the increasing interest in the subject of SC risk management as suggested by Christopher and Lee (2004), Manuj and Mentzer (2008), Stank et al. (2011), among others. The study also implicitly highlights some drawbacks of the practice of reducing the supplier base, which in recent decades has become frequent in the automotive industry, especially in the Brazilian case (Pires and Sacomano Neto, 2008; Scavarda et al., 2010). Thus, the dependence on a single supplier for a particular item can result in a major constraint to the SC as a whole. Avittathur and Swamidass (2007) report a similar constraint in India. Likewise, the absence of one type of flexibility in one member of the SC may negatively affect another type of flexibility downstream in the SC, limiting the overall supply chain's ability to demonstrate flexibility to its end-customers. For instance, the lack of relationship and sourcing flexibilities with OEMs' first- and second-tier suppliers may cause supplier's cooperation, trust and commitment constraints that result in volume and mix limitations to end-customers. Similarly, the presence of one flexibility type may improve the overall supply chain's ability to demonstrate flexibility to its end-customers: postponing part of the vehicle configuration downstream in the SC may improve the mix offered to end-customers. This reinforces the interdependence of the SC members highlighting the need to match/align customer flexibility to supplier flexibility, which is in line with the studies from Avittathur and Swamidass (2007) and Hua et al. (2009). Furthermore, the results highlight the importance of SCF to increase the supply chain's ability to change its overall production volume or its product mix. It reinforces volume flexibility as a way to respond to demand uncertainty (for similar conclusions see Sánchez and Pérez (2005)). However, this is not always correctly understood among the analysed firms. Increasing flexibility is not the same as increasing the supply chain's overall production plant capacity to react to a demand increase, as noted in our findings. If the demand decreases, this production capacity will became idle and stock will fill the pipeline. This concern was raised in the interviews regarding SCs A and B and was the downfall of SC C. A flexible SC should be designed so it can increase or decrease its production level to adapt to demand, which requires not only a firm flexibility level (e.g., internal manufacturing flexibility) but also a customer–supplier flexibility level. Our research findings point to SC members focused on the enhancement of flexibility at the firm level more than at the customer–supplier level. This corroborates findings by Sánchez and Pérez (2005) in the Spanish automotive industry and reinforces the fact that companies might miss opportunities to improve competitiveness by underestimating customer–supplier flexibility capabilities. From the perspective of usage flexibility, the analysed SCs have been reactive to uncertainty (although with low response), and none of the studied SCs used flexibility to completely seize opportunities proactively. Our results also call for additional research. The effective use of flexibility concurrently for both proactive and reactive purposes was

A.M.T. Thomé et al. / Int. J. Production Economics 158 (2014) 91–105

already studied in the printed circuit board industry (Sawhney, 2006) and should be better understood in the Brazilian automotive industry context, as it can become a competitive advantage from a SCF perspective. With the increasing synchronisation of SCs, bottlenecks in supplying operations can become serious constraints in delivering products to customers. Therefore, additional research on how a SC can be designed to increase the overall level of flexibility and how factors within and between the individual SC partners may restrict this flexibility is crucial for SC managers across many industries. In terms of contribution to the SCF empirical research, we can highlight at least three important points. Firstly, the object studied here was not restricted to just one or tier-pairs within the supply chains. Unlike other previous reported studies, this research analysed at least three tier levels within the supply chains. Therefore, this lack in the literature (analysing simultaneously the interrelations of at least three tier levels within the supply chains) required the creation of a new research design to conduct

103

and present the results of research, which helps in the search for new approaches to researching this emerging topic. In this sense, and secondly, this study offers an integrated SC constraints and SCF framework that can help both academics and practitioners in this understanding. Thirdly, flexibility and constraints were jointly approached from a SC perspective, which is not common among the extant literature on TOC and SCF. Finally, it is worth noting that although the case study was conducted on three relevant SCs of an emergent economy, the research approach provides evidence for a single industry and country, which limits the extent to which the findings can be generalised across a wider range of industries and countries. Because comparable studies in the literature are still lacking, this investigation is only an initial step towards the study of a topic that is likely to grow in importance as companies continue to reduce their supplier base and outsource their activities, making them more dependent on their SC members, both upstream and downstream.

Table A1 Interview protocol. Opening Demographic data Questions for dealers and OEMs' Regional sales offices

Questions for OEMs and suppliers

Additional unplanned/floating prompts

Introductions of the interviewer and the interviewee; overview of the study's scope; confidentiality assurance. Title of the interview participants; job history; background of the organisation, supply chain and industry. Could you indicate a significant case involving your supply chain that limited the ability to provide external flexibility to endcustomers? Could you provide its main reasons and supply chain members involved? What were the main consequences for the supply chain? How did you deal with this fact? What makes it easy (or difficult) for you to deal with flexibility within a supply chain perspective? What may help you in dealing with flexibility? Do you agree with the case identified by your downstream partner? Could you provide its main reasons and supply chain members involved? What were the main consequences for the supply chain? How did you deal with this fact? What makes it easy (or difficult) for you to deal with flexibility within a supply chain perspective? What may help you in dealing with flexibility? Will you describe that? Could you tell me more about that? Will you explain that in more detail? Can you give me examples or tell me about a related case? Could you provide the contacts of the executives companies that were involved so we could try to include them in the study?

Table A2 Observational guideline. Flexibility types

Product (mix and flexibility): ability to alter product mix (keeping overall output stable) and to change its overall production volume

Supply Chain constraints dimensions Network

Infrastructure

Institutional

Are there evidences of changes in the mix of products offered (e.g., models, options)? Did the network faced sudden rises or fall on the volume of products demanded by the market? If yes, how did the network respond? At which level in the SC?

Are there identified bottlenecks in the logistics capabilities of service providers and of the existing logistic infrastructure that hinds changes in product mix or product volumes? What are the constraints, when did it occur and for which products? How often?

Are there trade barriers, laws and regulations, organisational policies and cultural factors preventing the SC network to deliver product flexibility (mix and volume)?

Are the logistics providers, plant/seller location and logistics infrastructure perceived as problems to acquire products/components/materials/parts?

Are there trade barriers, laws and regulations, organisational policies and cultural factors preventing a diversification of suppliers? Are there trade barriers, laws and regulations, organisational policies and cultural factors hindering relational flexibility?

Sourcing: ability to find another Is the supplier base large or restricted (e.g., partner, upstream single sourcing)? Is there a well-stablished local basis of components and spare parts suppliers? Relational: ability to build collaborative relationships, both upstream and downstream

Ask for a description and review materials about the existing network of suppliers, distributors and retailers for the products selected for the research: are they stable? Since when were they stablished?

Are there any infrastructure related obstacles for relational flexibility, such as geographical distances, road conditions, ports that could affect the flow of products/material in the network?

Delivery: ability to receive and deliver products as sources of suppliers and distributors changes

Do the network experiences deliveries failures or delays (upstream/downstream) as a result of suppliers/distributors changes?

Do the logistics infrastructure and capabilities affect the ability to receive/send products in time and in the adequate quantities when suppliers and/or distributors change?

Are there trade barriers, laws and regulations, organisational policies and cultural factors hindering delivery flexibility?

Postponement: ability to keep products generic until late in the distribution network

Are there examples of postponement of product configuration in the network? For which products/models? Where the final configuration (assembling, accessories, etc.) take place?

Do the logistics infrastructure and capabilities affect the ability to deliver postponement flexibility (e.g., transportation and storage of components and/or spare parts; local assembling/installation of options...)?

Are there trade barriers, laws and regulations, organisational policies and cultural factors hindering postponement flexibility?

General instructions: (i) Review each constraint separately for each flexibility type. Probe with additional questions and ask for additional evidence when answers are not sufficiently clear; (ii) After identifying flexibility issues (constraints/opportunities) ask how the specific flexibility issues identified affect the introduction of new products and the ability of the SC to introduce new product flexibility or to be responsive in adapting/modifying existing products (modification flexibility).

104

A.M.T. Thomé et al. / Int. J. Production Economics 158 (2014) 91–105

Acknowledgements The Authors gratefully acknowledge all the top managers for their invaluable assistance with this research project. The authors are also very grateful to the anonymous referees for their constructive suggestions and to the research agencies CNPq, CAPES, DAAD and DFG for their support.

Appendix A See Tables A1 and A2.

References Arvis, J., Mustra, M., Ojala, L., Panzer, J., Naula, T., 2007. Connecting to Compete 2010: Trade Logistics in the Global Economy. World Bank, Washington (64 p.). Amrani, A., Deschamps, J., Bourrieres, J., 2011. The impact of supply contracts on supply chain product-flow management. J. Manuf. Syst. 31 (2), 253–266. Avittathur, B., Swamidass, P., 2007. Matching plant flexibility and supplier flexibility: lessons from small suppliers of U.S. manufacturing plants in India. J. Oper. Manag. 25, 717–735. Berry, W.L., Cooper, M.C., 1999. Manufacturing flexibility: methods for measuring the impact of product variety on performance in process industries. J. Oper. Manag. 17 (2), 163–178. Bloor, M., 1997. Techniques of validation in qualitative research: a critical commentary. In: Miller, G., Dingwall, R. (Eds.), Context and Method in Qualitative Research. SAGE, London, pp. 37–50. Bylinsky, G., 1983. The race to the automatic factory. Fortune 21, 52–64. Carlsson, B., 1989. Flexibility and the theory of the firm. Int. J. Ind. Organ. 7 (2), 179–220. Chandra, C., Grabis, J., 2009. Role of flexibility in supply chain design and modeling – introduction to the special issue. Int. J. Manag. Sci. 37 (4), 743–745. Chang, S., Chen, R., Linc, R., Shiaw-Wen Tien, S., Sheu, C., 2006. Supplier involvement and manufacturing flexibility. Technovation 26, 1136–1146. Choi, T.Y., Liker, J., 2002. Guest editorial: supply chain management as an emerging focus on technology management. IEEE Trans. Eng. Manag. 49 (3), 198–204. Choi, T.Y., Hong, Y., 2002. Unveiling the structure of supply networks: case studies in Honda, Acura and DaimlerChrysler. J. Oper. Manag. 20 (5), 469–493. Christopher, M., Lee, H., 2004. Mitigating supply chain risk through improved confidence. Int. J. Phys. Distrib. Logist. Manag. 34 (5), 388–396. Christopher, M., Holweg, M., 2011. Supply Chain 2.0: managing supply chains in the era of turbulence. Int. J. Phys. Distrib. Logist. Manag. 41 (1), 63–82. Costantino, N., Dotoli, M., Falagario, M., Fanti, M.P., Mangini, A.M., 2012. A model for supply management of agile manufacturing supply chains. Int. J. Prod. Econ. 135 (1), 451–457. Cox III, J.F., Schleier Jr., J.G., 2010. Theory of Constraints Handbook. McGraw Hill, New York (1214 p.). De Meyer, A., Nakane, J., Miller, J.G., Ferdows, K., 1989. Flexibility: the next competitive battle. Strat. Manag. J. 10 (2), 135–144. De Toni, A., Tonchia, S., 1998. Manufacturing flexibility: a literature review. Int. J. Prod. Res. 36 (6), 1587–1617. Denzin, N.K., 1978. The logic of naturalistic inquiry. In: Denzin, NK (Ed.), Sociological Methods, a Sourcebook. McGraw-Hill, New York, pp. 54–73. Duclos, L., Vokurka, R., Lummus, R., 2003. A conceptual model of supply chain flexibility. Ind. Manag. Data Syst. 103 (6), 446–456. Esmaeilikia, M., Fahimnia, B., Sarkis, J., Govindan, K., Kumar, A., Mo, J., 2014. Tactical Supply Chain Planning Models with Inherent Flexibility: Definition and Review, Annals of Operations Research, Published online: 18 February 2014. http://dx. doi.org/10.1007/s10479-014-1544-3. Fatemi, M., 2010. Supply chain flexibility: definition and review. Eur. J. Econ. Finan. Admin. Sci. 20 (20), 140–147. Fisher, M., 1997. What is the right supply chain for your product? Harv. Bus. Rev. 75 (2), 105–116. Gerwin, D., 1993. Manufacturing flexibility: a strategic perspective. Manag. Sci. 39 (4), 395–410. Goldratt, E., Cox, J., 1984. The Goal. North River Press, Croton-on-Hudson, NY. Goldratt, E., Cox, J., 1992. The Goal: A Process of Ongoing Improvement, 2nd rev. ed.. North River Press Publishing Co., Great Barrington, MA. Gosling, J., Purvis, L., Naim, M., 2010. Supply chain flexibility as a determinant of supplier selection. Int. J. Prod. Econ. 128 (1), 11–21. Graves, S., Tomlin, B., 2003. Process flexibility in supply chains. Manag. Science 49 (7), 907–919. Hallgren, M., Olhager, J., 2009. Flexibility configurations: empirical analysis of volume and product mix flexibility. Omega 37, 746–756. Holweg, M., Pil, F., 2008. Theoretical perspectives on the coordination of supply chains. J. Oper. Manag. 26, 389–406. Hua, S., Chatterjee, S.R., Kang-Kang, Y., 2009. Access flexibility, trust and performance in achieving competitiveness: an empirical study of Chinese suppliers and distributors. J. Chin. Econ. Foreign Trade Stud. 2 (1), 31–46.

Holt, J., 1999. TOC in supply chain management. In: 1999 Constraints Management Symposium Proceedings. Phoenix, AZ, USA, pp. 85–87. Holt, J., Button, S., 2000. Integrating Strategy and Tactics across Multiple Business Units: The Supply Chain Solution, SAE Technical Paper 2000-01-1767. Kekre, S., Srinivasan, K., 1990. Broader product line: a necessity to achieve success? Manag. Sci. 36 (10), 1216–1232. Kesen, S., Kanchanaphiboon, A., Das, S., 2010. Evaluating supply chain flexibility with order quantity constraints and lost sales. Int. J. Prod. Econ. 126 (2), 181–188. Koste, L., Malhotra, M., 1999. A theoretical framework for analyzing the dimensions of manufacturing flexibility. J. Oper. Manag. 18, 75–93. Koste, L., Malhotra, M., Sharma, S., 2004. Measuring dimensions of manufacturing flexibility”. J. Oper. Manag. 22 (2), 171–196. Kumar, V., Kamel, F., Kumar, U., Boyle, T., 2006. Implementation and management framework for supply chain flexibility. J. Enterp. Inf. Manag. 19 (3), 303–319. Kumar, P., Deshmukh, S.G., 2006. A model for flexible supply chain through flexible manufacturing. Glob. J. Flex. Syst. Manag. 3–4, 17–24. Lee, H., 2002. Aligning supply chain strategies with product uncertainties. Calif. Manag. Rev. 44 (3), 105–119. Lee, P., Yeung, A., Edwin Cheng, T., 2009. Supplier alliances and environmental uncertainty: an empirical study. Int. J. Prod. Econ. 120 (1), 190–204. Liao, Y., Hong, P., Rao, S.S., 2010. Supply management, supply flexibility and performance outcomes: an empirical investigation of manufacturing firms. J. Supply Chain Manag. 46 (3), 6–22. Lim, L., Alpan, G., Penz, B., 2014. Reconciling sales and operations management with distant suppliers in the automotive industry: a simulation approach. Int. J. Prod. Econ. 151, 20–36. Lorentz, H., 2009. (Doctoral thesis). Contextual Supply Chain Constraints in Emerging Markets – Exploring the Implications for Foreign Firms. Publications of Turku School of Economics. Lorentz, H., Shi, Y., Hilmola, O.P., Srai, J.S., 2012. Supply chain constraints, opportunities, and adjustments in emerging markets. Benchmark. Int. J. 19 (1), Available from: /http://www.emeraldinsight.com/journals.htm?articleid=17021713S. Lorentz, H., Kittipanya-ngam, P., Srai, J., 2013. Emerging market characteristics and supply network adjustments in internationalising food supply chains. Int. J. Prod. Econ. 145 (1), 220–232. Lowson, B., King, R., Hunter, A., 1999. Quick Response. Wiley, Chichester. Lummus, R., Duclos, L., Vokurka, R., 2003. Supply chain flexibility: building a new model. Glob. J. Flex. Syst. Manag. 4 (4), 1–13. Manuj, I., Sahin, F., 2011. A model of supply chain and supply chain decision-making complexity. Int. J. Phys. Distrib. Logist. Manag. 41 (5), 511–549. Manuj, I., Mentzer, J., 2008. Global supply chain risk management strategies. Int. J. Phys. Distrib. Logist. Manag. 38 (3), 192–223. Melo, M.T., Nickel, S., Saldanha-da-Gama, F., 2009. Facility location and supply chain management – a review. Eur. J. Oper. Res. 196 (2), 401–412. Merschmann, U., Thonemann, U., 2011. Supply chain flexibility, uncertainty and firm performance: an empirical analysis of German manufacturing firms. Int. J. Prod. Econ. 130 (1), 43–53. Meixell, M.J., Gargeya, V.B., 2005. Global supply chain design: a literature review and critique. Transp. Res. Part E Logist. Transp. Rev. 41, 531–550. Moon, K., Ying, C., Ngai, E., 2012. An instrument for measuring supply chain flexibility for the textile and clothing companies. Eur. J. Oper. Res. 222 (2), 191–203. Pagell, M., Krause, D., 2004. Re-exploring the relationship between flexibility and the external environment. J. Oper. Manag. 21 (6), 629–649. Pires, S., Sacomano Neto, M., 2008. Configurations in supply chains: the case of a condominium in Brazil's automotive industry. Supply Chain Manag. Int. J. 13 (4), 328–334. Qi, Y., Boyer, K., Zhao, X., 2009. Supply chain strategy, product characteristics, and performance impact: evidence from Chinese manufacturers. Decis. Sci. 40 (4), 667–695. Randall, T., Ulrich, K., 2001. Product variety, supply chain structure, and firm performance: analysis of the U.S. bicycle Industry. Manag. Sci. 47 (12), 1588–1604. Sawhney, R., 2006. Interplay between uncertainty and flexibility across the valuechain: towards a transformation model of manufacturing flexibility. J. Oper. Manag. 24 (5), 476–493. Sánchez, A., Pérez, M., 2005. Supply chain flexibility and firm performance: a conceptual model and empirical study in the automotive industry. Int. J. Oper. Prod. Manag. 7 (7), 681–700. Scavarda, L.F., Reichhart, A., Hamacher, S., Holweg, M., 2010. Managing product variety in emerging markets. Int. J. Oper. Prod. Manag. 30 (2), 205–224. Sethi, A.K., Sethi, S.P., 1990. Flexibility in manufacturing: a survey. Int. J. Flex. Manuf. Syst. 2 (4), 289–328. Simatupang, T.M., Wright, A.C., Sridharan, R., 2004. Applying the theory of constraints to supply chain collaboration. Supply Chain Manag. Int. J. 9 (1), 1–29. Souza, F., Pires, S., 2010. Theory of constraints contributions to outbound logistics. Manag. Res. Rev. 33 (7), 683–700. Slack, N., 1987. The flexibility of manufacturing systems. Int. J. Oper. Prod. Manag. 7 (4), 35–45. Stalk, G., 1988. Time – the next source of competitive advantage. Harv. Bus. Rev. 66 (4), 41–51. Stank, T., Ditmann, P., Autry, C., 2011. The new supply chain agenda: a synopsis and directions for future research. Int. J. Phys. Distrib. Logist. Manag. 41 (10), 940–955.

A.M.T. Thomé et al. / Int. J. Production Economics 158 (2014) 91–105 Stevenson, M., Spring, M., 2009. Supply chain flexibility: an inter-firm empirical study. Int. J. Oper. Prod. Manag. 29 (9), 946–971. Stevenson, M., Spring, M., 2007. Flexibility from a supply chain perspective: definition and review. Int. J. Oper. Prod. Manag. 27 (7), 685–713. Suarez, F., Cusumano, M., Fine, C., 1996. An empirical study of manufacturing flexibility in printed circuit board assembly. Oper. Res. 44 (1), 223–240. Swafford, P., Ghosh, S., Murthy, N., 2006. The antecedents of supply chain agility of a firm: scale development and model testing. J. Oper. Manag. 24 (2), 170–188. Swamidass, P., Newell, W., 1987. Manufacturing strategy, environmental uncertainty and performance: a path analytic model. Manag. Sci. 33 (4), 509–524. Tan, K.C., 2001. A framework of supply chain management literature. Eur. J. Purch. Supply Manag. 7 (1), 39–48. Taylor, S.J., Bogdan, R., 1984. Introduction to Qualitative Research Methods: The Search for Meanings. John Wiley & Sons, New York. Törnroos, J.Å., Nieminen, J., 1999. Business Entry in Eastern Europe. A Network and Learning Approach with Case Studies. Kikimora Publications, Saarijärvi (Finland), pp. 97–108. Tomino, T., Park, Y.W., Hong, P., Roh, J.J., 2009. Market Flexible customizing system (MFCS) of Japanese vehicle manufacturers: an analysis of Toyota, Nissan and Mitsubishi. Int. J. Prod. Econ. 118 (2), 375–386.

105

Tsay, A., Lovejoy, W., 1999. Quantity flexibility contracts and supply chain performance. Manuf. Serv. Oper. Manag. 1 (2), 89–111. Tynjälä, T., 2011. An Effective Tool for Supply Chain Decision Support during New Product Development Process (Doctoral thesis). School of Science Aalto University. Upton, D.M., 1994. The management of manufacturing flexibility. Calif. Manag. Rev. 36 (2), 72–89. Upton, D.M., 1995. What really makes factories flexible? Harv. Bus. Rev. 4, 74–84. Vickery, S., Calantone, R., Droge, C., 1999. Supply chain flexibility: an empirical study. J. Supply Chain Manag. Glob. Rev. Purchas. Supply 35 (3), 16–23. Voss, C., Tsikriktis, N., Frohlich, M., 2002. Case research in operations management. Int. J. Oper. Prod. Manag. 22 (2), 195–219. Winkler, H., 2008. How to improve supply chain flexibility using strategic supply chain networks. Logist. Res. 1 (1), 15–25. Yin, R., 2008. Case Study Research: Designs and Methods, 3rd ed.. Sage Publications, London. Zhang, Q., Vonderembse, M.A., Lim, J.S., 2003. Manufacturing flexibility: defining and analyzing relationships among competence, capability and customer satisfaction. J. Oper. Manag. 21 (2), 173–191. Zelenović, D.M., 1982. Flexibility: a condition for effective production systems. Int. J. Prod. Res. 20 (3), 319–337.