industry co-evolution: a comparison of taiwan and north america's

A Comparison of TaiwElectronics

Ji-Ren Lee

ITEC Research Paper Series04-03

June 2004

Industry Co-Evolution:an and North America’s Contract Manufacturers

Timothy J. Sturgeon

Industry Co-Evolution: A Comparison of Taiwan and North America’s Electronics

Contract Manufacturers

Institute for Technology, Enterprise and Competitiveness, Doshisha University Research Paper 04-03

Timothy J. Sturgeon Senior Research Affiliate

Industrial Performance Center Massachusetts Institute of Technology

Cambridge, MA, U.S.A. E-mail: [email protected]

Fax: +1-617-253-7570 and Research Fellow

Institute for Technology, Enterprise and Competitiveness (ITEC) Doshisha University

Ji-Ren Lee Professor, Department of International Business

National Taiwan University 50, Lane 144, Keelung Rd. Section 4

Taipei 106, Taiwan, R.O.C. E-mail: [email protected]: +886-2-2362-9970

mailto:[email protected]

mailto:[email protected]

Abstract:

In this paper we explore a case where increased strategic outsourcing by groups of lead firms has, over time, led to the rise of a shared supplier network, one that can be accessed by the industry as a whole. Our approach postulates an industry-wide virtuous cycle between lead firm strategic outsourcing and the development of supplier competencies. As an illustrative case we trace the evolution of the largest electronics contract manufacturers based in the United States and Taiwan, and discuss several trends which may be pushing Taiwan and North American contract manufacturers closer together in terms of competition and geographic location. Keywords: Outsourcing, contract manufacturing, modularity, transaction value JEL codes: L14, L22, L63, N65, O32 Acknowledgements： This paper will appear as Chapter Three in Global Taiwan, edited by Suzanne Berger and Richard Lester, accepted for publication by M.E. Sharpe. In the preparation of this paper, Timothy Sturgeon received generous financial and intellectual support from ITEC

21st Century COE (Centre of Excellence) Program (Synthetic Studies on Technology, Enterprise and

Competitiveness Project) at the Doshisha University in Kyoto, Japan. This paper has also benefited greatly from the authors’ participation in the discussions and field research carried out by participants in the Globalization Study at MIT’s Industrial Performance Center. We are especially indebted to the dozens of executives at ODM and EMS contract manufacturers who generously took the time to seriously consider and respond to our questions. Suzanne Berger, Jason Dedrick, Martin Kenney, and Greg Linden provided us with detailed and highly useful comments on the manuscript. All errors and omissions, of course, are the responsibility of the authors.

June 2004 ITEC Research Paper 04-03 p.2

Industry Co-Evolution: A Comparison of Taiwan and North America’s Electronics Contract Manufacturers

Timothy J. Sturgeon / Ji-Ren Lee Introduction

The geographic scale and intensity of economic integration has been increasing, if somewhat unsteadily, for many hundreds of years. Without geographic integration, there is autarky and isolation, and so markets and the scale of production remain relatively small, and firms stay largely unaware of and unresponsive to economic activities taking place in other places. Global-scale economic integration has two levels, the first rather shallow, the second deeper. The first is market integration, where firms sell to customers in very distant locations by responding to orders received for a specific good that has been advertised at a specific price. This may put the selling firm in direct competition with firms that are far away. Similarly, a firm selling locally might experience competition from imports. So, with the geographic expansion of markets, global integration is driven forward because firms become aware of and respond to customers and competitors that are distant. The second level of global integration is operations. This assumes a foreign direct investment and a geographically fragmented value chain with different activities scattered across various locations, and some degree of explicit operational control to bind these activities together. For a complex set of reasons that we will explore later, lead firms in the electronics industry have been outsourcing more, that is, turning more to suppliers and less to subsidiaries as they set up global operations. Because geographic fragmentation has increasingly been associated with organizational fragmentation, sales and competition exist not only at the level of markets, but at the level of operations as well, as when suppliers cultivate distant customers and vie with one another for business. This paper explores the dynamic features of global outsourcing and through an analysis of two sets of supplier firms in the electronics industry, contract manufacturers based in Taiwan and in North America.

The electronics industry is extremely global in an operational sense. Companies send, receive, and act on signals from subsidiaries and suppliers in far-flung places on a daily and sometimes hourly basis. While this has been true for 10 to 20 years for activities that deal solely with intangible products, such as currency trading, airline reservations, and voice communications, it is a more recent phenomenon in industries that produce complex tangible goods, such as apparel, motor vehicles, and electronics. In some ways, the low-cost, high-bandwidth voice and data communications network that has grown up to support the global integration in intangible products has laid the


groundwork for the operational integration of tangible products, especially as the network has been extended to places with very low labor costs. This infrastructure may in turn be laying the groundwork for the next wave of global outsourcing, the organizational and geographic fragmentation of service and professional activities. For this reason, we feel it is important to have an intellectual framework in place that will allow us to better characterize and understand how global-scale production systems evolve over time.

Part I: Strategic Outsourcing and Industry Co-Evolution In this section we develop a co-evolutionary framework at the industry level by

highlighting the interplay between increased strategic outsourcing and the rise of shared supplier networks. Recent literature (e.g., McKelvy, 1997; Lewin and Volberda, 1999; Lewin, Long, and Carrol, 1999) has advanced co-evolution as a framework for analyzing the dynamics within, and the interplay between, firm adaptation and environmental selection. Simply put, firm learning—micro-level evolution—both conditions and is conditioned by its context; the matrix of organizations, institutions, and more general social forces within which it is situated—macro-level evolution. While the literature on firm-co-evolution usefully situates firm learning within its broader competitive and institutional contexts, it fails to elaborate the co-evolution of lead firms (buyers) with either individual suppliers or their broader supplier networks. More recently, Nishiguchi (2001) has promoted co-evolution as an ideal framework for the study of inter-organizational relations, especially outsourcing. In Nishiguchi’s view, lead or “convening” firms co-evolve with their suppliers as they oscillate between exploitative and symbiotic inter-firm relationships. Competition, the third leg supporting the evolutionary conceptual stool, is left out of the inter-organizational part of his framework because firms in the same value chain are not portrayed as competing with each other but with discrete value chains convened by other lead firms. Supplier networks are conceptualized, for analytic purposes, as separate and not shared or overlapping. For example, Toyota and its largely symbiotic supplier network competes against General Motors and its largely exploited supplier network. Network evolution occurs as disturbances both internal and external to the symbiotic or exploitative relationships push lead firms and suppliers alike to adapt and innovate. Co-evolution occurs within each network in a process that might best be called value chain co-evolution.

This paper explores a case where increased strategic outsourcing by groups of lead firms has, over time, led to the rise of a shared supplier network, one that can be


accessed by the industry as a whole, even by lead firms that compete head-to-head in final product markets. Our case is the outsourcing of electronics production and the concomitant rise of a pool of highly competent contract manufacturers based in the United States and Taiwan. In this case inter-firm relations tend to be neither exploitative nor symbiotic in Nishiguchi’s sense, but flexible, modular, and self-reliant. We agree that a firm's competencies are powerfully conditioned by the complementary competencies that may or may not reside in the value chains in which it participates, and that the development (or loss) of firm-level competencies exerts a weak effect on the competency of the entire chain. However, we observe that, under certain conditions, groups of lead firms can interact with broadly shared supply-bases to drive the evolution of entire industries in a process we call industry co-evolution.

The emergence of shared supply networks can be understood as a co-evolutionary process encompassing series of strategic choices and interactions between lead firms and their suppliers that occur in the context of a dynamic environment. While such a conceptualization is in keeping with the line of research that adopts a co-evolutionary view of inter-organization relations (e.g., Nishiguchi, 2001; Madhok, 2000), our approach postulates an industry-wide virtuous cycle between lead firm strategic outsourcing and the development of supplier competencies. Suppliers often increase the quality and scope of their services in response to lead firm requests. Once new supplier competencies are in place, they can be used as a basis to develop relationships with other lead firms, and can influence future lead firm decision making regarding strategic outsourcing if, and this is the crucial element, lead firms and suppliers are not locked into dyadic relationships because of asset specificity, equity ties, geographic propinquity, or other factors.

Our conceptualization of the process of industry co-evolution is based on the dynamics of firm capabilities and the characteristics of inter-firm linkages. By contrasting our case with the co-evolution literature that centers on how micro- and macro-level factors enable or limit the firm-level decisions (e.g., Mckelvy, 1997; Lewin and Volberda, 1999), our intent is not to replace micro-level theory or reject the influence of macro-level forces, but to shift the focus to the meso-level processes that operate at the industry level. We view the development of the firm in both its immediate institutional milieu and in the context of its industry, that is, the sets of inter-linked and perhaps competing value chains that serve up specific products to final consumers. In this way, we look beyond the boundaries of the firm to understand the dynamics of firm capabilities, but not so far beyond as to lose sight of the ways in which the decisions made within the firm might in turn impact what the external environment has to offer.


The proposition of transaction value; a dynamic view of transactions

Prior to elaborating our model of industry co-evolution, one premise on the nature of inter-firm collaborations between upstream and downstream firms has to be established. The purpose of engaging in inter-firm collaboration along an industry value chain is to combine two sets of complementary but dissimilar resource configurations in a mutually beneficial manner. The mutual benefits of collaboration are not only those resulting from ex ante resource complementarity but also those that might be generated due to ex post investments in relation-specific assets by either party (Bensaou and Anderson, 1999; Dyer and Singh, 1998) and through inter-organizational learning over time, which in turn may help both firms upgrade their competence (Lee and Chen, 2000). Although there might be risks of increased costs associated with relation-specific investments due to the partner’s possible opportunistic behavior (Williamson, 1975, 1985), the perceived future value generated from interdependent complementarities might, in practice, overwhelm such risks (Celly, Spekman, and Kamauff, 1999; Heide and Miner, 1992). In other words, basing inter-organizational decisions on transaction costs alone could undermine the realization of collaborative benefits and hence the transaction value of inter-firm collaborations (Dyer, 1997; Madhok and Tallman, 1998; Zajac and Olsen, 1993). Concurring with this line of argument, we assert that the ultimate goal of inter-firm collaboration is to maximize the overall value, not merely limit the costs, of inter-firm linkages. Such a value-creation proposition is particularly valid for analyzing inter-organization initiatives from a co-evolutionary perspective because a certain portion of collaboration value can only be realized through the dynamic process of interaction among economic actors (Madhok, 2000, p. 280).

Specific patterns of inter-firm (vertical) relations come from the cumulative strategic choices made by both lead firms and suppliers. Such choices are not random. Each firm bases its decisions on its existing stock of competence, which Levinthal and Myatt (1994) refer to as a firm’s initial conditions of co-evolution. At the same time, the prior actions may constrain a firm’s subsequent decision-making (Arthur, 1989; Teece, Pisano, and Shuen, 1997). A lead firm will outsource as a way to enhance the value-creation opportunities derived from combining its own competence set with those of its partners. Suppliers do not respond to these lead firms initiatives automatically, but within the context of the demands placed on them by other existing customers and


their overall strategic goals. We regard strategic outsourcing as a value-enhancing initiative that seeks to utilize the linkages between two sets of complementary competencies, as well as initiatives aimed at increasing profits by leveraging low-cost inputs. In other words, a lead firm’s strategic choice about competence scope has to be complementary with those of its suppliers before any value can be achieved through collaboration.

Value chain co-evolution from the supplier’s view

In the past decade large vertically integrated firms in a range of industries have pursued strategic outsourcing to remain competitive, especially in highly contested and fast-moving markets (Quinn and Hilmer, 1994; Fine, 1998). By focusing on the competence areas that are essential to, or have the potential for, building competitive advantage, especially in product innovation, marketing, and other activities related to brand development, lead firms have increasingly come to rely upon specialized suppliers for providing best-in-class production services to quickly reap value from innovations while spreading risk in volatile markets (Venkatesan, 1992). The electronics industry is an exemplary case. Firms that develop, market, and sell electronics hardware products, from computers to communications switches to mobile phone handsets, have turned to contract manufacturers for production and even post-architectural design services.１

While it has been widely noted that widespread outsourcing results in the emergence of a de-verticalized industrial landscape (Hitt, Keats, and DeMarie, 1999), little attention has been paid to the evolution of the supply-base that has grown up to supply newly deverticalized lead firms (but see Herrigel and Wittke, forthcoming). This lacuna has led to the erroneous characterization of the deverticalization process as one where industries are evolving toward smaller, highly specialized firms, each of which has shed its “non-core” activities to focus on one or a few “core” competencies. The deverticalization trend looks very different from the supplier’s perspective. To meet the growing demand for full-service, or “turn-key” outsourcing solutions (Sturgeon, 1997, 2002; Gereffi, Humphrey, and Sturgeon, forthcoming), suppliers have in many cases had to add entirely new competence areas, increasing their scope of activities while improving quality, delivery, and cost performance. Increased outsourcing has also, in many instances, vastly increased the scale of suppliers’ operations. Thus, increased outsourcing has led to a deepening of competence scope and an increase in scale at supplier firms. As supply-bases come to be comprised of large, highly capable suppliers the prospects for increased outsourcing are improved.


In this way, suppliers and lead firms co-evolve in a recursive cycle of outsourcing and increasing supply-base capability and scale, which makes the prospects for additional outsourcing more attractive, not just to the lead firms that drove the upgrading of the supply base in the first instance, but for those lead firms just beginning to seriously consider large scale strategic outsourcing.

Too little attention has been paid to suppliers in the realm of theory as well. The central question of the transaction costs framework (Williamson, 1975, 1985) is whether lead firms choose to make or to buy a particular product, process, or service. The key analytic variable is the degree of asset specificity embodied in the inter-firm relationship. Suppliers are portrayed, largely, as either investing in specific assets or not. When asset specificity is present, lead firms will become locked into using particular suppliers, creating the risk of opportunism and high transaction costs. But highly competent contract manufacturers tend to mix specific and general assets. Instead of viewing supplier characteristics as a dyad consisting of suppliers that do and do not invest in specific assets (what we call commodity and captive suppliers in Figure 1) we see a third option: suppliers that mix specific and generic assets in ways that mitigate risk and make greater use of fixed capital (what we call a turn-key supplier in Figure 1). Specifically, specific assets are used for customer linkages, such as decoding requirements and responding to requests, while general assets are shared across a broad customer base and accessed via a set of standard protocols.


Figure 1. Comparison of Dyadic Supplier Archetypes with the Turn-key Supplier: General vs. Specific Assets

The turn-key supplier

Customer

interfaces

r

General Assets (shared resources)

Specific Assets(dedicated resources)

The commodity supplier

Critics of the transaction costs

costs matter (such as transaction valtrust) to problems of opportunism repeated interaction is taken into Johanson and Mattson, 1987; Poweagree with these critics, our view investing in specific assets, einter-organizational trust can be drelationships and make supplier invproducts an unwise proposition sinchange with breathtaking rapidity. over time, been forced to developpragmatism in the face of such volaand Wittke, forthcoming).

A key factor facilitating the gachieve greater flexibility and lowemanufacturing facilities. Such c

June 2004 ITEC Rese

The captive supplie

framework argue that factors more than transaction ue, see above), or that there are other solutions (e.g. than vertical integration when the long view of

consideration (Granovetter, 1985; Thorelli, 1986; ll, 1987; Jarillo, 1988; Lorenz, 1988). While we is that there are other important reasons to avoid specially in volatile markets. Even where

eveloped, market volatility can disrupt long-term estment in assets specific to any single lead firm’s ce both absolute and relative market positions can In many industries, lead firms and suppliers have, mechanisms to manage their relationships with tility (Helper, Macduffie, and Sabel 2000; Herrigel

rowth of outsourcing is the lead firm’s desire to r risk by reducing fixed investments, especially in onsiderations are especially important when

arch Paper 04-03 p.9

technological change is rapid and market volatility is high, contingencies that make effective capacity planning and efficient capacity utilization of in-house or externalized specific fixed assets extremely difficult. Lead firms use suppliers, then, to ramp output up and down according to changes in demand. For their part, suppliers are able to achieve relatively stable demand profiles, high capacity utilization rates, and low costs by pooling demand from a large number of customers provided that process technology depends not on specific assets, but on generic assets, assets that are widely applicable to the entire customer pool, and provided that suppliers have the capacity to capture complex requirements from multiple customers and utilize these data in a format usable by their generic processes. Value chains with such “plug-and-play” characteristics have been referred to as “modular value chains” (Baldwin and Clark, 2000; Schilling and Steemstra, 2001; Sturgeon, 2002; Langlois, 2003; Gereffi et al, forthcoming).

Because complementary competencies form the basis for much inter-firm collaboration, co-evolutionary processes are driven by each firm’s managerial discretion, and the performance feedback of existing inter-organization relations. But conditions external to the bi-lateral relationship matter as well. For example, frequent technological changes increase the cost of capital investment and competitive pressure in end markets may force lead firms focus on critical competency areas, the complexity and cost of which furthers the drive toward vertical specialization and a horizontal industry structure. When rapid technological change combines with competitive pressure in this way time-to-market becomes a key survival (not success) factor (Curry and Kenney, 1999). Yoffie (1997) highlights the critical nature of both technological change and competitive pressure in driving many electronics-related industries to toward a horizontal industry configuration, where competition occurs among firms specialized in specific stages of the value chain. These two external factors also serve as an impetus for lead firms to focus on specific areas of expertise and rely on suppliers for others. In sum, the lead firm’s response to a changing environment induces suppliers to enhance their competence scope, which in turn reinforces the drift toward horizontal industry structure.

The lead firm’s increasing reliance on the competence of its suppliers will often be balanced by an effort to exert some degree of control over inter-firm transactions. Specifically, any conflict of interest resulting from the supplier’s aggressive expansion of competence scope into areas deemed core to the lead firm will weaken collaborative activities. This potential damage to future business can in turn become a powerful feedback to the supplier’s competence development. In addition to the use of asymmetric competence levels to exercise external control, a lead firm can implement either multiple sourcing (e.g., McMillan, 1990; Porter, 1980) or parallel sourcing (e.g.,


Richardson, 1993) to enhance its bargaining power over suppliers. In the face of a lead firm’s desire to maintain both flexibility and control, supplier

management is often forced to focus on upgrading the firms manufacturing competence. However, there are risks associated with continuous investments in manufacturing, especially if they entail relation-specific investments and lead to long-term captive supply relationships. However, modular products and processes provide suppliers with opportunities to reduce asset specificity while maintaining the capability to offer flexible manufacturing services by putting in place generic capabilities. The development of generic manufacturing capabilities can expand a supplier’s customer base and open up new application areas. In other words, a supplier can reduce risk and expand markets through the development of a multiple-customer and/or multiple-business profile.

Let us discuss the tripartite characterization of suppliers in Figure 1 in more detail. Commodity suppliers depend on generalized assets (and often produce standard products), and because they connect to customers largely via arms-length market transactions, only a very thin customer interface is needed, such as a telephone order or mailed, faxed, or on-line order forms. Since price is the key factor structuring outsourcing decisions, and thin linkages and standardized products make supplier switching easy, commodity suppliers are often found in what Nishiguchi (2001) refers to as exploitative supplier networks. Captive suppliers depend on dedicated assets, and in extreme form would serve only one customer and therefore have a single, extremely thick customer interface optimized for high levels of collaboration. There is general agreement in the literature that parties to transactions in networks comprised of captive suppliers would be protected from opportunism by reliance on long-term relationships, trust, equity ties, or some other form of thick linkage (Richardson, 1972; Thorelli, 1986; Johanson and Matsson, 1987; Lorenz, 1988, 1992; Jarillo, 1988; Bradach and Eccles, 1989; Powell, 1987, 1990). In Nishiguchi’s framework, captive suppliers would tend to be found within symbiotic supplier networks.

In our view, this dyadic view of suppliers is misleading. In practice, many suppliers seek to limit interdependence with customers and diversify their business profiles, while at the same time building up generalized assets and thick customer interfaces relative to the commodity supplier archetype. We refer to such suppliers as turn-key suppliers (Sturgeon 1997, 2000) because of their relatively independent stance toward their customers, their high level of competence, and their leveraging of generalized assets to serve multiple-customers and/or multiple-businesses. The term is meant to highlight the relatively limited need for interaction and instruction prior to collaboration. A core competence of turn-key suppliers is an ability to use their


generalized assets to serve the very specific needs of their customers, which are understood and monitored via a customer interface that is relatively thin compared to the captive supplier and relatively thick in comparison to the commodity supplier. The customer interface at a turn-key supplier might be comprised, for example, of customer-specific program and product management teams. Turn-key suppliers represent more than a point on a continuum between commodity and captive suppliers. Their existence can change the character of entire industries by creating supply bases that are shared by a wide range of lead firms while allowing a great deal of non-price data to flow across the inter-firm linkage. The core competence of turn-key suppliers is the translation of diverse customer needs into a standardized format that can be easily fed into a set of generic processes. A further competence of turn-key suppliers is to scan groups of customer requests for opportunities to pool component and material purchases to low input costs. Turn-key suppliers, therefore, have four critical competency areas: 1) customer service, 2) the ability to create standardized instructions from diverse customer needs, 3) the creation and maintenance of generic capacity, and 4) the ability to scan customer requests for pooling opportunities.

A multiple-customer/multiple-business customer profile combined with a core of generalized assets not only reduces a supplier’s risk of fixed investment under-utilization but also enables industry-wide learning opportunities. Competence building through learning from partners has been indicated as one of the strategic goals of inter-organizational collaborations in general (e.g., Hamel, 1991; Lei and Slocum, 1992; Khanna, Gulati, and Nohria, 1998; Madhok, 2000), and vertical relations in particular (e.g., Bettis, Bradley, and Hamel, 1992; Lee and Chen, 2000). Depending upon the strategic intent (Hamel, 1991) and absorptive capabilities (Cohen and Levinthal, 1990) of respective partners, both explicit and embedded knowledge can flow through collaborative initiatives and be internalized by partner firms. In the case of vertical relationships, turn-key suppliers can acquire, among other things, market information, design concepts, technical specification, quality standards, and process parameters through working with customers with a higher level of competence. Learning can flow from suppliers to lead firms as well.

Two of the key pitfalls of strategic outsourcing are the creation of new competitors via supplier upgrading and the loss of critical resources through suppliers that might be shared with competitors. There is substantial research that highlights the potential threat to the buying firm from suppliers’ aggressive learning and subsequent entry into the buyer’s business (e.g., Bettis, et al., 1992; Makides and Berg, 1988). However, we argue that the common benefits that accrue to each partner by developing a shared supply network will be no less, and often greater, than the benefits the lead firm or


supplier could realize through bi-lateral relationships (Khanna, et al., 1998). The spread of knowledge about standards and general best practices, for example, can help push entire industries toward greater efficiency.

Similarly, because core assets are generic, turn-key suppliers can leverage knowledge gained from one customer to better serve others with less risk of compromising core intellectual property, typically product design and pricing data. The risks of intellectual property leakage through a shared supply-base are further reduced when product life cycles are extremely short, as they have become in a broad range of sectors, because the criticality of product data is extremely ephemeral. This it not to say that such risks are not present and do not influence the decision-making of lead firms. Intellectual property, customer data, pricing information, and market knowledge can and do leak though shared suppliers, and suppliers themselves can seek to capture this knowledge with the intention of eventually competing with their customers. The importance of trust as a governing mechanism is reduced, but not by any means eliminated in modular value chains. Judgments about trust have affected the co-evolution of the two groups of contract manufacturers we analyze in section two of this paper, for example.

Beyond the industry- and firm-level factors underlying co-evolutionary processes, different institutional arrangements in which firms are embedded will affect strategic choices concerning competence scope and inter-firm linkages. Institutional factors serve both as facilitating and restricting forces on a firm's strategic and organizational adaptation (Lewin, et al., 1999). Among others, geographic proximity matters for facilitating resource exchanges and the fulfillment of customization needs (Porter, 1998). Further, the differential capacities of capital markets could limit a firm’s strategic options in constructing new inter-organization relationships through acquisitions, geographic expansion, or both.

Industry co-evolution

The key micro-level features that allow turn-key suppliers to become accessible to a broad customer base are the relatively high level of codification in inter-firm transactions and the generic character of the products and services provided to lead firms. At the macro-level, the system requires widely accepted standards to describe product features and components, and process technology that resists the build-up of specific assets and thus remains effectively generic. But the existence of generic production technology, often embedded in sophisticated production machinery, and a broadly recognized set of standards for translating product data onto production data do


not make the development of generic capacity unproblematic for the firm. Adopting and keeping abreast of fast-moving standards often means significant competence enhancement. Likewise the process of choosing, buying, maintaining, staffing, and upgrading production equipment that provides a supplier with generic capacity can be very difficult and costly.

At the meso-level, which is the level we would like to emphasize here, the emergence of significant numbers of turn-key suppliers can create capacity pooling effects, in which lead firms can easily connect to—and disconnect from—the production and service capacities that reside externally in the supply-base. The existence of high bandwidth inter-firm linkage mechanisms allow for the build up of specific assets to the degree that inter-organizational collaboration and learning can take place (creating transaction value), while the generic nature of the core products and services provided puts limits on the build-up of asset specificity and keeps the threat of opportunism from hindering inter-firm collaboration because the barriers to exit are relatively low for both parties. Pooling effects and low barriers to relationship exit are created by the generic nature of the products and services provided and the multiple-customer/multiple-business profile of turn-key suppliers.

Within modular value chains the forces of symbiosis and exploitation are so entwined as to render the dualism less than useful. Furthermore, competition becomes part of the analysis because lead firms can compete head to head while sharing the same set of suppliers and in some cases, even the same facilities and production lines. We argue that shared supply-bases, as they come into existence, exert profound effects on entry conditions, the nature of subsequent competition, and future outsourcing decisions. These effects are distinct from those that can be expected to arise from cases where buyer-supplier production networks are discrete and evolve along classically exploitative or symbiotic lines. Shared supply bases, as in the modular production network, tend to generate a self-reinforcing dynamic—a classic network effect (Katz and Shapiro 1985, Farrell and Saloner 1985, Arthur 1996)—because pooling effects create large external economies of scale and scope and powerful learning effects that induce an increasing number of lead firms to tap the network, which in turn further enhances the competence, scale, and scope of the turn-key supply-base and induces more lead firms to participate. Thus a historical process is unleashed that is one of industry co-evolution, where the development of an external supply-base through outsourcing encourages further outsourcing, and so on until a shared supply-base emerges. Figure 2 compares an industry with discrete, competing supply-bases to an industry with a shared supply base.


Figure 2. Co-Evolution with Competing vs. Shared Supply Bases

A) Value chain co-evolution: discrete, competing, supply bases connected to specific lead firms

Competition

A B

n

Supply base A Supply base B

B) Industry co-evolution: A shared, co-evolving, supply-base connected to multiple lead firms

Co-evolution and

Competition

Lead firm” n” Lead firm A

n

Part II.United S

In examplecan be ucomparidifferenthereaftehereaftemanufac

June 200

Co-evolutio

Co-evolutio

The co-evolution of etates this section we movs of the process of insed to understand th

ng the leading electr institutional environr) based largely in r) firms based largelturers have played

4 ITE

Lead firm

Shared supply base

lectronics contract manufacturing

e from the abstract to the specific dustry co-evolution. We demonstrae dynamics of competition within a onics contract manufacturers that haments, namely “original design ma

Taiwan, and “electronic manufactury in the North America. These tvital roles in creating the tightly

C Research Paper 04-03

Lead firm

in Taiwan and the

with two real-world te how our approach specific industry by ve emerged in two

nufacturers” (ODM, ing service” (EMS, wo sets of contract

integrated external

p.15

production capacity that worked to alter the dynamics of competition, organization and location in electronics and computer-related industries during the 1990s. We make our comparison of ODM and EMS contract manufacturers through a brief account of their origins, development, current position in the industry. We argue that different starting points and strategic foci have led to each group of firms varying in specific ways along three dimensions of competence scope: value chain scope, product/customer scope, and geographic scope. We then discuss some recent trends that have worked to intensify competition between ODM and EMS contract manufacturers.

ODM Contract Manufacturers

Beginning in the early 1990s, the term ODM came into common usage in Taiwan to indicate a specific type of contract manufacturing service where suppliers provide design services along with volume production. The term was used to differentiate this service from manufacturing-only services, which had long been referred to as “original equipment manufacturing,” or “OEM.” ２ The emergence of ODM contract manufacturing was enabled by the high degree of modularity and standardization in personal computer product architecture and driven by very short product life cycles, which generate a need for a constant stream of design activities to bring new products to market. These conditions helped to create a market for the iterative, post-architectural design services that ODM contract manufacturers provide (Lee and Chen, 2000). Riding on surging demand, ODM contract manufacturers have become a major presence in the design and production of personal computers and related devices, including desktop personal computers, notebook personal computers, motherboards, and peripherals. Table 1 shows the value, volume and share of global manufacturing of various personal computer and peripheral devices produced by Taiwan companies in 2002 according to the Market Intelligence Center (MIC), a Taipei-based market research organization founded—and still in part funded—by the Taiwan government. In addition to personal computers, Taiwan firms produce a high share of the world output of many electronic sub-systems that find their way into or are connected to personal computers, such as analog modems (60.2%), cable modems (65.6%), DSL modems (51%), network interface cards (64.3%), simple local area network hubs (81.2%), switched power supplies (77%), scanners (91.2%), as well as computer keyboards and mice (MIC, 2003).


Table 1. IT Hardware Produced by Taiwan Firms, 2002

Product

Production value 2002

($M)

Production volume 2002

(units x 1,000)

Average value per unit

($)

Estimated share of global

production volume, 2002

Notebook PCs $13,922 18,199 $764.99 61%Desktop PCs $6,933 24,740 $280.23 23%LCD Monitors $5,646 18,254 $309.30 61%Motherboards $5,636 86,511 $65.15 75%CRT Monitors $4,544 42,910 $105.90 51%Optical Disk Drives $3,146 79,409 $39.62 45%Servers $1,324 1,485 $891.58 30%Total $41,151 271,508

Source: MIC, 2002 IT Industry Yearbook, December 2003.

A complete understanding of ODM contract manufacturing requires a broader view of Taiwan’s electronics industry, which includes a very strong components sector. In 2002 Taiwan firms, including industry leaders Taiwan Semiconductor Manufacturing and United Microelectronics, accounted for 72.5% of the world’s integrated circuit foundry services. Taiwan firms also provided 32% of the world’s integrated circuit packaging services and 38% of integrated circuit testing services (MIC, 2003). These service providers, essentially contract manufacturers for the integrated circuit sector, have in turn enabled the development of more than 200 integrated circuit design companies in Taiwan, known as “fabless IC design houses,” a number second only to the United States (ITRIEK, 2003). While there are significant personal computer components, sub-systems, and peripheral devices in which Taiwan firms are not active, namely software, printers, hard disk drives, and higher-value semiconductors such as microprocessors and memory, the sum of the capabilities in Taiwan comprises a powerful, agile supply-base for the design, manufacture, and delivery of personal computers and related products. Working in close geographic proximity, mostly along the Taipei-Hsinchu corridor in Taiwan, this supply-base grew to constitute an extremely efficient system that allowed ODMs to respond very rapidly to orders from lead firms (Dedrick and Kraemer, 1998). Notebook computers, for example, which generally have a high enough value to weight ratio to make air shipment viable, can be shipped to customers within two to three days of incoming orders.


This powerful productive engine has developed, almost in its entirety, in response to orders from lead firms based in other societies. At the same time, the development of contract manufacturing in Taiwan and elsewhere has provided lead firms with an increasing range of sourcing options. This process of co-evolution has meant that Taiwan’s electronics industry has been able to develop without a significant cadre of local lead firms. From the late 1970s to the present day, sourcing from Taiwan has expanded from computer monitors, to various components and sub-systems, to complete desktop and notebook personal computer and peripheral systems. Firms from the United States computer industry have played an especially important role. In the early 1980s IBM began sourcing personal computer monitors from television and television tube producers in Taiwan, including Tatung and Chung Hwa. As the demand for personal computers expanded rapidly and the open architecture of IBM-compatible PC became firmly established in 1984 with the IBM model AT, some entrepreneurial firms in Taiwan, such as Acer and Mitac, recognized the opportunities and moved aggressively to develop the capability to design personal computers and peripheral devices based on the emerging standard. IBM’s modular system architecture relied on a central processing unit supplied by Intel and an operating system from Microsoft, and because the contracts famously did not block Intel and Microsoft from selling to IBM’s competitors, a bevy of new entrants, intense price competition, and a series of boom and bust cycles soon followed. These conditions caused contract manufacturing to become a popular strategy for lead firms in the personal computer industry seeking to cut costs and reduce risk by limiting investments in fixed capital. In the late 1980s, a set of firms that had been focused on the design and manufacture of hand-held calculators entered the field of personal computer contract manufacturing. These firms, which include Quanta, Compal, and Inventec, eventually became the dominant notebook computer ODMs, in part because the design and assembly competencies that drove miniaturization in calculators were well suited to notebook computers, where small size, low weight, and efficient power consumption are key factors for success in the marketplace and in part because calculators, while much simpler, are similar to personal computers in that they are built around a standard central processing unit.

The modular system architecture of the IBM personal computer also stimulated specialized components firms in Taiwan, such as Hon Hai (Foxconn), Delta, and Asustek, to develop components and sub-systems compatible with the new standard, such as connectors, power supplies and motherboards. (Of these three firms, Hon Hai and Asustek have recently won large OEM orders for complete personal computer and other systems.) These moves were part of a larger worldwide trend in the computer


industry away from vertical integration and toward a horizontally organized industry comprised of specialized sub-industries. As Baldwin and Clark (2000) show, the modularity of computer architecture can be traced back to the introduction of the System 360 mainframe computer in the 1970s, and to several influential texts that promoted modularity in computer system design. Modularity effectively “marketized” various computer elements and enabled the emergence of a wide range of sub-industries to provide industry-standard components and sub-systems such as memory, storage, and displays. The openness of IBM personal computer architecture and the popularity of the machines served to accelerate this trend dramatically.

The modular system architecture of personal computers, and the dominant role of the central processing unit and software operating system in setting system architecture, along with intense competition and short product life cycles, created the conditions for the emergence of a set of firms to specialize in the iterative, post-architectural portions of product design, including the board-level operating system (BIOS), which determines how the machine handles the input and output from its main board to the other elements of the systems, such as storage and displays; and industrial design, which determines the physical appearance of the product. However, because most functionality resides in chips sets and software—things that ODMs do not design—control over the innovative trajectory of the industry has continued to reside in other firms and other societies.

Which companies are the largest buyers of electronics hardware from Taiwan? Table 2 shows the largest five buyers of IT hardware from Taiwan firms in 2003, and the value of these firms’ purchases back to 1998. IT hardware includes finished products of all kinds, but not components. The top five buyers are all personal computer companies, but the top ten list of American buyers includes non-personal computer companies such as Cisco (communications switches), Motorola (mobile phone handsets), Viewsonic (monitors), Solectron, a U.S.-based contract manufacturer that sources some of its personal computer motherboards from Taiwan firms, and Intel, which also sources motherboards in Taiwan. This broadening product mix suggests that the long-standing efforts by Taiwan contract manufacturers to move beyond the personal computer industry are beginning to bear fruit. Aside from Sony, the most important Japanese firms sourcing IT hardware—mostly personal computers—from Taiwan ODMs are, alphabetically, Fujitsu, Hitachi, NEC, and Toshiba. The most important European firms, again in alphabetical order, are Actebis, a distributor that assembles and configures personal computers, Fujitsu/Siemens (personal computers), Nokia, Philips, and Siemens (mobile phone handsets).

By comparing the value of foreign buyers’ purchases in 2002 to the total value of IT hardware produced by Taiwan firms that year, we can estimate that contract


manufacturing accounted for 91% of the value of all IT hardware produced by Taiwan firms, leaving only 9% accounted for by the sales of own-brand products, such as Acer’s line of branded personal computers. Taiwan’s electronics industry, then, can be said to be a clear example of “supplier-oriented industrial upgrading” (Sturgeon and Lester, 2004). Taiwan’s ODM contract manufacturers are highly dependent on a small number of buyers. The top five buyers listed in Table 2 accounted for 70 percent of all the information technology hardware produced by Taiwan firms in 2002, a figure that is estimated to have increase to 71% in 2003 (authors’ calculations based on MIC 2003).

Table 2. Top Five Buyers of IT* Hardware from Taiwan, 1998-2003, billions of USD

1998 1999 2000 2001 2002 2003** 2003 ShareHP/Compaq 7.1 10.0 14.5 14.4 15.0 15.0 34%DELL 1.9 2.8 4.5 5.5 6.5 7.0 16%Sony 0.0 0.0 0.0 0.0 3.5 4.0 9%Apple 0.8 1.0 2.5 2.5 2.8 3.0 7%IBM 2.1 3.0 3.5 3.0 2.5 2.5 6%Top 5 11.9 16.8 25.0 25.4 30.3 31.5 71%Others 3.7 4.2 7.6 8.2 13.0 13.0 29%Total 15.6 21.0 32.6 33.6 43.3 44.5 100%Source: authors’ calculations from MIC 2003. *IT hardware includes finished products, but not components. ** MIC estimate.

The story of lead firm-ODM co-evolution that we have outlined here reveals a powerful dynamic of outsourcing, upgrading, and subsequent outsourcing; the enabling role of open standards and modular product architecture; intense competition and rapid product life cycles that drove lead firms to seek spread risk and lower costs through outsourcing; as well as the entrepreneurial agility that many firms in Taiwan displayed in shifting to export production by recognizing and seizing new opportunities to specialize in narrow segments of the value chain. But there were two other important factors that have not yet been discussed. The first is the Taiwan government, which helped by licensing, refining, and disseminating foreign technology and encouraging, and in some cases underwriting, the entry of local firms into promising market areas, especially for IT-related products. The second is the role of Japanese technology partners, which provided critical technology that nevertheless came with little support


and restrictions that inhibited Taiwan firms from building on the technology to develop independent product development strategies. These sort of licensing agreements continue to be important today, in Taiwan’s emerging flat panel display industry, for example (see Akinwande et al, this volume).

How has this development path positioned Taiwan’s electronics industry within the global IT industry? In 2002 Taiwanese ODM firms supplied approximately 31% of all the personal computers sold worldwide, but since these machines tend to be models within the low to medium-price range, this output only accounted for about 13% of the value of sales.３ From these data, we can see that there are large swaths of the worldwide electronics industry and even the personal computer industry—especially higher value segments—where Taiwan firms have very little presence. Nevertheless, it is abundantly clear that Taiwan electronics firms have become an indispensable part of the global electronics industry, a remarkable achievement for a small island with an economy that was based largely on agriculture only 30 years ago.

The concentration of ODM on the personal computer industry, and dependence on several extremely large buyers such as Dell and Hewlett Packard/Compaq, comes with several potential threats and challenges to future prosperity. First, the constant erosion that has occurred in personal computer selling prices has placed a great deal of pressure on ODMs to lower costs. This pressure was greatly heightened in 1997 when Compaq decided to develop a line of personal computers that could retail for less than $1,000. Dell and other lead firms now market desktop computer systems for less than $500. The degree of price erosion in personal computers can be determined by comparing the production volume to value in the three largest product categories shown in Table 1. According to MIC, the total volume of notebook computer, desktop computer, and motherboards produced by Taiwan firms grew by 29.8%, 8.4%, and 7.4% from 2001 to 2002, while the value of these products only grew by 13.1%, 1.6%, and -0.2% respectively (MIC, 2003). Intense price pressure has triggered a huge capacity expansion in Mainland China where operating costs are lower.

The second risk has to do with the shrinking share of personal computers in worldwide IT spending. After exploding onto the scene in 1984 with widespread adoption of the IBM model AT, the personal computer share of total worldwide IT４ spending peaked in 1996 at 23%. Since then, the applications for IT products and services have continued to proliferate with the Internet, wireless and digital data communications infrastructure equipment, and the like. By 2003, sales of personal computers had dropped to just 18% of the value of world IT spending (IDC, 2004). In relation to total spending on data processing hardware, data communications, and computer peripherals, personal computers accounted for about 47% in 2001, a figure


that was expected to decline to 44% in 2003 (IDC, 2002). These risks help to explain why ODMs are urgently seeking to diversify their

product and customer scope beyond personal computer-related products. As already mentioned, Asustek, a leader in motherboard design and production, has successfully diversified into notebook computers, VGA cards, optical disk drives, and the manufacturing of the PS2 video game console for Sony. Revenues have increased dramatically while motherboard sales have dropped to only one third of sales. ODM manufacturers such as Hon Hai (Foxconn), Quanta, Compal, Asustek have all allocated significant resources to the design and production of mobile phone handsets, another product segment where iterative post-architectural design around a standard chip set and operating system and miniaturization are critical success factors. However, because neither detailed design competencies nor the personal computer-specific supply base in Taiwan are easily adaptable to electronics products beyond personal computers, product and customer diversification has been slow and the ODMs remain quite narrowly focused on personal computers. Still, except for the near monopolies held by Intel and Microsoft in central processing units and operating software, severe downward pressure on prices and margins is felt at all levels of the personal computer value chain: lead firms, contract manufacturers, and component suppliers alike.

In addition, because of the high degree of design modularity in personal computers, activities such as software, microprocessor, BIOS, and industrial design can be carried out with a large measure of independence, which lends itself to both the organizational (outsourcing) and geographic fragmentation that we see in the industry. In other words, various elements of the personal computer value chain can quite easily function within different firms located at great geographic distance from one another. This does not appear to be true for other segments of the electronics industry, especially for higher value segments, and the geographic concentration of ODM design activities in East Asia and the lack of design and new product introduction facilities in Europe and the United States may be a disadvantage (Sturgeon and Lester, 2004). And clearly, the technical and strategic barriers to outsourcing the design of extremely complex, high-value, and often non-standard products such as large computers, sophisticated communications switches, and medical devices remain extremely high.

The other strategy that ODMs have pursued to increase their leverage in the industry and profit margin has been to develop their own-brand products. As new product categories emerge there is a constant searching at many ODMs for opportunities to develop their own brands, as with new product categories such as “Internet appliances” that are just beginning to be established. Clearly, if managed successfully, maintaining both contract manufacturing and own-brand businesses can improve a


company’s financial performance (Lee, Chen, and Tang, forthcoming). For example Asustek and Gigabyte sell their own brands of motherboard, especially in Europe, where custom assembled, or “clone” personal computers are popular. However, ODM efforts to develop their own-brand businesses have inevitably created competitive tension with their major buyers. For example, Acer generates about 30-40% of its revenues from its own personal computer brand, but has historically struggled with combining its branded personal computer and contract manufacturing businesses (Sturgeon and Lester, 2004). To alleviate this tension, Acer has undergone a series of moves to separate the contract manufacturing and branded segments of its business, most recently by spinning off its contract manufacturing arm (as Wistron) and turning to other Taiwan ODMs (including the leading notebook computer ODM, Quanta) as a source for notebook computers. In 1993 Quanta surpassed Wistron as Acer’s largest supplier of notebook personal computers (Tzeng and Chang 2003). The market share of Acer brand of personal computers has recently improved. The company is currently the fifth largest seller of personal computers overall and the second largest seller of notebook computers in Europe. A common strategy is for ODMs to focus their own-brand strategies on serving markets that are of little interest to their customers, such Chinese-language markets in East Asia. But recent efforts by lead firms in the personal computer and mobile phone industries to capture the growing market on Mainland China have made it clear that such markets are indeed of great interest to the customers of ODMs. And so, by and large, ODM contract manufacturers have been forced by their customers to remain focused on detailed design and production only.

EMS Contract Manufacturers

EMS contract manufacturers exist in many places throughout the world, but have over time come to be dominated by the largest five firms, all based in North America, namely Solectron, Flextronics, Sanmina-SCI, Celestica, and Jabil Circuit. Solectron, Flextronics５, and Sanmina-SCI are based in Silicon Valley, California; while Celestica are based in Toronto, Canada; and and Jabil in St. Petersburg, Florida. These firms have heterogeneous origins; some began as contractors to the US space program in the 1970s, one was a contract assembler for mainframe computer and automotive electronics customers in the American Midwest, and still others originated as contract assemblers for the early personal computer and disk drive industries in Silicon Valley. All of the firms that have ultimately come to dominate the industry have placed an emphasis on quality, and several founders held an early fascination with systematic quality improvement schemes and Japanese production methods (Sturgeon, 1999).


These firms were, in effect, repositories of leading edge electronics manufacturing expertise in a society that in the late 1970s and 1980s seemed to be falling behind in manufacturing.

In the late 1980s and early 1990s, leading EMS firms grew by serving a set of lead firms increasingly focused on product innovation and decreasingly focused on manufacturing. As in the case of the ODMs discussed above, growth in the early days of the personal computer industry drove much of this expansion. Solectron provided manufacturing services to a variety of early personal computer companies, including Apple and Osborne. SCI Systems (now part of Sanmina-SCI) was selected to provide main-board assembly for the original IBM personal computer in 1981, and grew to dominate the early EMS industry as demand for IBM’s personal computer products mushroomed. In the latter half of the 1980s, uncertainty created by moves into new, automated production technologies (called surface mount technology, or SMT) and a devastating boom-bust cycle in the personal computer industry caused lead firms to become hesitant about making new investments in manufacturing capacity, and so they increased their orders from contract manufacturers, both in the United States and overseas. Pooling effects allowed contractors to better amortize rising investments in expensive SMT production equipment (Stanford Case Study, 2001). But EMS firms too had been stung by the sudden drop in demand for personal computers and disk drives, and so began a long and sustained effort to diversify their customer base, adding customers in industrial and medical electronics, and later, communications and data networking gear. As personal computer production shifted to Taiwan firms, the EMS firms also moved up-market within the computer industry to workstations, servers, and mainframes.

As late as the early 1990s contract manufacturers were still perceived in the United States as low-cost sweatshops, or “circuit board stuffers,” with few capabilities and little utility beyond providing supplemental capacity in times of peak demand. All this changed—and changed quickly—beginning in 1991 when Solectron, then the number two EMS firm, won a prestigious national quality award (the Malcom Baldrige Award) and the leading firm at the time, SCI Systems, broke the one billion dollar revenue mark. Both events did much to legitimate the EMS sector in the eyes of existing and potential customers, and a few pioneering lead firms, including IBM and Hewlett Packard, moved to sell off their in-house facilities to leading EMS firms. Thereafter the floodgates opened and business poured into the top EMS firms at a breakneck pace. By the late 1990s, lead firms moved consolidate their proliferating global outsourcing relationships by awarding global manufacturing responsibility and selling worldwide manufacturing plants, if they had any left, to the largest and most


capable EMS firms. Figure 3 shows that revenue growth at the top twenty EMS contract manufacturers between 1999 and 2002 was largely contained within the largest five firms.

Figure 3. Revenues at the Top Twenty EMS Contract Manufacturers, 1999 and 2002, thousands of US dollars.

$0

$2,000

$4,000

$6,000

$8,000

$10,000

$12,000

$14,000

$16,000

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20

2002 1999

Source: Electronic Business Top 100 Contract Manufacturers, various years.

The largest five EMS firms completed 50 acquisitions of competitor and customer facilities between 1995 and 2002, which, along with the establishment of many new facilities, gave them global operations. Capital for expansion was provided by huge share price increases that came in the stock market run up in the late 1990s, and international management capabilities were drawn from the deep pool of expertise resident in the United States, Europe, Singapore, and Australia. Solectron, for example, expanded its locations from a single campus in Silicon Valley in 1991 to an astonishing 135 by 2001. Seventy of these were in North America and 65 were located outside the continent, including West and East Europe, Asia, and Brazil. Table 3 compares the revenues, employment, facilities, and locations of the largest five EMS contract manufacturers to the largest 100. The data in Table 3 is drawn from an annual survey conducted by Electronic Business Magazine, and is heavily biased toward firms based in the United States; very few Taiwan firms—and none of the ODM contract manufacturers discussed in the previous section—are included. In 2002 the largest five firms in the survey accounted for approximately 70% of the revenues, 63% of the employment, 69% of the total facilities, and 82% of the facilities outside of North


America of the largest 100 EMS firms in the survey. Table 3. Comparison of Top Five to Top 100 EMS Contract Manufacturers; Revenues, Employment, and Facilities, and Location, 1999 and 2002; Compound Annual Growth Rate 1999-2002; and Top Five Share of Top 100, 2002

Revenues ($M) 1999 2002 CAGR ’99-‘02 Share of Top 100, 2002Flextronics $1,808 $13,615 96% 20%Solectron $8,391 $12,261 13% 18%Sanmina-SCI $8,624 $10,168 6% 15%Celestica $5,297 $8,272 16% 12%Jabil Circuit $2,400 $3,729 16% 5%Top 5 $26,520 $48,045 22% 70%Top 100 $46,029 $68,149 14% 100% Employment

1999 2002 CAGR ’99-‘02 Share of Top 100, 2002 Flextronics 18,147 95,000 74% 21%Solectron 37,963 73,000 24% 16%Sanmina-SCI 37,470 46,030 7% 10%Celestica 18,000 40,000 30% 9%Jabil Circuit 12,000 26,000 29% 6%Top 5 123,580 280,030 31% 63%Top 100 262,938 446,386 19% 100% Total Facilities 1999 2002 CAGR ’99-‘02 Share of Top 100, 2002Flextronics* 75 110 47% 18%Solectron** 52 135 61% 22%Sanmina-SCI 64 100 16% 17%Celestica 32 44 11% 7%Jabil Circuit 21 31 14% 5%Top 5 244 420 20% 69%Top 100 719 605 -6% 100%


Facilities Outside North America 1999 2002 CAGR ’99-‘02 Share of Top 100, 2002Flextronics* 55 75 36% 24%Solectron** 31 65 45% 21%Sanmina-SCI 21 63 44% 20%Celestica 15 30 26% 10%Jabil Circuit 9 24 39% 8%Top 5 131 257 25% 82%Top 100 295 313 2% 100%Source: Electronic Business Top 100 Contract Manufacturers, various years. *Flextronics facility figures are for 2000; growth rates have been adjusted accordingly.

**Solectron facility figures are for 2001; growth rates have been adjusted accordingly.

Unlike the ODM contractors, EMS firms have, historically, provided few design services. While this has kept their profit margins at very low levels, it has been easier for them to diversify their customer base (and increase revenues) because—unlike design—manufacturing processes have remained relatively applicable to a wide variety of electronics product categories. Still, there are a variety of new competencies that EMS firms have been struggling to develop related to their new, global operational footprint, such as global inventory management, logistics, and capacity planning. In the 1990s, while ODMs were increasing their capabilities in notebook computers, the EMS firms worked to diversify their customer and product portfolios to include a wide range of non-computer goods, such as medical electronics for both hospital (GE Medical) and home (Johnson and Johnson and Lifescan) settings, consumer electronics, and especially data communications equipment. The rise of the Internet meant a huge boom in the market for data communications gear, including routers, bridges, switches, and hubs. Cisco Systems, based in Silicon Valley, CA, was an early mover in this market, and captured nearly 80% of worldwide Internet protocol router sales. Cisco, as is the case with nearly all younger Silicon Valley firms, was and is organized according to a modular value chain model, and so had little in-house manufacturing capacity. Because of frequent engineering changes and high price points Cisco has chosen to work closely with local contract manufacturers, mostly the large EMS firms based in Silicon Valley such as Solectron, Sanmina-SCI, and Flextronics, on new product introduction. This model was highly influential with managers at traditional


telecommunications firms, such as Nortel, Lucent, Ericsson, and Alcatel, who’s intense desire to transform their organizations to participate in the Internet “revolution,” caused them to emulate Cisco’s strategies. These managers were also rewarded when stock prices increased with every announcement to shed fixed assets by selling plants to contract manufacturers. The product mix of the largest five firms in 2001 is shown in Figure 4. While product mix data is not yet available for 2002 at the time of this writing, it is very likely that the share of communications equipment in the product mix of these firms has decreased significantly because orders from these firms nearly ceased in 2001 when the Internet bubble burst. Still, the product diversity of the EMS contract manufacturers relative to the ODM contract manufacturers remains quite high, especially when we consider that the products in the category of computers and peripherals includes not only personal computers, but a very large measure of workstations and enterprise computing equipment as well.

Figure 4. Product Mix for the Largest Five EMS Contract Manufacturers, 2001

Computers and peripherals

28%

Communications40%

Industrial5%

Medical6%

Military4%

Consumer8%

Other9%

Source: Electronic Business Top 100 Contract Manufacturers, 2003.

The strategies pursued by the EMS contract manufacturers during the 1990s came with a set of significant risks. Most severe, perhaps, was the risk of overexpansion.


Capital outlays for acquisitions meant the accumulation of huge fixed costs, and in some cases debt, as well as a need to integrate dozens of new facilities, a task that continues to absorb large amounts of management attention at EMS contract manufacturers. The cost of acquiring customer facilities was ameliorated to some degree by deals that guaranteed the contract manufacturer a specified level of business by the lead firms selling the plants. But most such deals have now expired, and the long IT recession has meant that the largest EMS contract manufacturers have found themselves with huge amounts of underutilized capacity. Acquisitions of competitors and the construction of new facilities, of course, came with no such buffer. Because acquired plants had been initially established and operated by a variety of firms, and many for purposes other than contract manufacturing, the cost and difficulty of integrating these plants has been very high. For a significant number of acquired plants in high cost locations such as the United States and West Europe, acquisition was followed almost immediately by closure. With the drop off in IT spending, EMS contract manufacturers began incurring losses. In the cases of Solectron and Sanmina/SCI, these losses were very large indeed, amounting to about $10B between 2001 and 2003. In a less severe downturn, the multiple customer/multiple product stance of the EMS contract manufacturers might have allowed them to shift away from slower growing or declining products and customers toward faster growing market segments, but the IT recession of 2001-2003 was both broad and deep, and capacity utilization dropped precipitously as a result. Given the recent acquisition binge the EMS firms had been on, it is an understatement to say that this capacity glut was poorly timed.

Solectron, for example, had nearly 5,000 workers in France by late 2000. About half of these jobs were permanent and half were temporary. The lion’s share of employment was at the Bordeaux plant, which Solectron had acquired from IBM in 1992. By July 2003, Solectron’s employment in France had shrunk to less than 1,500, and the Bordeaux plant was slated for closure. Plants acquired in Brittany and Nord Pas de Calais were closed almost immediately. The remaining work from the Bordeaux plant was transferred to Solectron plants in Romania and China (Le Monde, July 12, 2003, p. 15). Of these sudden plants closures, coming as they did on the heels of an equally dramatic acquisition spree, Solectron’s President could only say, "Our customers, who are the ones to decide, want our production in low-wage countries...it's a question of survival." Solectron has lost more than $6B since 2001 and has had its entire management team replaced.


Differences and Similarities Between ODM and EMS Contract Manufacturers

Because of their different histories and institutional contexts, ODM and EMS contract manufacturers show significant differences in all three dimensions of business scope that we focus on in the paper, customer and product scope, value chain scope, and geographic scope. ODM firms provide a wider range of value-chain activities than EMS firms do, especially in the area of post-architectural product design and development. As already discussed, lead firms in the personal computer industry such as HP-Compaq and Dell have retained marketing, product strategy, and much of conceptual design; and ongoing innovation by component producers Intel and Microsoft dictate most of the standard machine architectures that post-architectural design must adhere to; but the ODM firms have become world leaders in bringing personal computer products to market quickly by joining product concepts developed by personal computer companies with the changing product architectures dictated by the component makers. EMS firms have so far concentrated mainly on the base manufacturing processes that are common to most electronics products and subsystems, particularly circuit board- and product-level assembly, and despite recent efforts, have been much slower to develop design businesses.

This difference in value chain scope is related to differences in product and customer scope. EMS firms are able to leverage the same production processes across a wide range of product categories, including computers, communications equipment, consumer electronics, electronic instruments, industrial electronics, medical electronics, and electronics for military and aerospace applications; while ODM firms have largely concentrated their portfolio in personal computer-related product categories, an area where they have built up design capabilities. ODM and EMS firms are also different in regard to the geographic scale and scope of their production facilities. Because of their aggressive acquisition of lead firms' manufacturing facilities and their response to customer requests to provide one-stop global manufacturing services, the largest EMS firms have many more manufacturing locations than do ODM firms. As EMS firms have set up global operations, logistics and supply-chain management have become important competencies. ODM firms, on the other hand, have their manufacturing sites geographically concentrated in Taiwan and Mainland China. The differences between ODM and EMS contract manufacturers are summarized in Figure 5.


Figure 5. Comparison of Typical ODM and EMS Electronics Contract Manufacturers: Value Chain Scope, Product/Customer Scope, and Geographic Scope

Value Chain Scope

Distribution/Service

Sales/Marketinｇ Product/Customer Scope

Product Strategy Key Lead Firm Functions

Product Definition etc.

Industrial Product Design Medical Datacom Manufacturing Telecom Servers Logistics Logistics

US-Based EMS

Firm

Geograpic Scope

Despite differenc

emergence of value cha

following key features:

1. Macro-environmencourage valustandardization;technological cproduct and magrowth.

2. An overlappingsystems firms, wproviders.

3. A quasi-merchcustomers and m

4. Reliance on a ge5. Reliance on mec

commodity supp

June 2004

Taiwan-

Based

ODM

Firm
PCs
East Asia • North Am • Eur • East Eur • South Am

es both ODM and EMS suppliers have contributed to the

in modularity in the electronics industry because they share the

ental factors in the electronics industry that facilitate and e chain modularity, including good product and component a highly codifiable output from the design process; rapid hange at the component, process, and final product levels; rket proliferation; and, during most time periods, rapid industry

set of customers, initially consisting of American electronics hich have ceded control over manufacturing to outside service

ant, non-captive multi-customer/multi-product stance toward arkets. neric, easily transferable, widely applicable core of fixed assets. hanisms to support a thick customer interface (relative to liers).

ITEC Research Paper 04-03 p.31

The difference in value chain scope between ODM and EMS contract manufacturers alters the source, but not the thickness of the inter-firm link. The ODM supplier’s involvement in post-architectural design requires thicker linkage mechanism relative to the commodity supplier, but the standard and modular characteristics of personal computer design architecture cause the link between ODMs and their customers to be relatively thin, relative to the captive supplier. Despite a sustained attempt to increase co-design activities, asset specificity in EMS firms is generally limited to customer-specific program and project management activities that strive to give the customer the impression that EMS production is “an extension of the customer’s organization.” But the more complex, less standardized, and less modular products that EMS firms generally produce work to thicken linkages with customers, and recent fieldwork has shown a great deal of interaction around new product introduction activities such as prototyping, board layout, and process validation. For the ODM’s, the link between conceptual design and iterative design tends to be relatively thin because of the unusual roles that Intel and Microsoft play in setting personal computer design parameters. But tendencies do not mean a lack of heterogeneity. In the course of our field research interviews, several of the largest ODM contract manufacturers stated that approximately one-third of their business was still derived from OEM contracts, one-third from pure ODM, where fully-designed products were offered to customers, and one-third from contracts where the they collaborated with their lead firm customers on product design.６

While industry-standard manufacturing capabilities (Sanchez, 1996; Baldwin and Clark, 1997, 2000) are a vital element of both the ODM and EMS firm’s competence profile, this variation shows that suppliers can have a range of relationships with different customers and with the same customers across different contracts (Herrigel and Wittke, forthcoming).７ These dynamics are illustrated by Figure 6, which highlights the variable thickness of the turn-key supplier’s customer interface in the context of co-design (thicker) and build-to-print (thinner).


Figure 6. Variation in the Thickness of the Inter-firm Link: General vs. Specific Assets in the Context of Co-design and Build-to-print

(

r

(T

Unlike to have someODM firms, hany one custothat do exist iof the captivecustomer swiagain to the f

Growing Con

During and the Nortexperienced l

June 2004

The commodity supplier
Customer
interfaces

urn-key (Tu

the commodity supplier, both ODM and EM resources dedicated to specific customers. owever, lies a core of general assets that camer and toward other existing and new c

n both the ODM and EMS firms, since they supplier and are combined with a core of tching because they ease the process of iniluid character of modular value chains.

vergence and Competition

the 1980s and most of the 1990s, contract h America co-evolved with a large meaittle direct competition with one another de

ITEC Research Paper 04-03

The captive supplie

Build-to-print

rn-key supplier, instance B)

Co-design

supplier, instance A)

General Assets potentially shared

resources)

SAtn bust argentia

masursp

Specific Assets(dedicated resources)

contract manufacturers tend the center of both EMS and e quickly shifted away from omers. The specific assets e much shallower than those eral asset, actually facilitate

l collaboration. This points

nufacturers based in Taiwan e of independence, and so ite sharing some of the same

p.33

customers, such as Dell, IBM, Hewlett Packard, and Compaq. As already discussed, ODM firms have concentrated on the design and manufacture of motherboards and fully assembled mass-market desktop personal computers and more recently the design and assembly of notebook computers, especially those models with low to medium price ranges. EMS firms have also had the production of computers and related products as their main business, but have tended to focus on higher-price-point personal computers, computer servers, workstations, and large machines for enterprise computing, as well as disk arrays, printers, and other higher-value computer peripherals.

There are several very recent trends that have both pushed ODM and EMS contract manufacturers toward direct competition and shifted the competitive advantage, at least temporarily, in the direction the ODM firms. The first is the rise in lead firm demand for design services. Lead firms have already outsourced manufacturing, and so the outsourcing of routine design activities, as well as the full design of low-cost product lines, seems to be the next logical step. The 2000-2003 market downturn caused these cost-cutting efforts to become especially urgent. Each and every large contract manufacturer has created a design services group in response to this interest, and some have successfully completed projects, albeit mainly for small, start-up firms with little if any design competence. Flextronics management has perhaps been the most aggressive in this regard, and has targeted ODM design services as a key strategic growth area (Baljko, 2002). Still, it is clear that the design competencies at EMS firms fall far short of those at the ODMs. There appears to be a fundamental difficulty that is inhibiting rapid expansion of design services at EMS firms, namely their broad product scope, which is an advantage in that allows manufacturing loads to be more easily balanced, but a drawback because it makes it difficult to develop deep design expertise in any single product area. Design, it seems, is not a generic activity.

A second factor increasing head-to-head competition and advantaging ODM contract manufacturers has been the rapid expansion of manufacturing capacity in Mainland China by both sets of firms. Customers have encouraged all of their contractors to expand in China, in part because of lower costs there relative to other East Asian centers of electronics production such as Singapore, Malaysia, and Thailand; in part because of the lure of the huge potential market in China; in part to increase scale economies; and in part to take advantage of critical mass of component manufacturers that have moved production there. But the ODM contract manufacturers have a very significant head-start in China. The migration of Taiwan’s electronics production base to Mainland China began in the mid-1990s with components and peripherals, and then spread to desktop personal computers and motherboards, with the latest stage being notebook computers. This migration has established two industry clusters for


electronics manufacturing, one in the Pearl River Delta near Hong Kong and the second in the Yangze River Delta region near Shanghai. Under the tight control of their Taiwan headquarters, the ODM’s manufacturing subsidiaries have co-located with their component suppliers to respond to the time-to-market, ramp-up, and cost reduction requirements set by lead firms. Table 4 clearly shows the increasing concentration of the ODM firm’s overseas production in Mainland China. As a result the share of the value of electronic hardware production by Taiwan firms on Mainland China increased from 35% in 1999, to 37% in 2001, to 47% in 2002. Accordingly, the share of production in Taiwan decreased from 47% in 2001 to 36% in 2002 (MIC, 2003).

Table 4. Taiwan ODM Firms Overseas and China Production Volume Ratio by Major Product Category, 2001 and 2002

2001 OverseasProduction

2001 ChinaProduction

2002 OverseasProduction

2002 ChinaProduction

Notebook PCs 12% 5% 43% 38%Desktop PCs 86% 48% 94% 52%Motherboards 55% 52% 63% 62%LCD Monitors 26% 23% 76% 68%CRT Monitors 98% 66% 98% 72%Optical Disk Drives 94% 92% 95% 92%Servers 49% 16% 54% 24%

Source: MIC, ICT Industry Yearbook, 2002 and 2003.

Some EMS firms, such as Flextronics and SCI, have had manufacturing plants in China since the early 1990s, and all of the top firms have scrambled to add capacity there since the 2001 industry downturn. Flextronics now has two large industrial parks in China. Table 5 provides an example of the recent geographic shift to China by a leading EMS firm. The share of Solectron’s geographic net sales generated in China jumped from 10% in the first three months of fiscal year 2002 (September through November 2003) to 14% in the first three months of fiscal year 2003. At the same time sales in all other locations except for Malaysia declined, especially in the United States, where the share dropped precipitously, from 33% to 28%. Still, the total capacity of EMS firms in China remains far below that of the ODMs. What had been a major advantage of the EMS firms relative to the ODMs, namely their global operational footprint (Sturgeon and Lester, 2004), became a liability as demand dropped off sharply for products produced by plants in high-cost locations, (e.g., enterprise


computers and communications infrastructure equipment) and products being produced in regional plants (such as Mexico and East Europe) for fast response were shifted to China to reduce costs. The earlier surge in investments from Taiwan in the 1990s has clearly put the ODMs at an advantage in China.

Table 5. Solectron’s Geographic Net Sales, Three Months Ended November 30 2002 and 2003

2002 Sales 2003 Sales Share 2002 Share 2003United States $869.80 $763.40 33% 28%Other North and Latin America $388.80 $382.50 15% 14%Europe $456.30 $435.30 17% 16%Malaysia $321.60 $395.10 12% 15%China $273.60 $368.90 10% 14%Other Asia Pacific $358.00 $351.60 13% 13%Total $2,668.10 $2,696.80 100% 100%

Source: Solectron Quarterly Report.

Another trend that is providing ODM contract manufacturers with an advantage is the recent surge in notebook computer demand for home use, driven by lower prices, increased functionality, the advent of low cost wireless routers for the home, and display quality that has begun to rival desktop machines. The big change came in the 1999-2000 time period, when the share of notebook computer sales, which had increased incrementally from about 10% of total personal computer sales in the late 1980s to about 16% in 1998, began to increase more rapidly from 17% in 1999 to 27% in 2003 (IDC, 2003). The historical strength of ODM firms in notebook computers means that they are being favored as the market shifts in their direction. Finally, to the extent that Japanese firms have embraced strategic outsourcing along the lines of value chain modularity—and most have been very reluctant and slow to do so—they have chosen Taiwan contract manufacturers over their North American rivals. As Table 2 shows, Sony did not begin sourcing from Taiwan ODMs until 2002, but MIC estimates that Sony purchases amounted to $4B in 2003, a figure which places the firm as the third largest purchaser of IT hardware from Taiwan firms. Toshiba is also a major customer, though not in the top five. While the European lead firms seem to have largely embraced American-style contract manufacturing, Japanese lead firms, after a


period of questioning in the late 1990s, appear to have retrenched in their traditional strategies for vertical integration. While this variation is striking, and may have to do with the product categories in which Japanese firms are concentrated, namely small-form factor consumer electronics where miniaturization necessitates product architectures that are highly integrated and product-specific, vertical integration appears to be a very difficult strategy to maintain in the face of heavy pressure to reduce costs.

The pattern of revenue and net income growth at the largest five EMS and ODM firms tells a dramatic story of heady growth, crisis, increased competition, and reversal of fortune. In the period 1993 to 2000, EMS revenues outpaced revenue growth at the largest five ODM firms by a significant margin. Net income was relatively low during this period—below five percent—but firms often choose to exchange short-term profitability for increases in market share. Notwithstanding the decline in EMS revenues in the 2001-2002 period due to a sudden drop in demand for communications infrastructure equipment, as shown in Figure 7, between 1993 and 2002 75% of the total revenues generated by the top five ODM and EMS firms combined ($336.7B) were captured by the top five EMS firms. There are two reasons for this. First, EMS firms were awarded larger and greater numbers of orders from lead firms. The larger geographic scope allowed EMS firms to take over global manufacturing responsibility for their customers. In addition, the narrower value chain scope of the EMS firms allowed them to rapidly add customers in new product segments, such as communications infrastructure equipment, while the ODM firms grew largely by capturing a larger share of the production within the expanding personal computer industry. Second, EMS firms grew both organically and through acquisition (of both competitors' and customers' facilities), while ODM firms grew only organically and scaled up largely on Mainland China.


Figure 7. Revenue Comparison of Largest Five ODM and EMS Electronics Contract Manufacturers: 1993-2003, Millions of Current US Dollars

$0

$10,000

$20,000

$30,000

$40,000

$50,000

$60,000

1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003

Top 5 ODM firms Top 5 EMS firms

Source: Company annual reports.

Note: The largest five EMS firms are Flextronics, Solectron, Sanmina-SCI, Celestica, and Jabil.

The largest five ODM firms are Hon Hai, Quanta, Acer, Compal, and Asustek. In other words, the ODM firms’ focus on a single electronics sub-sector, personal

computers, may have limited their growth relative to the EMS firms, which have been able to expand into a wide spectrum of electronics sub-sectors, especially the burgeoning data communication and medical electronics fields. Another possible scenario has to do with value chain scope. Because ODM firms have expertise in product design, lead firms may have been wary about granting them too much business as they increased their strategic outsourcing, instead choosing EMS firms which, because they do far less design work, were perceived as less of a potential threat. Similarly, ODM firms may have fallen into a “competency trap” of sorts (Levitt and March, 1988). Because of their higher degree of specific assets related to design, ODM firms may have necessarily had to move more slowly into non-computer sectors because of the time required to build up the requisite design expertise.

There are of course, possible causes of the growth differential between ODM and EMS firms that do not have to do with business scope. Macro-level environmental factors, such as the share price increases in the United States stock markets in the 1990s,


which provided capital for the largest EMS firms to acquire dozens of customer facilities and smaller competitors. ODM firms are mostly listed on the Taiwan stock market, and this makes growth through acquisition extremely difficult. Additional factors, such as the long experience by American managers with international operations and a desire on the part of American lead firms for spatial and social propinquity with their increasingly important outsourcing partners—at least on the headquarters level—may have also played a part.

Ironically, perhaps, the differences that led EMS firms to outgrow ODM firms in the 1990s have worked against them since the downturn. Firms may temporarily substitute market share for profitability, but this can turn out to be a very dangerous strategy in volatile industries such as electronics. EMS firms’ diversification into non-computer product categories, an effort to reduce demand cyclicality through diversification, placed them at the center of largest boom and bust cycle the electronics industry has ever witnessed. Moreover, their strategy of growth through acquisition left them with dozens of underutilized plants in high cost locations, and as the downturn dragged on, the situation was exacerbated by customer demands to shift production to low-cost locations. As already discussed, EMS firms have been scrambling to close plants in high-cost locations, a very expensive proposition, while at the same time investing heavily in new production capacity in China. Finally, the narrow value chain scope that had enabled rapid expansion into new product areas began to work against the EMS firms as customers began to look for outsourcing partners that could quickly take over the design for low-cost segments of their product lines. Much of this business has gone to ODM firms.

As a result of all of these trends, revenue growth at ODM firms far outpaced the EMS firms in the period 1993 to 2000. The top five EMS firms’ revenue grew at 50% per year during this period, while the top five ODM firms’ revenue grew at 36% per year. In the period 2000-2003, however, ODM firms revenue grew 34% per year while the top five EMS firms’ revenue, as a whole, declined by 5 percent. Much of these declines can be attributed to Solectron and Sanmina-SCI, which combined lost 90% of the more than 10 billion dollars lost by the top five EMS firms during this period. But even without this precipitous, and perhaps temporary drop in profitability, ODM firms consistently performed better than EMS firms in the 1990s and early 2000s. Figure 8 compares return on sales for the largest five ODM and EMS contract manufacturers from 1993 to 2003. During the latter half of the 1990s, it was rare for a major EMS contract manufacturer to exceed five percent return on sales, while it was equally rare for a major ODM firm to drop below five percent. The systematic difference in profitability might be explained by the design services provided by ODMs, which


generally have a higher rate of profit than manufacturing, and the fact that at least half of EMS revenues are derived from purchasing components from contracts negotiated by their customers, a service with very little, if any profit since customers are well aware of the prices that EMS firms are charged by component manufacturers and distributors.

Figure 8. Return on Sales Comparison of Largest Five ODM and EMS Electronics Contract Manufacturers: 1993-2003*

-15.0%

-10.0%

-5.0%

0.0%

5.0%

10.0%

15.0%

1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003

Top 5 ODM firmsTop 5 EMS firms

` * For ODMs data was collected for 1993-2003, for EMS firms data was collected for 1995-2003.

Source: Company annual reports.

Note: The largest five EMS firms are Flextronics, Solectron, Sanmina-SCI, Celestica, and Jabil.

The largest five ODM firms are Hon Hai, Quanta, Acer, Compal, and Asustek.

This reversal of fortunes may turn out to be permanent, and signal the beginning of

a long decline of the North American EMS firms relative to the Taiwan ODMs as outsourcing of design increases and production continues to concentrate in China, or it may not. There are two scenarios that could bolster the competitive position of the EMS contract manufacturers. The first is that the EMS firms develop competitive design competencies and increase their footprint in China to match that of the ODM firms. To us, this seems unlikely given the large head-start that the ODM firms have in these two areas. The second, and perhaps more plausible scenario is that a full-scale revival of worldwide IT spending, perhaps combined with a new “killer” application or major technological shift in production technology, will once again constrain capacity. This would work to alleviate extreme downward price pressure, and put a premium on


co-location in advanced economies to support co-design and collaboration in new product introduction. It may also revive interest in regional production (e.g., Mexico and East Europe) to support customization and rapid time-to-market. If such shifts do occur, and if they have as their innovative epicenter the places that have traditionally spawned them in the past (e.g., Silicon Valley and Boston), the North American EMS firms will likely still be in an excellent position to win this new business. A third scenario, and one that is perhaps the most likely, is for the continued co-existence and expansion of both sets of contract manufacturers as the demand for manufacturing services continues to expand, and a blurring of the distinctions between the two groups of firms. The firm that is now the largest Taiwan contract manufacturer, Hon Hai, provides few design services and thus might better be labeled an EMS contract manufacturer than an ODM. If the largest EMS contract manufacturer, Flextronics, is successful in developing its design services business, it may make more sense to refer to it as an ODM contract manufacturer. Only time will tell us which scenario is correct—-if indeed any of them are—but one can be sure that rapid innovation and volatility will continue to characterize the electronics industry for some time to come.

Conclusion: implications for strategy and further research

By exploring the emergence of the shared, modular supply-base from a co-evolutionary perspective we have tried to provide a framework for understanding the modern industrial landscape, especially as it comes to be characterized by strategic outsourcing, vertical specialization, and horizontal industry structure. Questions about the dynamics and evolutionary aspects of inter-firm vertical relationships across different value chains have implications for both the formulation of competitive strategy and for future research.

First, our argument suggests that the key organizational transformation underway should be thought of not as de-verticalization but as de-linking production-related activities from innovation-related activities (Sturgeon, 2000). In Schumpetarian competition, innovative firms that gain market share build up large productive assets specific to the product and create barriers to new entry. Building on Schumpeter’s conception of innovation, Nelson and Winter (1982) show that the dynamic process of industrial evolution tends to create larger firms and more concentrated market structures over time because successful innovations lead to higher profits and greater investments in productive capacity that put innovative firms ahead of their competitors and then protect them from new market entry. However, the emergence of modular value chains alters the relationships between innovative capacity, firm size and scope, and


entry conditions. Firms that outsource a large share of their manufacturing no longer have to carry the financial, administrative, and technical burdens of fixed capital related to production, allowing them to focus on innovation and become more organizationally and geographically flexible (Sturgeon, 2000: 16). On the other hand, the market position of dominant firms is no longer protected by large-scale firm-specific investments in plant and equipment, making market entry more feasible because existing firms are no longer buffered from competitive pressure by large, in-house fixed capital. Barriers to entry are lower because competitors can tap the same modular supply-base and therefore gain access to leading-edge production capacity. Thus, competition between lead firms becomes more tightly tied to product-level innovation, increasing the pressure on firms to execute in the areas of product development, brand development, and marketing. The rising costs associated with an intensified focus on innovation and marketing in turn makes it less likely that lead firms will invest in in-house productive capacity in the future, speeding the shift toward modular industry organization, a shift that augers well for the continued growth of contract manufacturers of all types.

Second, the emergence of a shared supply-base creates a fruitful arena to explore the relationship between organizational capabilities and production routines. To be capable of providing turn-key production services at competitive prices, a contract manufacturer has to specialize in a set of generic production routines that can be efficiently applied across a wide range of customers. By doing so, a supplier can effectively elevate the level of capacity utilization and hence reduce overall costs. On the other hand, a contract manufacturer may have to accommodate various degrees of customer-specific needs for differentiation purposes. Furthermore, as customers demand services at sites around the globe suppliers have to replicate and/or integrate their routines across national borders while maintaining service integrity. The sudden increase in demand for production services in China has been to the advantage of ODMs, but as we have already suggested, an upturn could easily turn the tables back in the EMS firms’ favor by increasing the value of regional production systems and their ability to tie these regional systems together in an integrated global operational footprint.

Third, research on the changing patterns of industrial outsourcing offers an excellent opportunity to examine the determinants of a firm's boundary decisions. The traditional analytical tool for explaining boundary decisions lies in the transaction cost theory (Williamson, 1975, 1985). This theory requires decision makers to consider the costs that may be incurred due to transaction-specific investments made by either party in order to decide whether to outsource or produce in-house. However, as even


complex transactions become easier to codify, and the buyer-supplier relationship within the context of the modular value chain becomes quasi-merchant in nature (Sturgeon, 2000), the outsourcing firm’s concern over the costs of governing external transactions with suppliers matters less than its desire to benefit from access to capabilities that are costly to develop internally and impossible to acquire through arms-length market transactions (Barney, 1999). In other words, the potential level of value that can be added by partners and potential partners becomes a critical determinant of decisions about boundaries between transacting firms and the extent of outsourcing activities (Zajac and Olsen, 1993).

Fourth, one of the central questions for policy-makers and entrepreneurs in Taiwan is how to develop lead firms with strong global brands and a firm grasp on the innovative trajectory of the products they make and the markets they serve. Taking the above discussion as a window into recent trends in global economic organization, it is easier to explain why the electronics industry in Taiwan has been extremely successful in some areas, namely large scale manufacturing and post-architectural design, while being so unsuccessful in other areas, namely product definition, brand development, and new market creation. This is because new forms of industrial organization, such as value chain modularity, allow these sets of functions to be efficiently coordinated among separate firms and separate locations. The policy questions that arise are profound. Should Taiwan be satisfied to be a vibrant hub for component and contract manufacturing, or should it strive to develop capabilities in the realms of branding and product strategy? In field interviews, respondents from EMS and ODM contractors, as well as at their customers, repeatedly stressed the pitfalls associated with contract manufacturers moving into branded products. Competing with customers, even in small ways, risks “killing the golden goose.”

If moving onto branded products is not to be a sequential step from contract manufacturing, then are there other ways that Taiwan, as a whole, can move in the direction of branding and other high value activities? One way could be for Taiwan hardware start-ups to rely on the contract manufacturing capacity in their backyard and across the globe instead of seeking to build-up in-house capacity, just as start-ups in Silicon Valley and as the many fabless semiconductor design houses in Taiwan do. But to do this, Taiwan entrepreneurs would have to think differently than they have in the past, moving from a “manufacturing-first” to an “idea-first” approach to new firm formation. These questions are especially pertinent as the bulk of Taiwan’s electronics manufacturing base shifts to Mainland China. And for Taiwan and North America alike, the critical issue may be, not the relative competitiveness of its firms, but the rapid shift of production—and perhaps even design—employment to China.


Notes:

１ The market research firm Technology Forecasters estimates that the revenue of contract manufacturing will share around 15% of the total cost of electronics goods sold in 2001. Moreover, the growth rate of contract manufacturing in the coming five years will be more than 20% a year, which is more than twice as quickly as the electronics industry as a whole (Economist, 02/12/2000). ２ While in Taiwan’s electronics industry the term “OEM,” is widely used to indicate manufacturing-only contracting, in other countries and industries the same term is used to refer to “lead” or “customer” firms. In the motor vehicle industry for example, OEM, or simply “OE,” is used to refer to automakers such as Ford, Volkswagen, and Toyota. In the American electronics industry, we often find the very confusing situation where firms that have outsourced most of their manufacturing, such as Hewlett Packard, or firms that have never had large-scale in-house manufacturing, such as Cisco, are nevertheless referred to as “OEMs.” See Sturgeon (2000) for an extended discussion. ３ Worldwide personal computer unit sales and value of sales are from International Data Corporation, a private market research firms based in Framingham, MA , as cited in a press release dated December 11, 2003. Taiwan ODM personal computer unit production and value of production figures are from Market Intelligence Center as cited in the PowerPoint presentation "The Taiwan ICT Industry Outlook", by Paulo S. Su ([email protected]), Manager, Market Intelligence Center, Institute for Information Industry (http://mic.iii.org.tw/english/). ４ The International Data Corporation’s definition of information technology includes hardware (servers, personal computers, networking equipment, printers, etc.), software (system infrastructure software, applications, and application development and deployment), and IT services (training and education, planning, implementation, maintenance and support, and operations). There are certain categories that may not be included due to lack of coverage, or that are not in International Data Corporation’s research scope, for example, telecommunications spending, and some hardware categories, such as scanners and digital cameras, among others. ５ While Flextronics International is incorporated in Singapore and performs some important global functions there, such as purchasing, the top management and most key functions, including strategic planning and IT development, are centered in San Jose, California. ６ This terminology has recently led to the emergence of a third term, “original brand manufacturer,” or “OBM,” which denotes a supplier that has begun to develop, manufacture, and sell its own brand of products. Because an increasing number of competencies are required to move from manufacturing, to design, to sales and marketing, there has been a widespread assumption that firms and nations that begin as suppliers to lead firms in other places might upgrade their position in the global economy by moving sequentially through OEM, ODM, and OBM stages of development. While movement along this trajectory certainly offers firms from developing countries an attractive learning and upgrading path, it also can set them on a collision course with their customers, which is why the contract manufacturers that have grown the fastest have been those which have adopted a “pure play” contract manufacturing strategy and have publicly and repeatedly renounced the aim of developing and marketing their own brand of products (Sturgeon and Lester, 2004). ７ Herrigel and Wittke (forthcoming) argue that such variations are increasing as buyers and suppliers search for advantage in the context of increased outsourcing and building supplier competence. Intense collaboration, in the context of co-design for example, builds interdependence and moves the relationship in the direction of captivity. Commodity relationships tend to flow from customer purchases of off-the-shelf designs from ODMs, which makes the relationship more market-like and tends to increase supplier power, and from build-to-print arrangements, which tend to increase buyer power. According to Herrigel and Wittke (p. 15), both parties tend to seek to mix these linkage


forms and change them over time to keep independencies and power asymmetries at reasonable levels, a process they call “sustained contingent collaboration.”

References:

Arthur, W. B. 1989. Competing technology, increasing returns, and lock-in by historical events,

Economic Journal, 97: 116-31.

Baldwin, C. Y., & K. B. Clark. 1997. Managing in an age of modularity. Harvard Business

Review, 75(5): 84-93.

Baljko, J. 2002. ODMs set to lead recovery in contract sector. Electronic Business News, December 19.

Barney, J. B. 1999. How a firm’s capabilities affect boundary decisions. Sloan Management Review. Spring, 137-145.

Bensaou, M. and E Anderson. 1999. Buyer-supplier relations in industrial markets: When do

buyers risk making idiosyncratic investments? Organization Science, 10 (4), p. 460-81.

Bettis, R., S. Bradley, and G. Hamel. 1992. Outsourcing and industrial decline. Academy of Management Executive, 6(1): 7-22.

Bradach, J. and Eccles, R. 1989. Price, authority, and trust: From ideal types to plural forms.

Annual Review of Sociology. 15: 97-118.

Celly, K. S., R. E. Spekman, and J. W. Kamauff. 1999. Technological uncertainty, buyer preferences and supplier assurances: An examination of Pacific Rim purchasing

arrangements, Journal of International Business Studies, 30 (2), p. 297-316.

Cohen, W. and D. Levinthal. 1990. Absorptive capacity: A new perspective on learning and

innovation. Administrative Science Quarterly, 35: 128-152.

Curry, J. and M. Kenney. 1999. Beating the clock: Corporation responses to rapid change in the

personal computer industry. California Management Review, Fall, 42(1): 8-36.

Dedrick, J. and K. Kraemer. 1998. Asia’s Computer Challenge: Threat or Opportunity for the United States and the World. Oxford, UK: Oxford University Press.

Dyer, J. H. & H. Singh. 1998. The relational view: Cooperative strategy and sources of

interorganizational competitive advantage. Academy of Management Review, 23/4: 660-679.


Dyer, J. H. 1997. Effective inter-firm collaboration: how firms minimize transaction costs and

maximize transaction value. Strategic Management Journal, 18(7): 535-556.

Economist. Successful manufacturers these days need not go anywhere near a factory. Feb. 12, 2000.

Farrell J. and Saloner, G. 1985 Standardization, compatibility, and innovation Rand Journal 16: 70-83.

Fine, C. 1998. Clock speed: Winning industry control in the age of temporary advantage. Reading, MA: Perseus Book.

Gereffi, Gary, John Humphrey, and Timothy Sturgeon. Forthcoming. “The Governance of Global Value Chains.” Review of International Political Economy.

Granovetter, M. 1985. Economic action and social structure: The problem of embeddedness,

American Journal of Sociology, 91, pp. 481-510.

Hamel, G. 1991. Competition for competence and inter-partner learning within international

strategic alliances. Strategic Management Journal, 12, Summer: 83-104.

Heide, J. B. and A. S. Miner. 1992. The shadow of the future: Effects of anticipated interaction

and frequency of contact on buyer-seller cooperation. Academy of Management Journal, 35(2): 265-

Helper, S.; Macduffie, J.P.; and Sabel, C. 2000. Pragmatic collaboration: Advancing knowledge

while controlling opportunism. Industrial and Corporate Change 9(3): 443-488.

Herrigel, G.and Wittke, V. forthcoming. Varieties of Vertical Disintegration: the Global Trend Toward Heterogeneous Supply Relations and the Reproduction of Difference in US and German Manufacturing, Chapter ten in: Glenn Morgan, Eli Moen & Richard Whitley, eds,

Changing Capitalisms: Internationalisation, Institutional Change and Systems of Economic Organization. Oxford University Press, Oxford, UK..

Hitt, M. A., B. W. Keats, and S. M. DeMarie. 1999. Navigating in the new competitive landscape:

Building strategic flexibility and competitive advantages in the 21st century. Academy of Management Executive, 12(4): 22-42.

Jarillo, J. 1988. On strategic networks. Strategic Management Journal. 9: 31-41.

Johanson, J. and Matsson, L. 1987. Interorganizational relations in industrial systems: A network

approach compared with the transaction-cost approach. International Studies of Management and Organization. 27(1): 34-48.


Katz M. L., and Shapiro C. 1985. Network externalities, competition, and compatibility. American Economic Review 75: 424-440.

Katz, M. L., and Shapiro, C. 1986. Technology adoption in the presence of network externalities. Journal of Political Economy 94: 822-841.

Khanna, T., R. Gulati, and N. Nohria. 1998. The dynamics of learning alliances: Competition,

cooperation, and relative scope, Strategic Management Journal, 19(3): 193-210.

Langlois, Richard, 2003. “The Vanishing Hand: The Changing Dynamics of Industrial Capitalism,” Industrial and Corporate Change 12(2): 351-385 (April).

Lee, J. R. and J. S. Chen. 2000. Dynamic synergy creation with multiple business activities: Toward a competence-based business model for contract manufacturers. In Ron Sanchez

and Aime Heene (eds), Advances in Applied Business Strategy, Volume 6A: Theory Development for Competence-based Management, p. 209-228, Stamford, CT: JAI Press Inc.

Lee, J. R., J. S. Chen, and M. Tang. forthcoming. Exploring the growth strategy of contract electronics manufacturers in Taiwan: A competence-based perspective. In Ron Sanchez and

Aime Heene (eds), Advances in Applied Business Strategy, Volume 7A: Competence Perspective on Managing Internal Processes, Stamford, CT: JAI Press Inc.

Lei, D. and J. Slocum, Jr. 1992. Global strategy, competence-building and strategic alliances.

California Management Review, 35(1): 81-97.

Le Monde, 2003. Solectron prepares a fifth wave of layoffs, July 12, p. 15. Paris, France.

Levinthal, D. A. and J. March. 1993. The myopia of learning. Strategic Management Journal, 14(special issue): 95-112.

Levinthal, D. A. and J. Myatt. 1994. Co-evolution of capabilities and industry: The evolution of

mutual fund processing. Strategic Management Journal, 15: 45-62.

Lewin, A. Y. and H. W. Volberda. 1999. Prolegomena on co-evolution: A framework for research

on strategy and new organizational forms. Organization Science, 10(5): 519-34.

Lewin, A. Y., C. P. Long, and T. N. Carroll. 1999. The co-evolution of new organizational forms.

Organization Science. 10(5): 535-50.

Lorenz, E. 1988. Neither friends nor strangers: Informal networks of subcontracting in French

industry. In Diego Gambetta (ed). Trust: Making and Breaking Cooperative Relations. Oxford, New York: Basil Blackwell.


Lorenz, E. 1992. The search for flexibility: Subcontracting networks in British and French

engineering. In Storper, M. and Scott, A. (eds.), Pathways to Industrialization and Regional Development. London and New York: Routledge, p. 122-132.

Madhok, A. 2000. Inter-firm collaboration: Contractual and capability-based perspectives, in N.

Foss and V. Mahnke (eds.) Competence, Governance, and Entrepreneurship, pp. 276-303, New York, NY: Oxford University Press, Inc.

Madhok, A. and S. B. Tallman. 1998. Resource, transactions and rents: Managing value through

inter-firm collaborative relationships. Organization Science, 9: 326-39.

Makides, C. and N. Berg. 1988. Manufacturing offshore is bad business. Harvard Business Review, 66(4): 113-20.

McMillan, J. 1990. Managing suppliers: Incentive systems in Japanese and United States industry,

California Management Review, Summer, 38-55.

McKelvey, B. 1997. Quasi-natural organization science. Organization Science, 8: 352-80.

Nelson, R. R. and S. G. Winter. 1982. An evolutionary theory of economic change. Cambridge, MA: Belknap.

Nishiguchi, T. 2001. Co-evolution of interorganizational relations. In I. Nonaka and T.

Nishiguchi (eds.) Knowledge emergence: Social, technical, and evolutionary dimensions of knowledge creation, pp. 197-222. New York, NY: Oxford University Press.

Porter, M. E. 1980. Competitive strategy: Techniques for analyzing industries and competitors, New York, NY: Free Press.

Porter, M. E. 1998. Clusters and the new economics of competition, Harvard Business Review, November-December, pp. 77-90.

Powell, W. 1987. Hybrid organizational arrangements: New form or transitional development?

California Management Review, Fall: 67-87.

Powell, W. 1990. Neither market nor hierarchy: Network forms of organization. Research in Organizational Behavior, 12: 295-336.

Quinn, J. B. and F. G. Hilmer. 1994. Strategic outsourcing, Sloan Management Review, 35(summer): 43-55.

Richardson, G. 1972. The organization of industry. The Economic Journal. 84: 883-96.

Richardson, J. 1993. Parallel sourcing and supplier performance in the Japanese automobile

industry, Strategic Management Journal, 14: 339-350.


Sanchez, R. 1996. Strategic product creation: managing new interactions of technology, markets,

and organizations, European Management Journal. 14(2): 121-138.

Schilling, M.A. and Steensma, K., 2001. ‘‘The Use of Modular Organizational Forms: An Industry Level Analysis,’’ Academy of Management Journal 44(6):1149–1168.

Sturgeon, T. J. 1997. Turnkey production networks: A new American model of industrial

organization? Working paper 92A, Berkeley Roundtable on the International Economy, University of California at Berkeley.

Sturgeon, T. 1999. Turn-key Production Networks: Industrial Organization, Economic Development, and the Globalization of the Electronics Manufacturing Supply-base. Ph.D. Dissertation, Department of Geography, University of California at Berkeley.

Sturgeon, T. 2000. Turn-key production networks: the organizational de-linking of production

from innovation, in U. Jürgens (Ed.), New Product Development and Production Networks, Berlin: Springer-Verlag.

Sturgeon, T. 2002. Modular production networks: A new American model of industrial

organization. Industrial and Corporate Change. 11(3): 451-496.

Sturgeon, T. and Lester, R. 2004. The new global supply-base: New challenges for local suppliers

in East Asia. In Shahid Yusuf, Anjum Altaf and Kaoru Nabeshima (eds.), Global Production Networking and Technological Change in East Asia, Oxford, UK: Oxford University Press.

Teece, D. J., Pisano, G. and Shuen, A. 1997. Dynamic capabilities and strategic management,

Strategic Management Journal, 18 (7): 509-533.

Thorelli, H. 1986. Networks: Between Markets and Hierarchies. Strategic Management Journal. 7: 37-51.

Tzeng,, David and Chang,Sarah 2003. Quanta to become Acer's biggest notebook supplier. DigiTimes.com. Apri 4.

Venkatesan, R. 1992. Strategic outsourcing: To make nor not to make, Harvard Business Review, Nov.-Dec., 1-11.

Williamson, O. E. 1975. Markets and hierarchies. New York, NY: Free Press.

Williamson, O. E. 1985. The economic institutions of capitalism. New York, NY: The Free Press.

Yoffie, D. 1997. Competing in the age of digital convergence. Boston, MA: Harvard Business School Press.


Zajac, E. J. and C. P. Olsen. 1993. From transaction cost to transactional value analysis:

implications for the study of inter-organizational strategies. Journal of Management Studies. 30(1): 131-145.


industry co-evolution: a comparison of taiwan and north america's

Documents