Advances in Applied Sociology
2013. Vol.3, No.1, 69-78
Published Online March 2013 in SciRes (http://www.scirp.org/journal/aasoci) http://dx.doi.org/10.4236/aasoci.2013.31009
The ODM Model and Co-Evolution in the Global Notebook PC
Industry: Evidence from Taiwan
Daniel You-Ren Yan g 1, Yun-Chung Chen2
1Department of Sociology, Tunghai Universi t y, Taichung, Taiwan
2Department of Sociology, Hong Kong Baptist University, Hong Kong, China
Received December 24th, 2012; revised January 25th, 2013; accepted February 9th, 2013
This paper aims to explore the evolutionary dynamics of the ODM model in the global notebook PC in-
dustry. We found that the “modular design” process of notebook products requires plenty of technical in-
teraction and tacit knowledge exchanges among branding companies, key component suppliers and
ODMs. Taiwanese ODMs serve as important sources of information and knowledge for specifications
formulation through their system integration and technological development abilities. With increasing
complexity and shorter design life-spans, the mutual dependency among the lead firms and ODMs in-
creases, leading to the regional agglomeration in Taiwan and China.
Keywords: Subcontract Manufacturing; Modularity; Innovation; Global Production Networks; Value
Chain Governance; Latecomer’s Upgrading
Introduction: The Absentness of ODMs?
ODM (original design manufacturer) is a popular model of
production in East Asian’s industrialization, especially for
Taiwanese enterprises. The model was considered as a main
strategy for Taiwanese companies to upgrade from the OEM
(own equipment manufacturer) that the firm produces the com-
plete, finished product (both consumer and industrial product)
to the specification and design provided by the foreign custom-
ers (such as buyers or TNCs) in a subcontract arrangement
(Amsden & Chu, 2003; Cy nn, 2002; Hobday, 1995, 2003). How-
ever, the ODM model seemed to be ignored in the discussions
of global value chains as well as strategic outsourcing (Berg-
gren & Bengtsson, 2004; Dankbaar, 2007; Steinle & Schiele,
2008). For example, in the study of iPod and notebook PC
value chain (Dedrick, Kenneth, & Linden, 2010), there was for
almost no discussion of the Taiwanese ODMs. The roles of
these ODMs were almost absent in the global PC industry that
seemed to be overly dominated by the lead firms, such as HP
and Wintel, as described by the article: “… PC makers carry
out systems integration at a functional level, but most of the im-
portant system-level decisions have already been made by Mi-
crosoft and Intel… Many microchip vendors pursue a similar
strategy to Intel’s, offering complete reference designs, includ-
ing recommended system layout and software, which can be im-
plemented rapidly by customers with limited internal exper-
tise…” (Dedrick, Kenneth, & Linden, 2010). Is the story really
true?
Under ODM, the latecomer carries out some or all of the
product design and process task needed to produce a good ac-
cording to a general design layout from the foreign buyers or
TNCs (Hobday, 1995, 2002). Nevertheless, we observed that
the ODMs had advanced their technological competence and
applied not only to incremental designs, but also the leadership
product innovation after late 90’s. This article aims to investi-
gate the notebook PCs’ product innovation of Taiwanese ICT
(information-communication technology) ODM manufacturers
as an illustration of the special upgrading pattern in East Asia.
Two research questions were raised: Is the notebook PC Indus-
try dominated by the production networks in which brand-name
companies practice worldwide sourcing? Is the lead firm from
advanced economies determined the modular system architec-
ture of the notebook PC products that facilitate the borderless
global sourcing?
The modularization thesis argues that the competitiveness of
the ODMs in the PC industry is resulted from the modulariza-
tion of PC’s system architecture. The thesis gives prominence
to the roles played by the lead firms and the platform leaders
(such as Intel and Microsoft) in setting up such architecture
(Borrus, 2000; Borrus & Zysman, 1997; Sturgeon & Lee, 2005;
Kawakami, 2011; Sturgeon & Kawakami, 2011). For instance,
emphasizing the dominant role of the Wintelism and the lead
firms (HP, Dell, IBM, etc.), Sturgeon and Lee (2005) illustrated
how the intense competition and short product life cycles in the
PC market motivate lead firms to spread risk and lower cost by
outsourcing production to the ODMs (Sturgeon & Lee, 2005).
Kawakami’s decent study acknowledges the Taiwanese ODMs
had exploited those inter-firm relationships in enhancing their
capabilities, she still highlights the role of the platform leader—
WINTEL, simplified the story with the large substitutability of
these ODMs in the modular system architecture (Kawakami,
2011). However, in other ICT products like iPhone, the ODMs’
collaboration with the brand companies is even more durable
than those platform providers like INTEL.
The PC industry is a paradigmatic example of technological
modularization, with modularity in product design increasing
the ease of use and maintenance for customers and affecting the
modularization of production1. Furthermore, it is also the case
utilized by the “strong modularity school” most often to support
1For example, module components (CPU, HDD, CD-ROM, Modem, LCD
panel, memory module, etc.) make up 60% of Quanta’s procurement cost
(Yung, Le e, & Lai, 2009).

D. Y.-R. YANG, Y.-C. CHEN
the “natural” convergence of industries toward a modular con-
figuration (Campagnolo & Camuffo, 2010). Taking notebook
PCs as an example, however, we find that Taiwanese ODMs’
share of the global market volume has increased continuously
from 40% in 1998 to 85% in 2005 (Dedrick & Kraemer, 2008)
during the process of further modular design, which is a pattern
that departs from the assumed high substitutability through in-
creasing modularity2. Moreover, these ODM activities are also
geographically concentrating in the industrial clusters in Tai-
wan and then China, instead of spreading evenly in the global
scale (Yang & Coe, 2009; Yang & Hsia, 2007).
We acknowledge that the open system architecture and tech-
nological modularization of the notebook PC industry did lay a
foundation of learning for Taiwanese OEM manufacturers in
the early stage. While what more significant for us is the trans-
formation of contract manufacturing model after late 90’s. In
the ODM model, we observed that the lead firms (including the
brand name companies and the key component suppliers) no
longer direct the industrial specification in the product design
of the notebook PC. The ODMs get deeply involved in the
process of technological codification and play even active roles
in the product innovation. Taiwanese ODMs serve as important
sources of information and knowledge for specifications for-
mulation and roadmap development through their system inte-
gration and technological design capabilities. Increasingly new
product innovation (not merely process innovation) is initiated
at the ODM-based technology cluster in Northern Taiwan Re-
gion, creating a new form of technology interdependency. With
increasing complexity and shorter design life-spans, the mutual
dependence among the lead firms and ODMs does not decrease
during the process of technological modularization, on the con-
trary, mutual dependency in the production networks actually
increases.
Through investigating the specification—the most concrete
technological modularization—and its historical transformation
in the ODM model after late 90’s, this research therefore claims
the “co-evolution” of technological development and organiza-
tional governance. The GVCs’ (global value chain) theorization
should take such evolutionary characteristic of product innova-
tion into account, as well as its influence on organizational and
industrial structure. To summarize, we are opposed to the mo-
dular organization theory with the style of technological deter-
minism, and provide concrete evidence based on the notebook
PC industry.
Methodologically, this article zooms in on the organizational
behaviors of those ODMs with regard to product innovation in
the industrial as well as firm level, and uses specifications (or
simply “specs” as commonly used in the industry) as the plat-
form to discuss the negotiations between the lead firms and the
ODMs. Our argument unfolds in five further stages. First we
position our conceptual argument within the wider literatures of
technology development and value chain analysis to provide a
theoretical framework of the empirical investigation. Second,
we discuss the basic pattern of notebook PC’s product devel-
opment in the ODM model. Third, we provide firm-level evi-
dences of the intense technical interaction between branding
companies, ODMs and key component manufacturers. Forth,
we illustrated the practice of these technological-organizational
features in the daily “search routine3” (Dosi & Nelson, 1994)
of the ODMs’ technical teamwork. In the final section we offer
some concluding comments and the wider implications of our
analysis.
Modularity, Value Chains and Technology
Evolution
The establishment of industrial standards and specifications
in global production networks not only affected geographical
industrialization but also influenced interactions and systems of
governance between companies. The notion of modularization
emerges from the post-Chandlerian industrial organization de-
bate in the organization economics field since the 1990s (Chan-
dler, 1962, 1977, 1990; Lamoreaux, Raff, & Temin, 2003; Lan-
glois, 2002, 2003, 2004). The modularization thesis highlights
that the industrial standards and specifications are important
mechanisms that unify various production phases, and the pro-
cess of standardization massively decreases the information
needed for coordination between companies (Langlois, 2002,
2003, 2004). Sturgeon (2002, 2003) examined the tacit and co-
dified knowledge exchange in leading American electronics
firms and suggested that the “modular production networks” is
rising. He stated that “key node” in the global value chain relies
on internal, tacit knowledge exchanges, while other nodes were
linked with the key node through codified, standardized know-
ledge and information, which drive the production networks
toward global expansion through speed, flexibility, and cost de-
creases. The growth of “EMS (electronic manufacturing ser-
vices)” companies (such as Flextronics, Jabil Circuit, & Ce-
lestica) illustrates a new American model of contract manufac-
turing under modular production networks. Furthermore, the
increased technological modularization and spec. codification
enhanced the electronics companies with the ability of global
expansion. The EMS companies that have global operations are
in a better position to serve the lead firms by providing stan-
dardized modules, and the lead firms are freed from mutual
dependence on their suppliers and could source from multiple
suppliers globally (Sturgeon, 2002, 2003).
The modularization thesis is showing its influence, for ex-
ample, in the conceptualization of global value chains (GVCs),
with the key idea of “value chain modularity”(Gereffi, Hum-
phrey, & Sturgeon, 2005; Sturgeon, Biesebroeck, & Gereffi,
2008; Sturgeon & Kawakami, 2011). The idea implicitly as-
sumes that the “technological modularity” (such as modular
product design) will inevitably lead to “organizational modula-
rity”, in which “… suppliers and customers can be easily linked
and de-linked, resulting in a very fluid and flexible network
structure…” (Gereffi, Humphrey, & Sturgeon, 2005). In the
follow-up studies that utilize the GVCs framework, the precon-
dition of modularity was taken for granted, and the roles of the
platform leaders were given overly prominence frequently in
explaining the complicated inter-organizational process. Nev-
ertheless, it is running the danger of oversimplification as well
as under-socialization. For example, Kawakami (2011) argued
that “… Intel possesses the power not only to set standards but
also to drive the chain dynamics by redistributing the value-
added in its favor”. However, even the Wintelism did deter-
mine everything, the narrowly defined theoretical framework of
2For example, as Kawakami (2011) shows, Intel had launched the low-
power Centrino platform in 2003, further increased the modularity of the
industry by integrating even more functions within the chip-set, including
wireless internet connection, power management and others.
3Search routines focuson one or another aspect of the firm’s behavior and
capabilities, and come up with proposed modifications which may or may
not be adopted (Dosi & Nelson, 199 4 ).
70

D. Y.-R. YANG, Y.-C. CHEN
GVCs provides very few conceptual tool to further explore
such relevant “black box” (Wen & Yang, 2010). Although the
“value chain modularity” thesis is full of insights, to what ex-
tend does the modularity explain the historical experience and
learning opportunity of the latecomer remains unclear.
The rationale of organizational modularity lies in that tech-
nological modularity lowers the need of embeddedness4 in en-
hancing communication and constraining opportunism. In the
case of modular products, component producers do not need
information about the inner workings of other components
(Baldwin & Clark, 2000). For a modular product, the need for
communication and governance is low and the additional tech-
nical capabilities more than offset the lack of prior experience.
In other words, modular product designs make the benefits of
prior transactions much less important, and firms are able to
reconfigure their supply chains for modular products more
freely. For example, Baldwin and Clark (2000) explored the
increasing modularity in the development of IBM’s System/360,
provided evidence for the impact of increasing product modu-
larity on organizational modularity in the PC industry.
Such rationale was challenged by recent studies yet. For ex-
ample, Brusoni and Principe (2001) traced the aircraft engine
and chemical engineering industries and found that modular
product architectures actually required highly interactive or-
ganizational set-ups to provide the information structure neces-
sary to coordinate the various organizational units involved in
production. Hoetker (2006) used the LCD panel design in note-
book PC as the example and found that while product modular-
ity allows firms to adopt new suppliers more freely, it does not
help firms move activities out of hierarchy. Finally, by illus-
trating the powerful counter-forces that causing organizational
structures to become more integrated in the chip design indus-
try, Ernst (2005) reminded us the danger of generalizing em-
pirical observations that are context-specific and to confound
them with prescription as well as pre d i ct i o n .
Moreover, the “strong modularization school” could under-
estimate the nature of product development as an evolutionary
process to a certain degree, wherein the knowledge involved is
path-dependent and context sensitive (Nelson & Winter, 1982).
Technology and new product development is a dynamic process
that connects a sophisticated circuit of knowledge users and
producers who interact, imitate, adjust, correct, exploit, adapt
and learn mutually. In their evolutionary model of growth, Dosi
and Nelson (1994) conceptualized the technology progress as
an explicit process of search and competition among hetero-
geneous actors, involving the periods of iteration, cumulating as
well as network externalities. Firms’ search process provides
the source of differential fitness and tends to bind them together
as a community (Dosi & Nelson, 1994). The standard inter-
face developed from the modularized technology may not be
able to cope with the “uncertainty” and “novelty” that inherited
from the design process; on the other hand, it may serve as just
the starting point of the “search R&D” assumed to uncover new
techniques or to improve prevailing ones.
From the similar perspective, Sabel and Zeitlin (2004) ar-
gued that the “iterated collaboration” between lead firm and
suppliers was utilized wherein the suppliers could contribute to
the redefinition of interface specifications for new products
based on their experience. That is, the de facto coordination and
system-wide integration alongside the modularized interface
would be required in the product design process, unfolding the
relevance of embeddedness. Economic geographer Michael
Storper also argued the mechanism of “conventions” is de-
voted in the process of technology development, especially in
solving the problems of uncertainty, jointing collective action
and collaborative innovation. It is the “untraded interdepend-
ence” among production networks that contributes to the re-
gional “relational assets”, including informal rules, routines and
socially embedded networks (Storper, 1997).
In short, the global notebook PC production network is not
just constituted of manufacturing process and flows of modular
products and materials. It also includes a complex of know-
ledge interaction and embedded relationships that are coopera-
tive and interdependent by nature. The potentiality of the lead
firms to “drive” production networks through modularization
doesn’t overshadow these inter-dependencies in the networks.
Based on the notion of technology evolution and embedded in-
terdependence, this study challenges two basic and inter-related
assumptions of the modularization thesis:
 The thesis claims that the lead firms and key component
suppliers dictate the bulk of technology designs and shape if
not determine the innovation patterns of other firms in the
production networks. We however, argue that the concrete
process of product development involves the abundant pra-
ctices of tacit knowledge within intra-firm and inter-firm
interactions. The technology standards and specifications
dictated by the lead firms are just the starting points of such
knowledge interac t io n s.
 The thesis assumes the modularization of technology de-
creases mutual dependence among the lead firms and the
contract manufacturers. In contrast, we argue that the mu-
tual dependence among the lead firms and ODMs does not
decrease during the process of modular design. With in-
creasing complexity and shorter design life-spans, the mu-
tual dependency actually increases, leading to regional ag-
glomeration in Taiwan and China.
In the following sections, we would investigate how the tech-
nology and product specifications were defined in the real
“world of production” by analyzing the interactions and organ-
izational governance among the lead firms, key component
suppliers and the ODMs. The empirical study was conducted
through qualitative interviews. From May 2004 to August 2011,
we interviewed 65 professionals from companies located in
Northern Taiwan Region, Greater Suzhou Region and Greater
Dongguan Region in China. The companies included 28 Tai-
wanese ODMs, such as Quanta, Compal, Wistron, Inventec,
ASUSTeK and Foxconn etc.; IPOs (International Procurement
Offices) of 16 name brand companies, including DELL, HP,
IBM, Toshiba, Fujitsu, Hitachi, IngramMicro, Nokia and Acti-
bit; as well as 12 key component suppliers including INTEL,
AMD, Nvidia, Atheros, Philips, Seagate and TI. The positions
of the interviewees ranged from manager to general manager.
The companies in the sample were carefully and systematically
selected to cover the most important firms in the market and in
each region. For example, all top twelve systems manufanctur-
ers, accounting for over 60 percent of the total value of PC-
related industrial output in Taiwan, were included. The inter-
views were designed to reveal both qualitative and quantitative
details about the key network relations in the firm, with a par-
ticular focus on the relationships between the lead firms and
ODMs, and between those ODMs and the key component sup-
pliers. The fieldwork was in part facilitated by ongoing contacts
4For the discussion of e mbeddedness, please refer to Hess (2004).

D. Y.-R. YANG, Y.-C. CHEN
72
with industry representatives (e.g., at technology seminars and
procurement events) through the first author’s position as a
consultant for the Taipei Computer Association (TCA), the lar-
gest PC industry trade union in Taiwan.
engage in constant negotiation with the lead firms and the key
component suppliers to reach a final consensus on the specifi-
cations. There are three key assertions regarding this negotia-
tion process.
 Firstly, innovation on the part of ODMs is based on massive
information and not merely on codified requirements from
the lead firms, especially with regard to negotiating the spe-
cifications.
ODMs in the Global PC Production Networks
In 2006, Taiwanese ODMs accounted for approximately 86%
and 99% of the worldwide production of notebook PCs and
motherboards respectively (Market Information Center, Insti-
tute for Information Industry, 2007). Today, the top 5 tier-one
ODMs—Quanta, Compal, Inventec, Wistron and ASUSTeK to-
gether provide over 70% of the notebook PC production in the
world (Kawakami, 2011). We start by understanding the global
PC production networks in functional and geographical terms
(Figure 1).
 Secondly, the core of innovation by ODMs has been to
integrate the scattered specifications into one adequate
frame and become a complete system, thus having a large
say in the direction of technology development, and not
merely dependent on the lead firm.
 Thirdly, these specifications are not static; they evolve very
rapidly. The innovation process of the ODMs requires con-
stant contact with related companies for information and
knowledge.
In the past, the lead firms focused primarily on product de-
sign and customer relations, while outsourcing the remainder of
the production process to other firms. Today, manufacturing
operations usually are subcontracted by the lead firms to the
ODMs that not only assemble the PC products but also design
the product specifications. Now, most lead firms (or brands)—
such as Dell, HP, Apple and Toshiba—rely on Taiwanese sub-
contractors for manufacturing and product design/development.
Each lead firm tends to contract to two or three Taiwanese
ODMs that have a major share in the lead firm’s total global
laptop production, such as approximate 90% for Dell and HP.
The Northern Taiwan Region (NTR) is a critical node in the
global PC production networks. Although Taiwanese electron-
ics firms have been moving production “offshore” to Southeast
Asia, Europe and, most importantly, China since the early
1990s, the headquarters of the ODMs in the NTR are responsi-
ble for product research and development, technology research,
materials procurement, financial management and marketing
(see Figure 1).
Our analysis involves each stage of product development in
the ODM model as shown in Figure 2. With an “RFP”, or
“Request for Proposal”, the lead firm submits a request to the
system ODMs in Taiwan. The lead firm defines the product’s
initial specifications, which include the whole system’s speci-
fications. Meanwhile the ODMs conduct feasibility studies on
potential market and technology trends. Simultaneously, ODMs
also assess the possible technical providers and feasible com-
ponents. In other words, Taiwanese ODMs must propose a
basic system framework to respond to the client’s requests and
then send a revised proposal to the client.
After handling the “RFP”, the ODMs wait for a Request for
Quote (RFQ) from the lead firm. At this stage, detailed specs
will have been framed; however, the lead firms do not com-
pletely control the drafting of the detailed specifications. The
detailed system specs consist of a “feature list”, among which
certain features are determined by the lead firm while others are
determined by the ODMs. During this process, lead firms and
ODMs exchange ideas frequently. A lead firm’s RFQ serves as
an aid to the ODMs in technical learning as lead firms devote a
lot of effort to the study of user demands. The strength of Tai-
wanese ODMs lies in creating a total solution through system
PC production requires the contracted manufacturer follows
the specifications set by the lead firm. In actual practice how-
ever, in contrast to the modularization thesis, ODMs do not
merely accept a specification from the lead firms. They actually
Figure 1.
The illustration of PC industry’s global production networks. Source: adapted from Yang and Coe (2009).

D. Y.-R. YANG, Y.-C. CHEN
Figure 2.
The product development process in the ODM model. Source: adapted from Yang and Hsia (2007).
integration of related components.
In products such as notebook PC, Taiwanese ODMs are ex-
pected to propose a “total system solution” and system specifi-
cations. In order to do that, ODMs form long-term working
partnerships with the lead firms. The ODMs will recognize the
lead firms’ future needs and undertake related development.
They interact with the component suppliers constantly, espe-
cially their key components suppliers and platform providers,
such as INTEL and AMD, to identify new product features and
functions, as well as to integrate a new system structure based
on these new components. Every month, ODMs develop new
product models regardless of whether or not orders have been
placed.
Is the innovation of ODMs limited to product assembly
based on specifications provided by key component suppliers?
No. Taiwanese ODMs have gradually come to occupy the co-
pilot seat in leading major design and research activities, espe-
cially in systems integration. The features, functions, com-
patibility, security, and cost of the components in a system re-
quire extensive study and familiarity, which is often accom-
plished through interactions and negotiations among firms. The
development of electronics products requires extensive integra-
tion among different components to ensure full support and
conflict-free functioning among all parts. The key is the so-
called “solution”. In terms of “modules”, a system should be
viewed as several modules, not only one.
Some scholar asserted that the PC industry is the exemplar
case of modular production because Intel and Microsoft set the
standards and specs that everyone else has to follow. These
platform leaders are in the driver’s seat so the contract manu-
facturers (including the ODMs and EMSs) have no other choice
but to follow those specs set by Intel and Windows. However,
our detail empirical study showed that the platform leaders
(such as Intel) do not completely control the development of the
entire system. For example, the motherboard might not be a
technology-intensive product in itself, but its integration into
the system remains difficult. This integration relies on thou-
sands of inputs from various components built on the mother-
board. A manager in charge of Intel’s Asia Pacific chipset busi-
ness unit commented:
What we really need is the total solution. What else is there
when our chipset and the clients’ motherboards are both ready?
With PCI-Express, for example, we can test bugs on the moth-
erboard, yet we can’t test the graphics card before it is ready.
Graphics card companies have to test the product themselves.
Therefore, in addition to the hard disk and graphics card com-
panies, equipment companies such as Agilent should have
ready testing equipment. A total supporting network is what we
need, because Intel can’t do everything. Taiwanese ODMs pro-
vide the total solution that integrates Intel’s Chipset.
To summarize, the ODMs retain a certain amount of auton-
omy and control over system technology development, and they
play the role of the integrator between the lead firms and the
component suppliers. Below we will go further to explore two
inter-related questions. Firstly, do the lead firms have complete
control over product technology and related modules? And
secondly, do their relations with Taiwanese ODMs push them
to arm’s length transactions?
Co-Evolution through Co-Location: The
Upgrading of Lead Firm’s IPOs in Taiwan
The lead firms use their IPOs (International Procurement Of-
fices) located in Northern Taiwan as a conduit to establish in-
teractions with ODMs. Instead of starting in the core region and
then sourcing worldwide, the PC industry started “globally”
with many lead firm’s branch offices co-operating with Tai-
wanese suppliers in the region. Starting with IPOs since the
1960s and upgrading to R&D centers in the 2000s, the PC lead
firms continue to maintain their branch offices in Taiwan to in-
teract closely with the ODMs there.
IBM established its first foreign electronics procurement of-
fice in Taiwan in 1966. Its primary function was to procure
related parts and perform certification. After IBM started its PC
business in 1982, the role of its procurement office changed.
IBM’s procurement office began to find Taiwanese products
and then sell them to IBM’s branch factories worldwide. IBM’s
procurement office also played another important role in the
1980’s by familiarizing the emerging PC industry in Taiwan
with future technological trends and with IBM’s management
and technical resources. In other words, IBM’s procurement
office not only served as a conduit for transferring overseas
orders and promoting Taiwanese electronic products, it also
acted as an important agent of technology transfer.
The procurement offices of major lead firms played signifi-
cant roles in the early stages of Taiwan’s PC industry take off.
Under the assistance of these foreign procurement offices,
many ODMs started out as component suppliers, accumulating
their own design capabilities. At the first stage, the contract
manufacturers followed specifications from the lead firms for
early-stage produc t assembly , which i s commonly referred to as
original equipment manufacturer (OEM). However, the “in-
structor” role of major foreign lead firms subsided as the manu-
facturers upgraded their technologies to become ODMs. The
following case studies illustrate the transformation of the lead
firms’ procurement offices after the late 1990’s.
Dell’s Taiwan procurement office was established in 1990,
and it served as a node in the company’s worldwide procure-
ment (WWP) system. The WWP served three functions: first, it
offered global commodity management (GCM) to supplier

D. Y.-R. YANG, Y.-C. CHEN
sourcing, quotation negotiations, and outsourcing management;
second, it acted as the global supply manager (GSM) in charge
of global supplier logistics, stock adjustment, and just-in-time
supplies; and third, it was the SQE (supplier quality engineer-
ing) responsible for quality control. Each of the three functions
weighed differently and varied according to global business
demands5. However, after 2000, Dell’s Taiwan procurement
office was given more decision-making, handling not only pro-
duct procurement but also outsourcing business management.
In 2003, the Dell’s Taiwan procurement office even made pro-
curement and RFQ decisions for Dell’s notebook PCs. In order
to support the new functions, Dell upgraded its presence in
Taiwan by establishing a “technology development center
(TDC)” in Taipei that focused on engineering development.
The co-evolution of the lead firms’ branch facilities with the
Taiwanese ODMs resulted in the upgrading of the lead firms’
facilities. For example, with Dell’s TDC model, the related spe-
cifications of Dell’s notebook PC and RFQ features were de-
termined jointly by the Taiwan procurement office, TDC and
ODMs.
The upgrading of HP’s Taiwan procurement office was no
less than that of Dell. Our interview with HP’s Taiwan pro-
curement office indicated that the office had strong influence
over HP headquarters’ decision-making on issues such as the
technical structure of the products and the choice of ODMs. HP
has also set up a “technology development center (TDC)” in
Taiwan for the development of niche technologies and solu-
tions6. HP and their ODM partners jointly developed the feature
list of a product. The ODM partners contributed the common
features that did not require patented technologies, while the
other features incorporated HP solutions.
“Roadmaps” are the basis of communication between the
lead firms and the ODMs. Generally, a roadmap is lead firm’s
future product plan based on the convergence of industry,
product and technology trends. Not only is roadmap a visual
display of technological performance, it is also confidential.
When the lead firms’ technology development centers in Tai-
wan plan new products after 2000s’, they begin to request a
roadmap from ODMs. The discussion of the roadmap requires
“face-to-face” communication; otherwise it is difficult to un-
derstand a roadmap. Many unique solutions are incorporated
into the map, and face-to-face communication can facilitate the
exchange of ideas and minimize misunderstandings. At this
stage, communication between lead firms and ODMs is not sim-
ply one way, i.e., from lead firms to ODMs. On the contrary,
the lead firms also receive new knowledge from ODMs. The
partnership between HP and the ODMs, for example, has fur-
thered the exchange of roadmaps, resulting in the “co-design”
projects.
Beyond Arm’s Length Relationship
Because of the importance of the ODMs, we wonder if the
transactional relationship between the lead firms and the ODMs
will shift toward the arm’s length mode as suggested by the
modularization thesis. In our fieldwork, we discovered that
there are at least three modes of governance between the lead
firms and the ODMs regarding system-level roadmap formation,
knowledge exchange, and the practical transaction mechanisms.
The modes are “long-term relationships,” “institutionalized price
competition,” and “arm’s length transactions”. These modes
overlap and interact with and affect each other.
Long-Term Relationships
As Taiwanese ODMs provide the lead firms with abundant
technology resources, more solutions are being accepted and
included in the lead firms’ specification requirements in this
pattern. Such long-term relationships result in more orders from
the lead firms to the ODMs. The key to a “long-term relation-
ship” rests in the ODMs’ proposed solutions being accepted by
the lead firms. This relationship is often found between Tai-
wanese ODMs and Japanese lead firms and also some Ameri-
can lead firms. For instance, Dell and former Gateway estab-
lished long-term relationships with Quanta and Compal in Tai-
wan in late 90’s. Under such relationships, these ODMs were
guaranteed to have a higher investment return rate. But even
under the long-term relationships, Taiwanese ODMs retain a
certain level of autonomy (instead of becoming “captive”),
which also ensures the openness of transaction networks and
technological reference resources.
Institutionalized Price Competition
In this pattern, the lead firms’ choice of ODMs is based on
price competition. However, this doesn’t mean that the cus-
tomer-supplier relationship is arm’s length. For instance, HP’s
outsourcing policy is the most “opportunistic” among American
companies, and its “online bidding system” forces Taiwanese
ODMs into intense price competition. However, similar to
long-term relationships, HP’s specification decisions also initi-
ate intensive information exchange and communication, as well
as roadmap exchanges with the ODMs. The difference is that
HP may add one ODM’s solutions to its RFQ but still open the
order bidding to all ODMs. In other words, HP may “capture”
an ODM’s intellectual property and then proceed with its cost-
based bidding system.
We observed that many American and Japanese lead firms
adopt “long-term relation ships” and “institutionalized price com-
petition” simultaneously through a quarterly business review
(QBR)—a benchmarking mechanism–to include several appro-
priate ODMs in its available list (AVL) for closer relationships.
For instance, former Gateway first established long-term rela-
tionships with Quanta, then after the acquisition of the E-Ma-
chine, its procurement policy was geared towards the “institu-
tionalized price competition” model. EPSON and HITACHI
also adopted the combination of the two modes under cost re-
duction and competitive pressure from their headquarters.
Arm’s Length Transaction
The third cooperation mode—“arm’s length transaction”
happens in the transaction between Taiwanese ODMs and the
“clones”. The “clone” refers to the non-brand sellers, regional
distributors or minor brands. When the leading brand compa-
nies do not accept an ODM’s proposed product solution, the
ODM would sell it to those clones. The clone buyers do not
develop system specifications such as the leading brand com-
panies do. Instead, the ODMs fully control all the specifications
5The SQE function had moved to the manufacturing location in China.
6In addition to DELL and HP, EPSON established a Sourcing and Manu-
facturi ng Service Cen ter in Tai pei in 20 01. The Tai wan bran ch is in ch arge
of selected corporate products at the design stage, its work includes setting
specifications, providing working samples, performing pilot runs and qua-
lity control. I BM has also established a server de s ign center in Taipei.
74

D. Y.-R. YANG, Y.-C. CHEN
and design for such clones. It is for the marketing purposes that
the ODMs cooperate with the clones (such as Actibit & Dixon).
These types of transactions often take place at COM-PUTEX
TAIPEI, the second largest computer trade show in the world,
where thousands of less-known or non-brand computer sellers
from around the world come to search for suitable ODMs. In
this mode, the ODMs have autonomy to decide which models
are appropriate for that country’s market.
With a historical perspective, although foreign lead firms
play an important role in Taiwanese ODMs’ technology learn-
ing, their dominance over product specifications is gradually
decreasing. In their cooperation with ODMs after 2000s’, for-
eign lead firms neither play their roles as the modular definer
nor do they engage in arm’s length transactions. Arm’s length
transactions do exist, but only in supplementing the “long-term
relationships” and “institutionalized price competition” modes
of governance.
Mutual Learning between Key Component
Suppliers and ODMs
In this section, we want to explore the interaction between
the key component suppliers (CPUs, chipsets, graphic chip,
display panels, memory, wireless chips, and hard drives) and
the ODMs. Furthermore, we also investigate if key component
suppliers hold full control over related specifications and if
Taiwanese ODMs merely follow the specifications.
In order to adequately understand component functionality
and to integrate proper solutions with the features list, ODMs
communicate and exchange knowledge with key component
suppliers frequently. In addition, since the ODMs are more
sensitive to the market than are the key component suppliers,
learning from the ODMs is critical to those suppliers. The in-
teraction and negotiation is part of a routine.
The transactions between key component suppliers and the
ODMs are not arm’s length transactions. When component
suppliers promote their products, this usually triggers a know-
ledge exchange on technology development and specifications.
Such communication must be done face-to-face because much
of the technology content needs to be explained in full detail.
The opinions and knowledge exchanges revolve around the de-
velopment of the technology roadmap and affect co-develop-
ment of technology in the future as well. We argue that cu-
mulated product innovation and system integration knowledge
of Taiwanese ODMs is helpful to key component suppliers not
only in terms of technology learning but also in terms of speci-
fication developments. We also found the capabilities of the
ODMs were result from long-term research on extensive com-
ponent specifications and future technological trends. The
scope of technical knowledge for system integration is a valu-
able reference for key component suppliers who focus on their
own product development. With such reciprocal knowledge
exchanges, some ODMs actually take the lead in certain speci-
fication developments of key components. As a R&D executive
at Quanta said:
A Nvidia CEO will bring its R&D managers to meet with us
for future projects and the specifications required. We lead the
process, but most people don’t know this.
In short, although key component suppliers—potential plat-
form leader—own innovation resources to magnify their lead-
ing role in technology specification, they still need the consent
and mutual understanding from major ODMs. These ODMs are
the “first-tier followers”, usually have influence on key com-
ponent suppliers’ specification decisions.
We then turn to another assertion by the modularization the-
sis that ODMs merely build their innovation activities on fea-
tures provided by key component suppliers. As seen in Figure
2, if this assertion was true, ODMs would have been slacking
off for the three to six months starting from “kick-off” to mass
production. This is obviousl y an unlikely scenar i o.
For example, to customize or modify WLAN requires a large
amount of development work as well as collaborative innova-
tion. In assembling notebook PCs with strong multimedia ap-
plications, ODMs must think carefully about how to integrate
the WLAN module with the CPU and OS specifications. On the
other hand, WLAN chip suppliers such as TI, Broadcom and
Atheros sometimes have to collaborate with their ODM clients’
R&D staffs to develop various special applications based on
their chip design. In other words, when it comes to modulariza-
tion, the final integration of key components is not fully domi-
nated by the key component suppliers; the R&D resources of
the ODMs also play a critical role in developing applications
and solutions.
WLAN is not the only example of collaborative innovation.
Many developments and applications for key components share
this feature, including the most standardized hard drives. From
the preliminary design by the key component suppliers to the
applications of the ODMs, the process consists of a series of
knowledge exchanges. Again, the ODMs play the role of the
relevant technology integrator. They integrate parts and solu-
tions from a variety of key component suppliers into comer-
cialized technologies; then they provide feedback on future
technology development to these suppliers. The ODMs provide
valuable tacit knowledge for the key component suppliers and
influence their technology paths.
From the above observation, we learn that the initial speci-
fications of component suppliers need to undergo a complex
process of modification to be applied in system design, and that
the process involves intense knowledge exchanges between
component suppliers and the ODMs. Their relationships are far
from what is defined as the arm’s length transactions. On the
contrary, their relationships should be regarded as “technical
partnerships of co-design” that are beneficial to both parties.
We will further demonstrate this point by examining the role
of several Taiwan branch offices of overseas key component
suppliers. We take the TI WLAN chip as an example. The
company’s WLAN chip design team is located in France, while
the field application engineers (FAE) and the associate engi-
neers of its marketing and technical support departments are
located in Taipei. The goal of its Taipei office is to relay cus-
tomer needs to the TI headquarters. Another example is the
major video chip company ATI, which has its Asia-Pacific
headquarters in Taipei. ATI’s R&D centers are located in To-
ronto, Silicon Valley, and on Route 128 in Boston, with around
60 technical support engineers in Taipei playing the important
role of giving the feedback learned from the co-design process
with the ODMs to the headquarter. An executive from the com-
pany remarked:
The biggest benefit in setting up our Asia-Pacific headquar-
ters in Taiwan is taking our partners into consideration. Our
headquarters, of course, will launch new product roadmaps
every quarter, and we will then give feedback on products that
cater to local demand, making for reciprocal product planning.
Though the primary function of AMD Taipei is to promote

D. Y.-R. YANG, Y.-C. CHEN
its CPU, it also has a design department called TATS (Taiwan
Application and Technical Support). AMD’s headquarters re-
tains large control over the CPU design, while TATS partici-
pates in some aspects of product design and specification plan-
ning. Intel Taipei is similar in this respect. The interviewees
from these two companies considered the Taipei branches as
development centers, as they both assist their customers to de-
velop products based on their technical platforms and then give
feedback on the future demands of Taiwanese customers to
headquarters. Hence, the functions of the Taipei branch offices
would have extensive influence on further technology deve-
lopments for these two platform providers.
Co-Evolution in Practice: The Routine
We’d like to illustrate the technological-organizational fea-
tures of the mutual dependence and co-evolution discussed
above by exploring the practices of Taiwanese research team-
work from the perspective of “search R&D” and “search rou-
tine” (Dosi & Nelson, 1994). According to the real world prac-
tice, we make a basic distinction between product/project man-
agers (PM) and those staffs who work in first line research and
development (RD).
The main responsibility of a PM is to search for technical
resources provided by suppliers, as well as feasible solutions. In
the beginning, PM needs to research on the technical resources,
the primitive definition and integration. Most PMs of Taiwan-
ese ODMs have the R&D backgrounds; it’s usually the senior
RD engineers have sufficient knowledge to be the PMs. The
PMS spend most of their time researching technological trends
and technical resources, and must communicate with the com-
pany’s marketing department and their component suppliers. In
particular, a PM would research new technical resources and
suggest primitive solutions for further execution by RD. The
RD engineers test potential solutions for features such as com-
patibility, electronics and heating ability. The RDs usually
communicate with the FAEs (field application engineers) of
their component suppliers in order to solve more detailed prob-
lems.
PMs are the important agents of innovation for Taiwanese
ODMs, as they provide the window for technology exchanges
among companies. When component suppliers sell products to
the ODMs, they contact the ODMs’ PMs, who then carry out a
routine business interaction. The routine is not merely one-way
communication for component suppliers to provide technical
information; the suppliers also gather knowledge about future
technology development planning and learn the market accep-
tance of their products from the ODMs.
Computer systems, for example, have thousands of compo-
nents, and the RD engineer’s role is to make them work to-
gether to reach expected performance. This requires sufficient
knowledge of each component’s characteristics and their inte-
grated functions, as well as the ability to perform cost-per-
formance evaluations. Hence, RDs often need to try different
combinations of components, and decide on the best solution
based on actual testing results. Therefore, the real “modulariza-
tion” takes place at the internal design stage performed by
ODMs. In other words, those solutions require RD engineers to
integrate them into a system. A primitive design can also be
re-modularized after the R&D process. In other words, although
the key component suppliers such as INTEL did offer complete
reference designs, including recommended system layout and
software, large amounts of modification and customization
were regarded often as necessary for the ODMs with internal
expertise to finish a competitive product innovation.
With regard to the “iterated collaboration” (Sable & Zeitlin,
2004), we categorize the process into intra-firm and extra-firm
learning. With the intra-firm learning, most companies arrange
training courses for RD engineers; however, the majority of
technical learning happens through “on-site traini ng” and “learn-
ing by doing”. In other words, senior engineers in a company
are important because they pass on the tacit knowledge inter-
nally. The accumulation of tacit knowledge becomes the “own-
ership advantage” of each company. Such experience is diffi-
cult to codify and require new personnel to become accustomed
through experimental trial and error.
We observed that a mature R&D engineer must have three
years of hands-on experience in order to independently lead
new product development. Through the “learning by doing”
teamwork, each company can produce many different solutions
for a problem instead of just one. Even though the key compo-
nent suppliers provide a detailed design guide or circuit map, or
even with Intel’s detail “data sheet”, not many ODMs’ RDs do
their job by following these given information. For example, a
RD manager at Asustek said,
Aggressive RD engineers try to solve problems on their own
and making something good out of something ordinary. The
most difficult part is to create the best performance at the low-
est cost.
On the other hand, the extra-firm learning for ODMs’ RDs is
often based on the occupation-based social networks, most fre-
quently on social exchange with key component suppliers’
FAEs for the object of “debugging”. A “bug” is a problem that
arises when a component proves incompatible with the system
designed by ODMs or interferes with certain functions. Bugs
often occur in the development of a new product, and the
ODMs will refuse a supplier’s components if the supplier can-
not solve the problem, unless it is client-appointed. The trou-
ble shooting interaction between ODMs’ RDs and key compo-
nent suppliers’ FAEs is mutually beneficial for technical learn-
ing. For example, a chipset may all show the same performance
under the Intel framework, but the system design might differ
among the ODMs. When a bug occurs, the FAEs usually work
together with the RDs firstly to modify the ODM’s system de-
sign before changing their original chipset design. For the
ODMs, FAEs provide the equivalent of technology support,
and are even considered as part of the ODMs’ R&D team.
Moreover, the key component suppliers usually feel they better
understand the features and market potential of their products
through the direct interactions with the ODMs’ RDs.
In brief, those “learning by searching” and “learning by de-
tection” behaviors within the production networks are essential
elements for the ODM model. Institutionalized learning net-
works are created through both formal organizational arrange-
ment (such as periodical roadmap updates in conjunction with
component suppliers) and informal social networks. These
“pragmatic collaborations” (Helper, MacDuffie, & Sabel, 2000)
have developed into routines among the technical teamwork
between the key component suppliers and the ODMs.
Conclusion
Whether there were limits to global sourcing? This research
explores one of the limitations with the perspective of organiza-
76

D. Y.-R. YANG, Y.-C. CHEN
tional co-evolution. We argue the mutual dependence among
the lead firms and ODMs does not decrease during the process
of technological modularization in the notebook PC industry.
The process of new product development requires plenty of
technical interaction and tacit knowledge exchanges among
lead firms, key component suppliers and ODMs. The mode of
inter-firm governance is far beyond the “arm’s length transac-
tions”. Moreover, Taiwanese ODMs play an irreplacable role in
system and component specifications due to the rapid technol-
ogy chang and the inherited complexity of system integration in
the notebook PC industry.
Furthermore, the “strong modularity thesis” fails to explain
the competitiveness of the ODMs. The ODMs’ innovative pat-
tern drives the widely organizational and technological interac-
tion in order to avoid the bottleneck of system innovation with
standardized technical interface and the limitation of an out-
dated industrial structure. The knowledge exchanges occurring
through single modular and inter-modular integration have
become an important mechanism for mutual technical learning
and developing core competences. Taiwanese ODMs are able
to serve as important sources of information and knowledge for
specifications formulation through their system integration and
technological development abilities. Under this circumstance,
key component suppliers such as Intel would have to rely more
on the intensive collaborations with the ODMs.
This article contributes to update the understanding of East
Asia’ ODM development7, an upgrading pattern for the late-
comers besides the branding strategy (Chu, 2009) and differs
from South Korea’s experience (Cyhn, 2002; Hobday, 2003).
The implication for world development is the latecomers might
get engaged in the technology development through the inte-
gration of absorptive capacity, cumulative innovation and de-
sign capability as well as participating in the global production
networks; while the “value chain modularity” would not auto-
matically contribute to the learning opportunity for latecomers
(Frigant & Layan, 2009). The organizational co-evolution, mu-
tual dependence and the need for intensive face-to-face in-
teraction in the “modular” design process further explain the
territorialization of the industry in the “design cluster” in Nor-
thern Taiwan Region. The region has developed routines from
intense learning networks and become a crucial technology
district in bridging the United States and Mainland China (Hsu
& Saxenian, 2000; Saxenian & Hsu, 2001).
Finally, we have to notice the ODM model has its limitation
in catching-up for the small economy without significant do-
mestic ma rket: the weak brandi ng capabilities, weak bargaining
power with the TNCs, the “IP barriers” that constraint the de-
velopment of home-made intellectual property rights (Ernst,
2010) and the job loss owing to the relocation of manufacturing
activities. In addition, the national innovation system in Taiwan
does not have integrated promotion strategy for industry spe-
cifications. Contributions from investing in fundamental re-
search and human resources, initiating the standard-setting col-
laboration as well as promoting the own-brand strategy that
complement the ODM production networks would be the dy-
namic forces for upgrading the technology district in the future.
Acknowledgements
Daniel You-Ren Yang acknowledges the funding provided
by National Science Council of Taiwan ROC (NSC 101-2628-
H-029-002-MY2).
REFERENCES
Amsden, A. H., & Chu, W. W. (2003). Beyond late development: Tai-
wan’s upgrading policies. Cambridge, MA: MIT Press.
Baldwin, C. Y., & Clark, K. B. (2000). Design rules: The power of
modularity. Cambridge, MA: MIT Press.
Berggren, C., & Bengts so n, L. (2004). Rethinking outsourcing in ma nu-
facturing: A tale of two telecom firms. European Management Jour-
nal, 22, 211-223. doi:10.1016/j.emj.2004.01.011
Borrus, M. (2000). The resurgence of US electronics: Asian production
networks and the rise of Wintelism. In: M. Borrus, D. Ernst, & S.
Haggard (Eds.), International production networks in Asia: Rivalry
or riches (pp. 57-79). London and New York: Rout l e d g e .
Borrus, M., & Zysman, J. (1997). Globalization with borders. Industry
and Innovation, 4, 141-166. doi:10.1080/13662719700000008
Brusoni, S., & Prencipe, A. (2001). Unpacking the black box of modu-
larity: Technologies, products and organizations. Industrial and Cor-
porate Change, 10, 179-205. doi:10.1093/icc/10.1.179
Campagnolo, D., & Camuffo, A. (2010). The concept of modularity in
management studies: A literature review. International Journal of
Management Reviews, 12, 259-283.
Chandler, A. (1962). Strategy and structure: Chapters in the history of
american industrial enterprise. Cambridge, MA: Harvard University
Press.
Chandler, A. (1977). The visible hand: The managerial revolution in
american business. Cam bridge, MA: Harvard University Press.
Chandler, A. (1990). Scale and scope: The dynamics of industrial capi-
talism. Cambridge, MA: Harvard University Press.
Chen, S. H. (2002). Global production networks and information tech-
nology: The case of Taiwan. Industry and Innovation, 9, 249-265.
doi:10.1080/1366271022000034480
Chu, W. W. (2009). Can Taiwan’s second movers upgrade via branding?
Research Policy, 38, 1054-1065. doi:10.1016/j.respol.2008.12.014
Cyhn, J. W. (2002). Technology transfer and international production:
The development of the electronics industry in Korea. Cheltenham:
Edward Elgar.
Dedrick, J., & Kraemer, K. L. (1998). Asia’s computer challenge:
Threat or opportunity for the United States and the world? New
York: Oxford University Press.
Dedrick, J., Kraemer, K. L., & Linden, G. (2010). Who profits from
innovation in global value chains? A study of the iPod and notebook
PCs. Industrial and Corporate Change, 22, 1-36.
Dosi, G., & Nelson, R. R. (1994). An introduction to evolutionary the-
ories in economics. Journal of Evolu t i on a r y Economics, 4, 153- 172.
doi:10.1007/BF01236366
Ernst, D. (2005). Limits to modularity: Reflections on recent develop-
ments in chip design. Industry and Innovation, 12, 303-335.
doi:10.1080/13662710500195918
Ernst, D. (2010). Upgrading through innovation in a small network eco-
nomy: Insights from Taiwan’s IT industry. Economics of Innovation
and New Technology, 19, 295-324. doi:10.1080/10438590802469560
Frigant, V., & Layan, J. B. (2009). Modular production and the new
division of labour within Europe: The perspective of French automo-
tive parts suppliers. European Urban and Regional Studies, 16, 11-
25. doi:10.1177/0969776408098930
Gereffi, G., Humphrey, J., & Sturgeon, T. J. (2005). The governance of
global value chains. Review of International Political Economy, 12,
78-104. doi:10.1080/09692290500049805
Helper, S., MacDuffie, J. P., & Sabel, C. (2000). Pragmatic collabora-
tions: Advancing knowledge while controlling opportunism. Indus-
trial and Corporate Change, 9, 443-488. doi:10.1093/icc/9.3.443
Hobday, M. (1995). East Asian latecomer firms: Learning the technol-
ogy of electronics. World Development, 23, 1171-11 93.
doi:10.1016/0305-750X(95)00035-B
Hobday, M. (2003). Innovation in Asian industrialization: A Gerschen-
kronian perspective. Oxford Development Studies, 31, 293-314.
Hoetker, G. (2006). Do modular products lead to modular organizations?
Strategic Management Journal, 27, 501-51 8. doi:10.1002/smj.528
7For the development of the earlier OEM stage, please refer to Chen (2002).

D. Y.-R. YANG, Y.-C. CHEN
78
Hsu, J. Y., & Saxenian, A. (2000). The limits of guanxi capitalism:
Transnational collaboration between Taiwan and the USA. Environ-
ment and Planning A, 32, 1991-2005. doi:10.1068/a3376
Kawakami, M. (2011). Inter-firm dynamics in notebook PC value
chains and the rise of Taiwanese original design manufacturing firms.
In: M. Kawakami, & T. J. Sturgeon (Eds.), The dynamics of local
learning in global value chains: Experiences from East Asia (pp.
16-42). Basingstoke & New York: Palgrave Macmi l l an.
Lamoreaux, N. R., Raff, D. M. G., & Temin, P. (2003). Beyond mar-
kets and hierarchies: Toward a new synthesis of American business
history. American Historical Review, 108, 404-433.
doi:10.1086/533240
Langlosi, R. N. (2002). Modularity in technology and organization.
Journal of Economic Behavior and Corporate Change, 49, 19-37.
doi:10.1016/S0167-2681(02)00056-2
Langlosi, R. N. (2003). The vanishing hand: The changing dynamics of
industrial capitalism . Industrial an d Corporate Change, 12, 351-385.
doi:10.1093/icc/12.2.351
Langlosi, R. N. (2004). Chandler in a larger frame: Markets, transaction
costs, and organizational form in history. Enterprise and Society, 5,
355-375. doi:10.1093/es/khh055
Nelson, R., & Winter, S. (1982). An evolutionary theory of economic
change. Cambridge, MA: Beknap Press.
Sabel, C. F., & Zeitlin, J. (2004). Neither modularity nor relational
contracting: Inter-firm collaboration in the new economy. Enterprise
and Society, 5, 388-403. doi:10.1093/es/khh057
Saxenian, A., & Hsu, J. Y. (2001). The Silicon Valley-hsinchu connec-
tion: Technical communities and industrial upgrading. Industrial and
Corporate Change, 10, 893-920. doi:10.1093/icc/10.4.893
Steinle, C., & Schiele, H. (2008). Limits to global sourcing? Strategic
consequences of dependency on international suppliers: Cluster the-
ory, resource-based view and case studies. Journal of Purchasing &
Supply Management, 14, 3-14. doi:10.1016/j.pursup.2008.01.001
Storper, M. (1997). The regional world: Territorial development in a
global economy. New York: Guilford Press.
Sturgeon, T. J. (2002). Modular production networks: A new American
model of industrial organization. Industrial and Corporate Change,
11, 451-496. doi:10.1093/icc/11.3.451
Sturgeon, T. J. (2003). What really goes on in Silicon Valley? Spatial
clustering and dispersal in modular production networks. Journal of
Economic Geography, 3, 199-225. doi:10.1093/jeg/3.2.199
Sturgeon, T. J., & Kawakami, M. (2011). Global value chains in the
electronics industry: Characteristics, crisis, and upgrading opportuni-
ties for firms from developing countries. International Journal of
Technological Learning, Innovation and Development, 4, 120-147.
doi:10.1504/IJTLID.2011.041902
Sturgeon, T. J., & Lee, J. R. (2005). Industrial co-evolution: A com-
parison of Taiwan and north American electronics contract manu-
facturers. In: S. Berger, & R. K. Lester (Eds.), Global Taiwan: Build-
ing competitive strengths in a n ew international economy (pp. 33-75).
Armonk: An East Gate Book.
Sturgeon, T. J., Biesebroeck, V., & Gereffi, G. (2008). Value chains,
networks and clusters: Reframing the global automotive industry.
Journal of Economic Geog rap hy , 8, 297-321. doi:10.1093/jeg/lbn007
Wen, H., & Yang, D. Y. R. (2010). The missing link between techno-
logical standards and value-chain governance: The case of patent-dis-
tribution strategies in the mobile-communication industry. Environ-
ment and Planning A, 42, 2109-2130. doi:10.1068/a41203
Yang, Y. R., & Hsia, C. J. (2007). Spatial clustering and organizational
governance of trans-border production networks: A case study of
Taiwanese information-technology companies in the Greater Suzhou
Area, China. Environment and Planning A, 39, 1346-1363.
Yang, D. Y. R., & Coe, N. M. (2009). The governance of global pro-
duction networks and regional development: A case study of Tai-
wanese PC production networks. Growth and Change, 40, 30-53.
doi:10.1111/j.1468-2257.2008.00460.x
Yung, I. S., Lee, H. W., & Lai, M. H. (2009). Competitive advantages
created by a cluster collaboration network for supplier management
in notebook PC production. T ot a l Q uality Management, 20, 763-775.
doi:10.1080/14783360903037358