You are on page 1of 15

Technological Forecasting & Social Change 74 (2007) 298 – 312

Technological interdependence and knowledge diffusion in the

building of national innovative capacity: The role of
Taiwan’s chemical industry
Mei-Chih Hu a,*, Chun-Yao Tseng b
Graduate Institute of Management of Technology, Feng Chia University, Taichung 407, Taiwan
Department of Business Administration, Tunghai University, Taichung 407, Taiwan
Received 3 December 2005; received in revised form 20 March 2006; accepted 26 April 2006


While the significance and effectiveness of patents in the chemical industry has been demonstrated in many
industrialized countries, this study examines the role of the chemical industry and knowledge diffusion in building
the innovative capacity of a nation in latecomer country Taiwan. The development of process innovation plays an
integral role in the strategic industries of Taiwan, but few attempts have been made to address how the efficiency of
process development can be enhanced. As a latecomer, Taiwan has built its national innovative capability on
strategic industries such as semiconductors, consumer electronics and flat panel displays. Through patent data
analysis, this study demonstrates the significant and indispensable role played by the chemical industry in
technological interdependence and knowledge diffusion with other Taiwanese strategic industries. This study
suggests that while the public resources of Taiwan are focused on accelerating the development of emerging
sectors and technologies, the chemical industry serves as an effective linkage and catalyst in problem-solving.
D 2006 Elsevier Inc. All rights reserved.

Keywords: Technological interdependence; Knowledge diffusion; National innovative capacity; Process innovation; Chemical

* Corresponding author.
E-mail addresses: (M.-C. Hu)8 (C.-Y. Tseng).

0040-1625/$ - see front matter D 2006 Elsevier Inc. All rights reserved.
M.-C. Hu, C.-Y. Tseng / Technological Forecasting & Social Change 74 (2007) 298–312 299

1. Introduction

Numerous studies have demonstrated, by cross-country comparisons over time, that future rate of
growth depends on the present rate of growth, which in turn depends on past decisions (see Refs [1–3]).
It is a fact of economic development that investment, whether in terms of physical, intangible and human
resources, stimulates further investment. Accordingly, industries apply a specific bindustrial policyQ on
the assumption that the government can guarantee future prosperity by providing further resources and
support. From supporting basic research to the building infrastructure and establishing regulations,
government policy makers can define industries in terms of national economic development and
influence the fortunes of individual firms, something especially true in the East Asian latecomer
countries given that government intervention has been a critical contributor to the East Asian dmiracleT.
The chemical industry plays a role as a general source of technologies for the entire economic system,
and its knowledge diffusion network has expanded encompassing a broad range of industries [4]. Since
the chemical industry produces numerous industrial intermediates that generate direct and indirect links
with sectors such as textiles, health, agriculture, automobiles, housing, and consumer durables, cyclicity
via co-evolution reinforces its network structure associated with its related industries. Co-evolution
involves interlinked and dependent growth, which is dominated by adaptation and selection of
technology options, and has become the basis of competitiveness in the chemical industry.
The history of the chemical industry in industrialized countries since the 1840s has clearly
demonstrated the interdependence between the governmental-initiated policy and the actions of
individual firms [5,6]. Therefore, identifying a method of stimulating new technology-based industries
has been challenging for both industrialized and industrializing countries, particularly in relation to
devising public policies that would significantly impact launching or accelerating new technology-based
industry development] [7].
Technological developments in the chemical industry have been cumulative innovation driven [8].
The synthetic organics industry based on coal tar revolutionized dyestuffs, eventually paving the road to
plastics, synthetic fibers, and modern pharmaceuticals. Moreover, growth in demand for polymers,
fertilizers, drugs, pharmaceuticals, plastics, cosmetics, dyes and detergents, has driven the industry as a
whole, demonstrating that the strength of the chemical industry derives from the co-evolution of network
relationships among firms resulting from interlinked and dependent growth. For instance, the traditional
and labor-intensive textile industry developed into a high-tech and capital-intensive industry from the
late 1960s because of innovations in the machinery and the chemical industries.
The chemical industry is the oldest high-tech industries, and remains one of the largest manufacturing
industries. The chemical industry is also one of the few industries in which patents have traditionally
been important and effective (see Ref. [9] for an example)1. One possible reason for this is that the
development of chemical engineering improved the usefulness of process patents and encouraged
process patenting.
The history of the chemical industry also demonstrates that the role of patents is considerably broader
than simply excluding competitors, and has changed in response to changes in industry structure [10,11].

The well-known Yale survey suggested that for most bhigh-techQ industries (except chemicals) patents are less effective.
Respondents from the chemical industry rated product patents as extremely important. Furthermore, while process patents were
perceived as ineffective in other industries, in the chemical industry they were rated as effective.
300 M.-C. Hu, C.-Y. Tseng / Technological Forecasting & Social Change 74 (2007) 298–312

How patents are used, or not used, impacts entry opportunities and industry structure, as well as the rate
of technological change itself.
In latecomer countries such as Taiwan, the chemical industry supplies its traditional giants textiles and
plastic products for new emerging industries such as semiconductors, computers, and electronics,
illustrating some of the ways that policy makers can exploit the development of emerging technologies
to create new industries. While petrochemical-related industries comprised a substantial percentage of
Taiwanese manufacturing output over the past years, and the technological advance of the chemical
industry has been cumulative and co-evolution with numerous industries, this study examines the
question of what path government and industry should follow to create a sustainable competitive
advantage in the chemical industry for the nation.
The distinctive drivers of building national innovative capacity in the East Asian latecomer
countries (including Taiwan, South Korea, Hong Kong, Singapore, and China) are illustrated in the
empirical work by Hu and Mathews [12,13], and are distinguished from those in advanced countries
[14]. When the drivers of national innovative capacity in the advanced country are broadly relied on
various elements in innovative infrastructure, industrial clusters and the linkages between innovative
infrastructure and industrial clusters, the national innovative capacity in latecomer countries is more
limited and dependent on previous knowledge stock, the R&D manpower, private R&D expenditure,
and specialization in the high-tech industry along with public R&D funding. The most notable finding
of HU and Mathews’s study was that the role of public R&D expenditure is significant, and changes
over time. While public R&D expenditure is found to strongly enhance the contribution of patenting
specialization in three key industrial sectors (namely Chemical, Electrical and Mechanical), which are
the drivers of innovation and knowledge diffusion in Taiwan, the role of the chemical industry is
found evolving from the innovative infrastructure provider into a function of linkages between new
emerging industries and innovation capacity; in particular, chemical engineering and process provide
numerous possibilities for process innovation.2 The most significant scientific and technological
challenges facing the chemical industry are realizing the considerable economic and societal benefits in
the form of applications [16].
This study thus addresses the implications of such diverse changes in product, process and
application innovations in terms of altering the role of the chemical industry within the national
innovation system in latecomer countries (such as Taiwan). The rest of this paper is organized as
follows. Section 2 identifies the significance of the chemical industry for knowledge diffusion, process
innovation and co-evolution in Taiwan, and is followed by an empirical study in Section 3 examining
technological interdependence in the establishment of national innovative capability between the
Taiwanese chemical industry and other strategic industries, which is be further demonstrated through a
case study of the Taiwanese TFT-LCD (Thin Film Transistor-Liquid Crystal Display) industry in Section
4. Given the importance of the public sector in the East Asian latecomer countries, Section 5 discusses
the role of the Taiwanese public sector in building national innovative capacity based on the perspective
of the co-evolution of the chemical industry, and Section 6 presents conclusions as well as discussing the
implications of the changes and linkages between the Taiwanese chemical industry and emerging

The incentives to adopt emerging technologies for the Taiwanese chemical industry are also demonstrated in the survey by
MOEA [15], in which the main incentive driving investment in the chemical industry is developing new products and
M.-C. Hu, C.-Y. Tseng / Technological Forecasting & Social Change 74 (2007) 298–312 301

2. Knowledge diffusion and process innovation

Knowledge diffusion might be as important as knowledge creation to ensuring the effective working
of national innovation system [17]. In this respect, the evolution of chemical industry knowledge and its
transfer to the emerging high-tech industries (for example semiconductors, TFT-LCD, pharmaceuticals
and biotechnology) is especially important for latecomer countries such as Taiwan. Various attempts
have been made to characterize knowledge variation among contexts [18–20]. While not all inventions
are patentable, pharmaceuticals, chemicals, semiconductors, and electronics rely on patents to protect
their inventions [9]. Particularly, the chemical industry has enormous complementarities with sectors
such as agriculture, automobiles, petroleum, and semiconductors, and is characterized by the fact that
overall industry development has accumulated underlying theoretical and practical knowledge, which is
sufficiently deep to enable the development of process innovation.
The chemical industry has become one of the most influential industries in Taiwan in terms of diffusing
knowledge to downstream industries as well as in promoting upstream production, particularly in the chemical
materials sector [21]. Since other chemical-related sectors such as petrochemicals, plastic and rubber are all
important drivers of the industrial development of Taiwan, the chemical industry is a highly interdependent with
the overall industry and thus acts as a driver of innovation. This demonstrates the importance of fostering the
development of the upstream chemical industry, especially in the chemical materials sector, to increase the
competitive advantage of Taiwan, which is a continuous supplier to the thriving downstream export sectors.
Notably, the development of chemical engineering caused process innovations to be separated from product
innovations, which is important to the viability of process specialists in latecomer countries such as Taiwan.
The chemical industry has historically been a cyclical industry, limiting its development and
interdependence. In this situation, numerous chemical firms have shifted their focus to specialty
chemicals such as lubricants, paints, adhesives, and catalysts, which have high value and low cyclicity
[22]. However, for a latecomer country such as Taiwan, the advantage of upgrading from basic industrial
chemicals to specialty chemicals is minimal, and basic industrial chemicals continue to comprise 80% of
industry production, with only 17% contributed by specialty chemicals and 3% by pharmaceuticals [23].

Export Import
Petrochemicals Feedstock and plastics products Specialty chemicals

Fig. 1. Export and Import of Taiwan’s Chemical Industry, 2002. Source: ITRI [23].
302 M.-C. Hu, C.-Y. Tseng / Technological Forecasting & Social Change 74 (2007) 298–312

The unbalanced development of the Taiwanese chemical industry has caused the trade disequilibrium of
specialty chemicals in supporting the rapid growth of advanced technologies. Fig. 1 displays the demand
for specialty chemicals in Taiwan in 2002, showing they comprise 68.1% of total chemical imports,
making them a major contributor to the trade deficit of the Taiwanese chemical industry.
The benefits of developing high value-added specialty chemicals have been demonstrated in many
cases, for example Japanese Sumitomo and Japan Synthetic Rubber (JSR). Within just 5 years, from
1998 to 2002, JSR increased its operating profit by 40 times on an increase of just 1.1 times in turnover.
This achievement is attributed to its diverse strategies of developing high value-added specialty
chemicals, particularly chemical materials, which fostered emerging technologies and subsidiaries,
resulting in nearly 70% of JSR’s profit coming from its specialty chemicals business, with accounted for
just 36% of sales [23]. While the role of government intervention in East Asian latecomer countries is
essential to developing national innovative capacity, the following empirical study attempts to provide a
clear picture for the interdependence and co-innovation of the chemical industry, and the innovative
incentives of industrial development between Taiwanese chemical industry and other strategic industries.

3. Technological interdependence between Taiwan chemical industry and other strategic industries

This section employs an empirical investigation based on patent data to examine the technological
interdependence between the chemical industry and other strategic industries in Taiwan. The reasons why this
study chose patent data to examine technological interdependence are included the following. First, the
information contained in patents is more accessible than other forms of data, is not subject to the problems of
imprecise definition and lack of comparability between firms associated with R&D data, and can be allocated
directly to more detailed analysis fields, for example technology, products or inventors. Second, patents are an
objective measure of technological output, since patents will be examined and eventually granted by the patent
office. A large amount of innovation information are contained in patents, they are uniformly classified, which
eases the analysis of specific technology aspects. Finally, in comparison to other sources, patents are often the
only source for the timely recognition of technological changes, which gains importance in the light of
increasing global competition [24].
Grupp and Schmooh’s study reviewed patents in the USA (United States Patent and Trademark Office;
USTPO) and Europe (Europe Patent Office; EPO), and Japan (Japanese Patent Office; JPO) for reliable
indicators for use in international comparative analysis, and postulated the idea of the dtriad patentT[25]. This
study finally selects US patents as the most reliable proxy of innovation quality, owing to USPTO having the
following advantages. First, USPTO includes statistics filtered by the US Patent Office during the registration
process, which is currently becoming a global standard. Second, US patents have been found to be an adequate
indicator of the economic importance of patents. Finally, USPTO has a unique feature which captures the
linkage between an invention and the most closely related prior knowledge. Patents assigned to more patent
subclass were more likely to be cited in subsequent patent documents and litigated for infringements [26].
One way to capture technological interdependence among different industries is to classify patents into
different technological clusters based on the technological classifications of patents. Patent classifications are
determined via a careful process. Widely used patent classification schemes include International Patent
Classification (IPC) and U.S. Patent Classification (USPC). According to a review of the two classification
systems by Lerner [26] and Fung and Chow [27], this investigation counts the number of USPC classes to which
patent examiners assign each patent. The USPTO has developed a highly elaborate system for classifying the
M.-C. Hu, C.-Y. Tseng / Technological Forecasting & Social Change 74 (2007) 298–312 303

Patent granted the USPTO in Electronics, Mechanical and

Chemical Industry, 1979-2004
Patent count

Mechanical Chemical Electronics

Fig. 2. Source: [29].

technologies to which patented inventions belong, comprising approximately 400 main (3-digit) patent classes,
and over 120,000 subclasses. Hall et al. [28] developed a higher-level classification, by which the 400 classes are
aggregated into 36 two-digit technological sub-categories, and these in turn are further aggregated into six main
categories: Chemical (excluding Drugs); Computers and Communications; Drugs and Medical; Electrical and
Electronics; Mechanical; and Others. However, this study simplifies these six categories into three distinct
industries3: Chemical (including Drugs and Medical), Electrical (including Computers and Communi-
cation) and Mechanical. Because the patent examiners assign patents to one or more U.S. patent
subclasses, a patent may simultaneously belong to two different industries, and the relationship of
technological interdependence amongst the three industries can be examined.
Fig. 2 shows that the total number of patents granted had a low level of international patents production and
innovation in the earlier years but increased markedly over time, particularly after the mid 1990s, peaking in the
final year of the sample period. From the view of patenting history in the USPTO, Taiwanese chemical patents
exhibit a highly correlation with the patenting trend in the electronics and mechanical industries with correlation
coefficients of 99.90% and 99.30%, implying chemical patents are typically associated with specialized
technological areas in Taiwan.
A similar investigation was conducted by Fung and Chow in which they constructed the technological
overlap between firms [27]. To clarify the relationship of technological interdependence between firms, this
study constructs the specific measure based on modifying the technological overlap developed by Fung &
Chow, namely the degree of interdependence (DOI) between industries, which is defined as follows:
DOIk;j;t ¼
where DOIk,j,t denotes the degree of technological overlap between two industries, restated, it is measured by the
number of patents simultaneously granted to industries, k and j at time t; while PTj,t is the number of patents
granted to industry j at time t.
The USPC and IPC (International Patent Classification) are different in the philosophies of dealing with a high number of
parallel entries. One difference between USPC and IPC is that there are seldom a high number of parallel entries in the IPC,
whereas the USPC often contains several dozens of parallel entries at the same level of indentation and results in the mechanical
sector that has been leading the other two sectors. However, this effect is minimized in the chemicals sector [29].
304 M.-C. Hu, C.-Y. Tseng / Technological Forecasting & Social Change 74 (2007) 298–312

Degree of interdependence - Chemical vs. Electronics and











Degree of interdependence between Electronics and Chemical

Degree of interdependence between Mechanical and Chemical

Fig. 3. Source: USPTO and authors calculation.

As some latecomer countries focus on so-called dstrategic industriesT, the current strategic industries
or technologies of Taiwan, such as semiconductors, TFT-LCD and nanotechnology, are expected to
become primary downstream industries and applications for the chemical industry. While the
relationship of knowledge embodied between the chemical industry and commercial commodities has
been proved in the industrial development of Germany [30,31], the inter-industry interdependencies
between the chemical industry and strategic industries based on input–output would be the main source
of knowledge accumulation and diffusion. In this respect, the downstream industry determines the R&D
directions executed by chemicals producers. Conversely, the strength of process innovations in the
chemical industry affects not only the incentive to invest in research, but also the efficiency with which
new knowledge is transferred to the downstream industry.
As shown in Fig. 3, the course of the industrial development of Taiwan over the past 25 years
supports the above argument, in which our empirical results by examining the technological
interdependence through US patent classifications amongst Taiwan’s three major industries indicate a
close relationship among the chemical, mechanical and electronics industries. Remarkably, the degree of
interdependence between the chemical and mechanical industries has remained above 50% during the
last 25 years, and nearly reached 80% during the late 1980s, corresponding to Taiwan’s brilliant
downstream production performance in mechanical industry in the early 1990s (see Fig. 3).4
Fig. 4 illustrates that during the major changes of the past half century, the chemical industry has
continued to fulfill a supporting role, and has evolved together with the technological changes in food
and textile products during the 1980s, machinery in the early 1990s, and electronics in the mid-1990s.
This reflects the fact demonstrated in Fig. 3 that the degree of interdependence between the chemical and
electronics industries has fluctuated while rapidly increasing overall since the mid-1990s and the degree
of interdependence between the chemical and electronics broke through 50% for the first time in 1997, to

Please note that there is a laggard effect between patent granting and downstream production performance.
M.-C. Hu, C.-Y. Tseng / Technological Forecasting & Social Change 74 (2007) 298–312 305

Electronics the Engine of Manufacturing in Taiwan

% of Production Value
35 LCD

30 NB
15 IC

10 PC

Metal machinery Chemical industry
Electronic information Food, textile & other products

Fig. 4. Source: Industrial Development Bureau, Ministry of Economic Affairs, Taiwan. Note: PC refers to Personal Computer;
IC refers to Integrated Circuit; NB refers to Notebook or laptop; and LCD refers to Liquid Crystal Display.

overtake that of chemical and mechanical industries. Again, this pattern corresponds to the emphasis on
the development of strategic industries and demonstrates the essential role of the chemical industry in
evolving together with industrial growth.
Under this circumstance, when the Taiwanese electronics industry (e.g. the semiconductor and FPD
industry) become one of the primary strategic industries in the 1990s and nanomaterials by design
became one of the primary technologies for national development in the 2000s, chemical material and
processing application become indispensable. Thus finding a way to stimulate new technology-based
industries, such as the chemical processing/engineering industry, has been a key challenge for both
industrialized and industrializing countries.
The following section utilizes a case study approach to examine the relationship of technological
interdependence and networking between the development of the chemical and electrical industries, and
the role of the chemical industry in facilitating knowledge diffusion in the industrial development of

4. Case study: building national innovative capacity through knowledge diffusion

When Taiwan’s FPD (Flat Panel Display) industry is trying to establish its new industry lines, namely
TFT-LCDs, following the successful development of PC and semiconductor industries, chemical process
innovation provides a great chance to maintain and extend its innovative capability.5 As a fast follower
initially focusing on Japanese large-panel TFT-LCD mass production and new commodities, one of the
competitive advantages of the Taiwanese TFT-LCD industry (like its Japanese and Korean competitors)
is competing in value chain restructuring — either reducing the cost of production, particularly material
and equipment costs, or moving the most easily detachable activities to lower-cost countries (such as
China). Besides the cost competition, value-added creation in the process innovation has become an

The definition of innovation in the latecomer country uses the term dnew-to-the-countryT, which is different from the term
dnew-to-the-worldT used in advanced countries.
306 M.-C. Hu, C.-Y. Tseng / Technological Forecasting & Social Change 74 (2007) 298–312

alternative to maintaining sustainable competitiveness, whereas chemical process innovation is critical

and indispensable.
Since chemical process innovation lines with the specialty of bembedded-inQ in most emerging
industries (such as in the opto-electronics) and technology (such as in the nanotechnology), R&D
problem-solving can occur before a new product or process design is introduced to the factory via
computer simulations, laboratory experiments, prototype testing, pilot production and other experiments.
With much of the literature focusing on blearning-by-doingQ6 in relation to manufacturing improvement
and technological innovation [32], the specialty of bembedded-inQ in the chemical industry creates the
need for blearning-before-doingQ [6], which leads to significant opportunities for innovation, especially
for process innovation in emerging industries or technologies.
The need for dlearning-before-doingT is especially important in the emerging technologies of
latecomer countries such as Taiwan, in order to build national innovative capacity through diffusing
the knowledge into industries as soon as possible. The technology leverage associates with
knowledge diffusion of Taiwan is one of the critical drivers in achieving successful industrial
development during past decades, which can be effected via the establishment of strategic
institutions and combinations of institutions and firms. While the strategy is to be generalized as an
experience and can be expected to apply again and again, Taiwan has attempted to create high-tech
industries through applying similar leverage strategies to those that are effective in semiconductors
to biotechnology and aerospace, but with little success to date. However, there are limits to the
success of leverage strategies. For example, even in the IT and electronics sectors, Taiwan has had
patchy success. While Taiwanese firms ride high on PC and PC components such as motherboard
and scanners, they have made little headway in such ultra sophisticated products as hard disc
drives, and until the late 1990s they were also making little headway in CD-ROMs and flat panel
displays. What is required in these products is not simply digital electronics and video technology,
but also capabilities in precision and chemical engineering [33]. In the absence of a strong
infrastructure of precision and chemical engineering firms, which themselves may have been
leveraged from supplying contracts to incumbent players, it has proven extremely difficult for
Taiwan to leverage and implement the technologies associated with these precision and chemical-
drive-based products.
dLearning-by-doingT on the learning curve involves starting from the introduction of a new product
or process into the problem-solving stage of the R&D process. In this situation, the chemical industry
may play a different role from the traditional source inputs and to reverse the route of dlearning-by-
doingT in the latecomer country. Distinct from the development in the advanced country, the
development of an industry in the latecomer country such as Taiwan is well known emerged in the
growth stage of a product or process life cycle when leading international players in advanced countries
start to outsource manufacturing capacity for their mature products. Through building mass production
capability to drive down the price in the global market, technology knowledge diffusion in the
latecomer country presents a unique pattern in the industry value chain. Not until the mature product or
manufacturing process are brought into the latecomer countries do other related components and
materials industries (such as chemical engineering) get the opportunity to involve the problem-solving
R&D process and production. In other words, the industrial value chain in latecomer countries is

Learning-by-doing occurs after a process is transferred to a commercial production environment.
M.-C. Hu, C.-Y. Tseng / Technological Forecasting & Social Change 74 (2007) 298–312 307

dbottom-upT, being based on mature product capability and extending backwards to build supportive
However, the difficulty faced by this dbottom-upT model of industrial development in Taiwan is that
the supporting industries must act as a blearning-by-doingQ partnership since the product emerged in
Taiwan (namely new-to-the-country). In this situation, transaction cost and risk are very high for the
development of full-stream industrial value chains in latecomer countries. In the TFT-LCD industry, the
technological origin transferred from Japan in 1997, and thus its key component industries (in Japan)
control specifications and have abundant blearning-by-doingTT experience. Therefore some key
components, such as color filter, resistor and developer, which contain substantial chemical materials,
are mostly imported from Japan or, if not imported, are produced by Japanese ventures in Taiwan. To
achieve the goal of mass production as soon as possible, this established industrial relationship must be
maintained to reduce the probability of failure, where local Taiwanese firms have found no role for
This established industrial network derived from the technology suppliers indicates the difficulty
faced by technology recipients in catching up with or fast-following the advanced countries due to
high risk and transaction costs. The existing industrial cluster in TFT-LCDs (linked by Japanese
companies) create higher entry barriers for local Taiwanese companies wishing to enter the value
chain as suppliers, and this difficulty remains difficult to overcome until new next generation
technologies emerge. However, the next generation of technology (for example from 3.5G to 4G or
4G to 5G) provides an entry opportunity for ready-to-go suppliers. The more knowledge accumulated,
the greater the chance that a new industrial cluster may be established via the blearning-by-doingQ
effect in developing the new process technology of the new generation. But the dembedded-inT
capability of the key components must be well-prepared whenever a new generation of TFT-LCDs has
been introduced. In such situations, latecomer countries must ensure its supportive sectors in the
strategic industry which have accumulated enough knowledge and capabilities to take advantage of the
opportunities when they arrive.
Based on technological evolution, innovation capability is built through conversion between tacit
and explicit knowledge [34]. In interacting around a new product or process, innovation would be
derived from the knowledge about how best to integrate the product or process it during the blearning-
before-doingQ activity. While chemicals and chemical processing are essential for the color filter in the
process innovation (e.g. resistor and coating technology) and takes the most essential 28%–32%
materials cost in TFT-LCDs, the development of Taiwan’s color filter sector thus presents how
Taiwan’s chemical industry helps build national innovative capacity through knowledge internalization
and diffusion.
When TFT-LCDs were introduced to Taiwan during their 3.5G in 1997, the entire range of key
components and materials, such as color filter had not been developed in Taiwan. The exception is
Taiwan’s second largest TFT-LCD player, CMO (Chi-Mei Optelectronics), who is originated from
chemical industry with the world No. 1 ABS production. Resulting from knowledge accumulation and
diffusion from its chemical sector, CMO is Taiwan’s only one color filter producer before 2000 until the
second color filter company, Sintek, was established in 1999 by means of technology transfer from
Japanese DNP and started mass production in 2000. The third Taiwanese color filter producer, AMTC
(Allied Material Technology Corporation) was set up in 2000 with technology transferred from Japanese
Toppan Printing and started mass production in 2001 (see Table 1). The effective knowledge diffusion
through chemical industry associates with the need of value chain integration in TFT-LCD industry
308 M.-C. Hu, C.-Y. Tseng / Technological Forecasting & Social Change 74 (2007) 298–312

Table 1
Color filter production capacity for Taiwan’s makers
1999 2000 2001 2002 2003 2004 2005
CMO 45 153 314 684 1036 2184 4250
Sintek – 70 375 818 1367 1994 2373
AMTC – – 22 336 571 1063 1300
Cando – – 14 195 293 370 733
AU Optronics – – – – 64 1094 3673
HannStar – – – – – 281 983
CPT – – – – – – 836
Innolux – – – – – – 408
Quanta Display Mass production in the first quarter, 2006
Toppoly Mass production in the first quarter, 2006
Total 45 223 725 2033 3331 6986 14556
Self-sufficiency rate 5% 10% 27% 55% 69% 81% 88%
Unit: 000 m2.
Source: DisplaySearch [35].

enable Taiwan’s top five TFT-LCD players (i.e. AU Optronics, CMO, CPT, HannStar and Quanta
Display) to build their in-house color filter production lines, which greatly increase Taiwan’s innovative
capacity in the global TFT-LCD industry.
Table 1 also demonstrates how the Taiwan’s color filter sector internalized and diffused external
knowledge (mostly from Japan) to build endogenous innovative capacity in the TFT-LCD industry,
whereas the self-sufficiency rate significantly jumped up from merely 4.9% in 1999 to 88% in 2005.
Associated with the growth of the global TFT-LCD industry, the effect of knowledge diffusion initiated
from the chemical industry has rapidly made Taiwan surpassing Korea and Japan to become the world’s
No. 1 color filter producer since the first quarter in 2006 [35].
This achievement indicates that the key incentive to building a full-stream industry somehow depends
on government industrial policy, in which the mechanism of blearning-before-doingQ is essentially
required in order to launch the product on the market in time. Thus determining how public policy
influences the launch and acceleration of the development of new technology-based industries through
knowledge diffusion to reinforce national innovative capacity has become a major challenge.

5. Discussions: linking emerging industries and national innovative capacity

Many studies have shown that the chemical industry benefits long-term economic health, and the
chemical industry has remained the largest manufacturing industry in the EU and the second trade-
surplus industry in the US. The emphasis of previous research for Taiwan’s industrial development has
been on manufacturing, and on efforts to constantly upgrade and internalize its technological capacities
from external sources. Fig. 4 illustrates how in Taiwan these efforts led to the industrialization of
agriculture, and a reduction in its role as a major employer, while manufacturing increased its
contribution to total added value to a maximum of 35% in the mid-1980s.
Fig. 4 above reveals how the emphasis on electronics shifted from consumer electronics in the 1970s
to PCs and IT products in the early 1980s, to chips and integrated circuits in the 1990s, reaching the top
in recent products like notebook PCs and LCD screens. Among these changes, the output of the
M.-C. Hu, C.-Y. Tseng / Technological Forecasting & Social Change 74 (2007) 298–312 309

Table 2
R&D productivity and its related, Taiwan’s three major industries
Industry Chemicals Mechanical Electronics
Number of firms (1) 31,341 66,689 16,277
Employees (thousand) (2) 506 735 755
Production (NT $ billion) (3) 2078 2000 2961
R&D manpower (4) 3721 8957 22,893
R&D expenditure (NT $billion) (5) 10.8 46.8 62.4
R&D manpower productivity (NT $ million) (3)/(4) 560 220 130
R&D $ Productivity (3)/(5) 192.4 42.7 47.5
Average Productivity/per firm (3)/(1) 66,333 29,990 181,925
Average Productivity/per employee (3)/(2) 4106 2720 3923
Source: ITRI [23].

chemical industry has remained constant despite it being highly co-evolved with the high-tech industries,
demonstrating that the factor inputs (such as R&D manpower and $) in the chemical industry are as little
as around 30% of electronic and 40% of mechanical inputs, but the productivity of the R&D inputs of
the industry exceeds four times electronics and around 3 to 4 times mechanical outputs. (see Table 2).
Table 2 also indicates that the Taiwanese chemical industry has the highest average productivity per
employee comparing to the electronics and mechanical industries. However, its average productivity per
firm is just 36% of that of the electrical industry. This finding reflects the trend in the global chemical
industry towards consolidation and mergers and acquisition, as firms reduce costs and focus their
businesses, which may be the direction followed by the Taiwanese chemical industry.
While chemicals are the building blocks of numerous products that meet fundamental human needs
for food, shelter, and health, and being vital to such advanced technologies as computing,
telecommunications and biotechnology, its contribution toward national innovative capacity in East
Asian latecomer countries is measured and particularly demonstrated in the finding by Hu and Mathews
[12]. When some latecomer countries such as Taiwan and South Korea reach the stage where they can
easily develop specialty chemicals, as discussed previously, their ability to dlearning-by-doingT is
expected to accelerate the development of emerging industries on process innovation in the future. The
development of process innovation plays an integral role in the strategic industries of Taiwan, but few
studies have examined how the efficiency of process development can be reinforced. The results of this
study suggest that the chemical industry can act as an effective problem-solving intermediator and
catalyst for Taiwan’s emerging industries or technologies, particularly in relation to public R&D

6. Concluding remarks

Previous studies have acknowledged the multi-faceted nature of industrial development in latecomer
countries that focus on scientific and technical activities aimed at innovation, particularly R&D. Through
the building of mass production capability in latecomer countries to drive down global market prices,
Taiwan’s technology knowledge diffusion in the value chain of an industry presents a unique pattern
unlike that of the advanced countries. However, this study took a single industry as the unit of analysis,
and further investigated its interactions with the industry as a whole, by using an empirical study based
310 M.-C. Hu, C.-Y. Tseng / Technological Forecasting & Social Change 74 (2007) 298–312

on patent data, to examine the role of the chemical industry and its technological interdependence in the
building of national innovative capacity in Taiwan — a latecomer country.
The development of process innovation is integral in the strategic industries of Taiwan, but few
studies have addressed how process development efficiency can be reinforced. However, history has
demonstrated the importance of sustainable development in the chemical industry, which plays an
indispensable role in building the national innovative capacity for the advanced countries. The chemical
industry is not only highly interdependent with Taiwan’s economic development, but is also one of the
most influential industries in terms of diffusing knowledge into other strategic industries. In this respect,
how the knowledge of the chemical industry evolved and transferred into emerging industries and
technologies is critical for latecomer or fast-follower countries like Taiwan.
Research has proved that high value-added specialty chemicals are the primary profit resource for
individual companies and also for the entire industry as well as helping national competitiveness, whereas
latecomer countries like Taiwan and South Korea may play a significant role in the area of specialty
chemicals because of the prosperous IT industry. The chemical industry thus enables the accelerated
development for emerging industries such as TFT-LCD and nanomaterials in process innovation, which is
also a major factor inspiring the numerous innovations in interdependent nanotechnology.
National innovative capacity in latecomer countries is limited and is dependent on previous
knowledge stock, R&D manpower, private R&D expenditure and specialization in the high-tech industry
along with public R&D funding. Particularly, the role of public R&D expenditure is significant, and
changes over time from that of an innovative infrastructure provider to a function of linkage between
innovation infrastructure and environment. While public R&D expenditure is found to significantly
enhance the contribution of specialization of patenting in three key industrial sectors (e.g. chemical,
electrical and mechanical), which are the drivers of the innovation capabilities and knowledge diffusion
of Taiwan, the chemical industry is providing numerous possibilities for process innovation. For the
chemical industry, the most significant scientific and technological challenges are realizing the enormous
economic and societal benefits from various applications. The significant influence of public policy in
launching and accelerating the development of new technology-based industries thus reinforce a nation’s
innovation capacity.
The relevance and effectiveness of patents in the chemical industry has been demonstrated in many
studies. This study demonstrates the significant and useful role of the chemical industry in the
technological interdependence and knowledge diffusion with strategic industries in the latecomer
country such as Taiwan. This study suggests that while the public resources of Taiwan are focused on
accelerating the development of emerging industries and technologies and upgrading the so-called
traditional industries, the chemical industry can act as an effective problem-solving intermediary and
catalyst. However, questions that still need to be answered, including how the involvement of the
chemical industry impacts downstream competitiveness and how does the impact differ across industries
in the latecomer country? In either case, important insights of great practical significance are likely to be


[1] P. Romer, Increasing returns and long run growth, J. Polit. Econ. XCIV (1986) 1002 – 1037.
[2] J.E. Stiglitz, Endogenous growth and cycles, in: NBER Working Paper No. 4286, 1993.
M.-C. Hu, C.-Y. Tseng / Technological Forecasting & Social Change 74 (2007) 298–312 311

[3] R.J. Barro, Economic growth in a cross section of countries, Q. J. Econ. CVI (1991) 407 – 443.
[4] B. Bowonder, Innovation and convergence: expanding boundaries of chemical industry, Interdiscip. Sci. Rev. 26 (1)
[5] R. Landau, Economic growth and the chemical industry, Res. Policy 23 (1994) 583 – 599.
[6] G.P. Pisano, Learning-before-doing in the development of new process technology, Res. Policy 25 (1996) 1097 – 1119.
[7] S. Hung, Y. Chu, Stimulating new industries from emerging technologies: challenges for the public sector, Technovation
(2004) 1 – 7.
[8] D.C. Mowery, N. Rosenberg, Paths of Innovation, Cambridge University Press, 1998.
[9] R.C. Levin, A.K. Klevorick, R.R. Nelson, S.G. Winter, Appropriating the returns from industrial research and
development, Brookings Paper on Economic Activity 3 (1987) 783 – 820.
[10] A. Arora, Patents, licensing, and market structure in the chemical industry, Res. Policy 26 (1997) 391 – 403.
[11] S. Klepper, Entry, exit, and innovation over the product life cycle, Am. Econ. Rev. 86 (3) (1996) 562 – 583.
[12] M.C. Hu, J. Mathews, National innovative capacity in East Asia, Res. Policy 34 (9) (2005) 1322 – 1349.
[13] M.C. Hu, Explaining the drivers of innovation in East Asia. Measuring rates and drivers of innovation: in countries and
companies. Symposium in Melbourne Business School, September, Australia, 2004.
[14] J. Furman, M. Poter, S. Stern, Determinants of national innovative capacity, Res. Policy 31 (2002) 899 – 933.
[15] MOEA (Ministry of Economic Affairs). Annual Reports. Taipei, Taiwan, 2002.
[16] Cleanese Chemical Group, Drivers of change in the global environment of the chemical industry, http://www.celanese.
com, 2000.
[17] B.A. Lundvall (Ed.), Innovation as an interactive process: from user–producer interaction to the national system of
innovation. Technical Change and Economic Theory, Printer Publishers, London, UK, 1988.
[18] D. Teece, Managing Intellectual Capital, Oxford University Press, New York, 2000.
[19] B. Kogut, U. Zander, Knowledge of firm, combinative capabilities, and the replication of technology, Organ. Sci. 3 (1992)
383 – 397.
[20] S.G. Winter, Knowledge and competence as strategic assets, in: D.J. Teece (Ed.), The Competitive Challenge, Cambridge,
MA, 1987.
[21] W.W. Chu, Import substitution and export-led growth: a study of Taiwan’s petrochemical industry, World Dev. 22 (5)
(1994) 781 – 794.
[22] A.M. Thayer, Firms find a new field of dreams, Chem. Eng. News 78 (42) (2000) 36 – 38.
[23] ITRI (Industrial Technology Research Institute), Chemical Industry Reports, Union Chemical Laboratory, ITRI, Taiwan,
[24] H. Ernst, Patent applications and subsequent changes of performance: evidence from time-series cross-section analyses on
the firm level, Res. Policy 30 (2001) 143 – 157.
[25] H. Grupp, U. Schmooh, Patent statistics in the age of globalization: new legal procedures, new analytical methods, new
economic interpretation, Res. Policy 28 (1999) 377 – 396.
[26] J. Lerner, The importance of patent scope: an empirical analysis, Rand J. Econ. 25 (2) (1994) 319 – 333.
[27] M.K. Fung, W.W. Chow, Identification of technological structures using patent statistics, Econ. Innov. New Technol. 12
(4) (2003) 293 – 313.
[28] B.H. Hall, A. Jaffe, M. Trajtenberg, The NBER patent citation data file : lessons, insight, methodological tools. UC
Berkeley, Brandeis University, Tel Aviv University, and NBER, 2001.
[29] J. Streb, Shaping the national system of inter-industry knowledge exchange: vertical integration, licensing and repeated
knowledge transfer in the German plastics industry, Res. Policy 32 (2002) 1125 – 1140.
[30] S. Subramoney, Carbon nanotubes, Interface 8 (4) (1999) 34 – 37.
[31] P. Adler, K. Clark, Behind the learning curve, Manage. Sci. 37 (3) (1991) 267 – 281.
[32] E. Von Hippel, M. Tyre, How the dlearning by doingT is done: problem identification in novel process equipment, Res.
Policy 24 (1) (1995) 1 – 12.
[33] J.A. Mathews, Competitive advantages of the latecomer firm: a resource-based account of industrial catch-up strategies,
Asian Pac. J. Manage. 19 (4) (2002).
[34] E.M. Rogers, Diffusion of Innovation, 4th ed., Free Press, New York, 1995.
[35] DisplaySearch. Quarterly TFT color filter component report. Austin, USA, 2006.
312 M.-C. Hu, C.-Y. Tseng / Technological Forecasting & Social Change 74 (2007) 298–312

Mei-Chih Hu is currently working as an Associate Professor at the Graduate Institute of Management of Technology, Feng
Chia University, Taiwan. She holds a PhD in Management from Macquarie Graduate School of Management, Australia. Her
papers have been published in journals like Research Policy and a variety of conferences like DRUID (Danish Research Unit
for Industrial Dynamics), Denmark.

Chun-Yao Tseng is an Associate Professor at the department of Business Administration, Tunghai University, Taiwan. He
holds a PhD in Business Administration from the Taipei University, Taiwan. His research is in the area of intellectual capital and
technology management. His papers have been published (or accepted) in a variety of journals including R and D Management,
International Journal of Technology Management, Applied Economics Letters, and etc.