THE LIMITS OF BRAIN-CIRCULATION: Chinese returnees and technological development in Beijing

By

CHEN, Yun-chung Assistant Professor

Hong Kong University of Science and Technology ycchen@ust.hk

Revised and Resubmit to Pacific Affairs

Center on China’s Transnational Relation*

Working Paper No. 15

The Hong Kong University of Science and Technology

* CCTR Corporate Sponsor 2006-2007: Ronnie C. Chan; Shui On Holdings Limited

1

Abstract

Saxenian’s theory of “brain-circulation,” which argues for a two way flow of technical communities instead of a one way brain drain or reverse brain drain, has turned from a theory to a popular formula for many policy makers of latecomer regions trying to jump-start their high-tech industries. Since the brain-circulation theory is primarily based on the experience of the Silicon Valley-Hsinchu connection, this article compares the Hsinchu experience with the Zhongguancun (ZGC) experience in Beijing to explain why ZGC is not a case for brain-circulation. By re-contextualizing the often neglected hypotheses of the brain-circulation theory, namely (i) decentralized industrial structure with specialized producer networks;(ii) trust-based inter-firm network that induces learning; (iii) critical financial infrastructure for high-tech startups; and (iv) the role of the state in facilitating technology transfer, this article explains why the effect of reciprocal industrial upgrading is not happening in ZGC.

Through case studies, this article details a typology to understand the real role of Chinese returnees in the technological development of ZGC. Managing transnational innovation networks is a complex challenge that requires a long history of accumulation. Since the history of transnationalism in ZGC is 30 years shorter than Hsinchu, only a few very experienced Chinese returnees in ZGC have the accumulated ability to manage its transnational innovation networks. Moreover, knowledge assets (venture capital, research labs, research grant …etc) in ZGC are mostly institutionalized and controlled by the Chinese state. Only a limited number of Chinese returnees can tap into them. These two barriers have limited the Chinese version of brain-circulation to only a few very experienced Chinese returnees with great political skills. The contribution of other types of Chinese returnees to the technological development of ZGC is limited and indirect. At this moment, they are to be able to create an informal industry-university linkage for exchange of human resources and technological information.

2

Table content

INTRODUCTION ........................................................................................................................................................ 3

A. WHY THE CHINESE RETURNEES?............................................................................................................................ 4 Chinese officials are suspicious of genuine MNC technology transfer ................................................................. 5 National champions are slow in creating proprietary technologies...................................................................... 6

B. CHINESE RETURNEES AS THE NEW VEHICLE OF TECHNOLOGICAL DEVELOPMENT ................................................... 7

I. ZGC VARIES FROM THE BRAIN-CIRCULATION THESIS ........................................................................... 8

A. RE-CONTEXTUALIZING THE BRAIN-CIRCULATION THESIS ....................................................................................... 9 1. Industrial structure: networked vs. non-networked........................................................................................... 9 2. The trust and learning mechanism: collaboration vs. lock-in......................................................................... 10 3. Venture capital for startups: established vs. weak........................................................................................... 13 4. The role of the State: Technology facilitator vs. technology seeker................................................................. 14

B. DIFFERENT PATHS OF TECHNOLOGICAL DEVELOPMENT: COMPARING HSINCHU AND ZGC.................................... 15

II. THE CHINESE RETURNEES AND TECHNOLOGICAL DEVELOPMENT .............................................. 18

A. FOUR TYPES OF CHINESE RETURNEES .................................................................................................................. 18 Type 1: Transnationalism with strong local connections .................................................................................... 19 Type 2: Transnationalism with weak local connections ...................................................................................... 22 Type 3: Localism with strong local connections ................................................................................................. 26 Type 4: Localism with weak local connections ................................................................................................... 28

B. COMPARING THE DIFFERENT TYPES OF CHINESE RETURNEES................................................................................ 28

CONCLUSION ........................................................................................................................................................... 32

3

Introduction

The Beijing Zhongguancun (ZGC1) Science Park authority is adopting the “reverse

brain-drain” policy by aggressively attracting the Chinese returnees to establish their startups in

one of the 23 “Overseas Student Venture Parks” in ZGC. Currently, there are around 6,000

Chinese returnees working in these parks.2

This article questions how far Saxenian’s “brain-circulation”3 thesis applies to the role of

the transnational technical community (or the so called Chinese returnees4 in the PRC context)

in ZGC in terms of technology transfer and industrial upgrading. The brain-circulation

argument is based primarily on the Silicon Valley-Hsinchu (Taiwan) experience. How well

does it explain the ZGC experience? Our ability to understand the limitation of this circulation

1 Zhougguancun in Beijing is often considered the “Silicon Valley” of China, with a high-concentration of over 8000 IT

firms.

2 人民网-人民日报海外版, 2005/05/31

http://edu.sina.com.cn/a/2005-05-31/1756116748.html

3 In contrast to the one way “brain-drain” or “reverse brain-drain” theories, brain-circulation – a term coined by Saxenian –

emphasizes the two way flow of skilled workers between home and host countries.

4 Chinese returnees are commonly refers to Chinese with PRC nationality who goes to the West (mainly the U.S., but also

including Europe and Japan) for higher education. They might or might not have working experience in the West after their

graduation. Increasingly, they also include who was hired directly by the companies in the West from China and return after

a few years. Today, the halo of Chinese returnees has lost its luster because the term has been overloaded with cadres who

claim to be “Chinese returnees” after only a short visit (less than 6 months) to the West. For the purpose of our paper, we

would narrowly defined Chinese returnees as those who have substantial working experience in the West. They usually has

a permanent resident or even citizenship of the Western countries. These Chinese returnees used to dream of working for

the Fortune 500 companies in the US under H1 visa (with the opportunity to obtain a green card in three years). Now, many

of them are heading back to China especially the NASDAQ crash in 2001. Some has return earlier in the 1990s for the rare

‘window of opportunity’ to start their own businesses as they are less technological competitor.

4

thesis will help us understand the real and limited effects of the Chinese returnees in the

technological development5 of Beijing’s Zhongguancun, and perhaps of China in general as

well.

This article argues that due to the different key aspects of the two regions, and also the

difference of these two transnational technological networks, the impact of the ZGC-Silicon

Valley connection is quite different from the Taiwan case. While acknowledging the positive

“brain-gain” in ZGC (with the increase in Chinese-returnee startups for example), the effect of

reciprocal industrial upgrading of the Hsinchu and Silicon Valley case due to brain-circulation

is not yet noticeable in ZGC. This is inconclusive at best.

This article consists of two main parts. Part one examines the crucial difference in the

historical formation of the high-tech regions in Hsinchu and ZGC. This shows the limits of the

brain-circulation thesis in explaining the ZGC experience, despite the seemingly “similar”

transnational technology networks between Taiwan and China. Part two uses case studies to

analyze the potential impact of the Chinese returnees in the technological development of ZGC.

The fieldwork of this research was done during two periods: 2002 June to Jan 2003 and

June 2005. I had interviewed around 100 informants from 40 companies (both foreign and local

firms), government agencies (mostly local), Chinese universities and their technology transfer

arms (like incubators, Technology Transfer Offices, Joint research labs).

A. Why the Chinese returnees?

The policy shift to attract Chinese returnees is the outcome of two frustrations faced by 5 The term industrial upgrading implies a collective phenomenon, which may be too restrictive for ZGC. Therefore, I use

the term technological development, which does not assume a collective phenomenon, admits great unevenness among

individual firms, but still acknowledges that certain types of Chinese returnees do have positive impacts on the

technological accumulation in the region.

5

the Beijing technocrats in trying to speed up the industrial upgrading in ZGC, Beijing.

Chinese officials are suspicious of genuine MNC technology transfer

While the official public policy to attract multinational corporations to Beijing remains

strong, Chinese technocrats in ZGC continue to be skeptical, or at least ambivalent, about the

role of the MNCs in Beijing. They doubt whether the profit-oriented MNCs ever genuinely

transfer technologies.6 This ambivalence is much stronger among the technocrats in ZGC than

in a highly commercial city like Shanghai.7

The ambivalence increases with the increasing presence of multinationals in ZGC,

especially with the arrival of multinational R&D centers since the mid 1990s. While the official

rhetoric is always positive and welcoming, worries about the inability of local firms to compete

with the “wolves” (a local word for multinationals) often emerge in the closed-door meetings I

had with technocrats, journalists, scholars, engineers and even students in ZGC. The climax

was the establishment of Microsoft Research China in 1998, which ignited a collective fear in

ZGC. The common anxiety projected in the mass-media in 1998 showed that the public was

worried about MNCs exploiting the skilled labor of China after they have successfully

exploited the unskilled labor in southern China.

The deep anxiety and mistrust against MNCs has a long history. Since the Opium War, the

local intellectual desire was for China to develop its own proprietary technology and science to

6 During my fieldwork in Beijing, I frequently encounter technocrats asking me if the MNC R&D centers will take the

technology invented in Beijing and commercialize them somewhere else. They think the R&D centers are using the

Chinese cheap skilled labor to develop new technology, and they will commercialize this new technology in their home

country, and not in China, and thus, not contributing to the technological development of ZGC.

7 Information based on interviews with technocrats of both Beijing and Shanghai between June 2002 to Jan 2003.

6

compete with the “colonizing” West.8 In fact, nurturing China’s own national champions in the

area of communication, biochemistry, new materials, software and the internet has been the

contemporary high-tech strategy. This strong “technological nationalism,” as opposed to

“technological globalism,”9 also legitimates the explicit “market for technology” policy that is

commonly used by the Chinese central authorities to “force” (or strongly invite) large MNCs to

transfer more technology to China. Examples such as requesting MNCs to set up R&D centers

in China, to transfer their latest product or latest manufacturing technology to the joint-venture

factory in China are abound. However, the real effect of “market for technology” in technology

transfer is still debatable.

National champions are slow in creating proprietary technologies

In April 2002, a book by an embedded reporter in ZGC described the historical success of

the local pioneers in ZGC since the 1980s.10 It tells the story of locally trained IT heroes from

the Chinese Academy of Science (CAS) and the Beijing universities succeeding in building

their IT empires. These local heroes manage to break through the straight-jacket of the public

research institutes11 to “jump into the sea” and set up their institutional-affiliated startups

8 Cao, C. (2004) China’s Scientific Elites. London: Routledge. Feigenbaum, E (2003) China’s Techno-Warriors: National

Security and Strategic Competition from the Nuclear to the Information Age. California: Stanford University Press.

9 See Ostry, Sylvia and Nelson, Richard R. (1995), Techno-Nationalism and Techno-Globalism. Washington, D.C.:

Brookings Institution

10 Liu, R. (2002), Knowledge Heros 2.0, Chongqing: Chongqing Publication (in Chinese,刘韧 (2002), 知识英雄 2.0 重

庆: 重庆出版社)

11 These public research institutes include the research labs of universities and of the Chinese Academic of Science (CAS).

7

(IAS)12 to provide IT products and services to the local market. Leveno13, Stone, and Founder

are constantly quoted as the heroic IAS run by visionary scientists against all odds. Until today,

these pioneers still inspire IT entrepreneurship all over China.14

Unfortunately, the true national champions are more exceptional cases than a common

phenomenon. For example, 75% of the total revenue among the 200 CAS IAS is still coming

from the most famous CAS spin-off, Leveno. Even for Leveno (and other IAS), its strength is

not proprietary technological innovation. It only excels in distribution and service innovation.

B. Chinese returnees as the new vehicle of technological development

The distrust against MNCs and the frustration with unproductive national champions

prompted ZGC technocrats to look for new agents of industrial upgrading15 in ZGC – the

Chinese returnees.

Learning from the successful story of the Hsinchu Science Park, ZGC technocrats try to

mimic Taiwan’s process of rapid high-tech development.16 The ZGC Science Park (ZSP)

12 The IAS is not a spin-off in the usual sense under capitalism, because of the problem of private ownership in China.

Spin-offs in other capitalist countries have clear ownership, but private ownership was unthinkable in Beijing in the 1980s.

So, the IAS had to be of mixed ownership, and was a formal institution without clear guidelines. The struggle on ownership

between the institution and the IAS founders were common whenever these IAS companies become very successful in the

1990s.

13 It is a local computer company, the most famous among the Chinese Academic of Science (CAS) IAS. It was called

“Legend Computer” before it bought the IBM PC department.

14 When I was interviewing young IT entrepreneurs in Shanghai, they always refer to these ZGC heroes as their inspiration

for starting their own businesses.

15 I use industrial upgrading, instead of technological development, because it is the desire of the Beijing technocrats even

though it might be too idealistic.

16 Hsu, J-Y and Saxenian, A. (2000) The Limits of Guanxi Capitalism: Transnational Collaboration Between Taiwan and

8

authority mimics the Hsinchu model by setting up five ZSP service branches in five major

technopoles in the West. They organize recruitment tours to attract Chinese returnees to invest

in ZSP. In fact, the ZSP authority even goes beyond what the Taiwan Hsinchu Science Park has

done by offering the Chinese returnees tailor-made “overseas student venture parks” with

bundled incentive packages.

Thus, the question is whether the Chinese returnees can repeat the success of Hsinchu as

indicated by Saxenian’s brain-circulation theory. To answer this question, we need to compare

the role of the Chinese returnees to their Taiwanese counterparts in their own regional contexts.

I. ZGC varies from the brain-circulation thesis

Saxenian and Hsu convincingly argue that there is effective technology transfer between

the Silicon Valley and Hsinchu via the (Taiwanese) transnational technical communities.17

Saxenian goes on to argue that “brain circulation” has resulted in the reciprocal industrial

upgrading of the host and the home regions.18

On the surface, Taiwanese and Chinese transnational networks might look quite similar, as

they are all returnees from the West (including Japan), especially from the Silicon Valley.

However, a closer look reveals their major structural differences in their respective regional

economies. In fact, this paper identifies four major prerequisites (or hypotheses) of the

brain-circulation theory, which are often neglected by policy makers and academics because

the USA. Environment and Planning a 32, 1991-2005.

17 Saxenian, L. and Hsu, J-Y (2001) The Silicon Valley-Hsinchu Connection: Technical Communities and Industrial

Upgrading, Industrial and Corporation Change, 10(4):2001: 893-202.

18 Saxenian, A. (2002) Transnational Communities and the Evolution of Global Production Networks: The Cases of

Taiwan, China and India, Industry and Innovation, 9(3): 183-202.

9

they are too eager to adopt the brain-circulation theory as “best-practice” to jump-start their

high-tech industry.

A. Re-contextualizing the brain-circulation thesis

The four major structural factors of the brain-circulation thesis are: (i) the industrial structure

of both home and host regions of the returnees, e.g. networked vs. non-networked; (ii) trust,

learning and entrepreneurship: collaboration vs. lock-in; (iii) the quality of the financial

infrastructure for startups: established vs. weak; (iv) the role of the state as a technology producer

vs. technology facilitator.

Below is the comparison of these main hypotheses behind the brain-circulation argument in contrast with the ZGC reality.

1. Industrial structure: networked vs. non-networked

Hsu (2004) argues that the high-tech industrial development in Taiwan and Silicon Valley

are similar because they belong to the decentralized industrial system with strong division of

labor and innovation among networked firms.19 Through their formal and informal networks,

the Taiwanese transnational technological communities link the two decentralized regions,

Silicon Valley and Hsinchu together by exploiting the regional comparative advantage20. In

other words, the friction of distance in technology transfer is overcome by the transnational

social networks of specialized producers.

However, ZGC does not have a decentralized industrial structure. Moreover, the inter-firm

19 Hsu, J.Y. (2004) The Evolving Institutional Embeddedness of a Late-Industrial District in Taiwan. Tijdschrift Voor

Economische En Sociale Geografie 95, 218-232.

20 Hsu, J.Y. and Saxenian, A. (2000) “The Limits of Guanxi Capitalism: Transnational Collaboration Between Taiwan and

the USA” Environment and Planning A 32, 1991-2005. page 1999

10

networks are very weak.21 Besides a few large companies, Beijing’s ZGC has over eight

thousand relatively small IT firms. Unlike the Silicon Valley and the Taiwanese case, there is

insignificant division of labor among the firms because of the lack of trust inherited from the

legacy of planned economy. 22 (See next section.)

Without a network of specialized suppliers, the start-up cost for new firms is high because

they need to produce almost everything in-house. In addition, new startups find it difficult to

position their goods and services in a niche market because of the lack of resource to do market

research. As a result, “small and complete”23 becomes the convention among ZGC startups.

“Small and complete” means no matter how small the startups, they still have to do everything

themselves. This convention drains their resources and makes it difficult for startups to focus

on specialized technology or product innovation. In short, the failure rate among the startups –

local or Chinese returnees – is very high. Thus, the survival of individual firms is already a

challenge, let alone genuine technology innovation and industrial upgrading.

2. The trust and learning mechanism: collaboration vs. lock-in

According to Hsu and Saxenian (2000), the agglomerated economies in Hsinchu and the

21 Wang, J.C. and Wang, J.X. (1998) “An analysis of new-tech agglomeration in Beijing: a new industrial district in the

making?” Environment & Planning A 30, 681-702. ; Wang J (2002) “Strategies of Local Industrial Clusters. Zhongguo

Gongye Jingji (Chinese Industrial Economics) 3 (in Chinese).

22 The IT startup boom in the mid 1980s was a historical accident in Beijing. On the demand side, there was a high demand

for localized products and services behind the high-trade barrier in the 1980s. On the supply side, there was the overflow of

scientists from public research institutions into the market place, who used to be institutionally bounded. See Chen, Y-C

(2006) Why do multinational corporations locate their advanced R&D centres in Beijing? Journal of Development Studies.

forthcoming

23 Wang, J. (2001) Innovation Spaces: Enterprise Clusters and Regional Development . Beijing: Peking University Press

(in Chinese, 王缉慈, 创新的空间, 北京: 北京大学出版社)

11

Silicon Valley encourage entrepreneurship and learning between the firms on the regional level.

The strong inter-firm spillover effects embedded in this decentralized industrial structure make

the role of the technical community more important than in the case of a centralized industrial

structure like Korea.24.

Taiwan’s tradition of collaborative entrepreneurship, relationship based business and

resource-sharing among SMEs25 was adopted into the high-tech sector. While keeping the

trust-based business relations inherited from the traditional consumer goods sector, the

Taiwanese high-tech sector (located in both the Silicon Valley and the Hsinchu regions)

abandoned the heavy reliance on family ties or primary social relations.26 Giving up family ties

while keeping the trust make it possible for Taiwanese high-tech firms to avoid being locked

into the social relations while continuing to enjoy the collaborative division of labor and

inter-firm learning in the high-tech sector of Hsinchu-Silicon Valley.27

Unfortunately, the existing guanxi-based relation in ZGC is different from the trust-based

social relations built over three decades in Taiwan and the Silicon Valley. In Beijing, and China

in general, it is difficult for the existing vertical guanxi-based relations (in both social and firm

relations) of the origin SOE28-based industrial structure to transit to the horizontal inter-firm 24 Hsu, J.Y. and Saxenian, A. (2000) “The Limits of Guanxi Capitalism: Transnational Collaboration Between Taiwan and

the USA” Environment and Planning A 32, 1991-2005.

25 SMEs stands for small and medium size enterprises.

26 Saxenian, L. and Hsu, J-Y (2001) The Silicon Valley-Hsinchu Connection: Technical Communities and Industrial

Upgrading, Industrial and Corporation Change, 10(4):2001: 893-202. Page 910.

27 Hsu, J.Y. and Saxenian, A. (2000) “The Limits of Guanxi Capitalism: Transnational Collaboration Between Taiwan and

the USA” Environment and Planning A 32, 1991-2005.

28 SOE stands for state-owned enterprises. The SOE-based industrial structure is vertical because all the instructions come

from the government directly. This is a legacy of the planned economy. The horizontal relations between SOEs are not

12

trust-based relations. In a similar fashion, the existing top-down guanxi between the Chinese

state agencies and their IAS29 also hinder the IAS from building horizontal inter-firm relations

after they went private.

Moreover, the limits of the private capital market (including venture capital) forces the

IAS to rely heavily on the state agencies for all kinds of resources (such as start-up fund, office

space, experienced employees), approval documents, and even orders. In fact, the main orders

of the IAS normally come from their original state agency through politicking with existing

guanxi among stage agencies.

This dependency of IAS on vertical guanxi has two negative lock-in effects. Firstly, when

the IAS really take off and make profit, they have a hard time breaking free from the state

agencies that used to support them. Secondly, the dependency of IAS on vertical guanxi

continues to exacerbate the distrust among IAS, and thus hinder the formation of horizontal

trust-based relations in ZGC.

Without the horizontal and market-based trust, as in the Taiwanese case, strategic alliances

and joint R&D among firms in ZGC is not common. Every firm (including IAS) holds back

critical technological secrets and special institutional guanxi. Thus, these isolated firms are

locked-into their own innovation without full knowledge of what others are doing. As a result,

“low-level repetitive innovation” and “making tanks behind closed doors (閉門造車)” hinder

the reciprocal learning that is prominent in Hsinchu and Silicon Valley.

based on trust, but on orders from above.

29 The state agencies still own the “spin-off” in the early days of the 1980s, because private ownership was not acceptable

both morally and institutionally among the scientists and scholars in ZGC.

13

3. Venture capital for startups: established vs. weak

Among all the specialized services that are crucial to the IT entrepreneurial economy, the

venture capital mechanism is always the top priority. Venture capital is essential to

technologically risky startups with super-profit compensation when it does IPO or is acquired

by other companies.

The successful transfer and implementation of venture capital (VC) – Silicon Valley

style – in Taiwan is often credited for the successful high-tech development in the

Hsinchu-Taipei region.30 The availability of transnational VC is in fact the key strength of

Taipei in the comparative advantage of these two regions. Moreover, Taiwanese venture

capitalists do not merely provide hard capital. They often coach the IT startup firms in making

realistic business plans, and introduce useful third parties into the collaborations. Transnational

venture capitalists become the source of knowledge and learning platform critical in the

commercialization of technology for many Taiwanese startups (in addition to the transnational

engineer networks).

While the availability and quality of venture capitalists is the strength of the

Hsinchu-Taipei region, the lack thereof is ZGC’s weakness. The underdevelopment of VC is

always blamed on the delay of government approval of the technological stock exchange for

startups. (It is also called the “second exchange board”, which is similar to the Nasdaq).

Without a Chinese local technological stock exchange board, venture capital cannot function

effectively because it cannot exit easily.

Moreover, after the NASDAQ crash in 2001, the limited VC in ZGC – mostly owned by

state agencies – turn conservative. They refuse to invest in the early stage of technological 30 Saxenian, A. and Li, C-Y (2003) “Bay-to-bay strategic alliances: The network linkages between Taiwan and the US

venture capital industries.” International Journal of Technology Management 25:136.

14

development of any startups. They tend to invest only at a later stage in order to reduce the risk

of failure. This conservative turn in the VC policy further limits the risk-taking innovative

activities of the startups in ZGC.

Without this crucial financial infrastructure, even if startups are determined to focus on

technology development (in a Silicon Valley style), they are required to maintain positive cash

flows simultaneously. Thus, the need to trade third party products, provide technical services or

low-end subcontracted manufacturing and do other money making activities drain away their

attention on their own core product/technology development.

4. The role of the State: Technology facilitator vs. technology seeker

The Taiwanese ITRI31 experience has become the high-tech development model for

latecomer countries. The significance of the ITRI lies not in the entity itself, but in its ability to

reflect the changing role of the Taiwan developmentalist state.32

After the successful spin-off of the semiconductor manufacturing companies TSMC and

UMC from the ITRI, the rapidly building up Taiwan IT private sector becomes critical of

government R&D spending. In the mid 1990s, private companies collectively asked the

government to stop the wasteful research funding lavished on the public research labs (such as

ITRI) to fund the private sector instead. Hsu (2004) argues that if the Taiwanese government

agrees to the collective demand and move away from the role of the sole technology provider

to that of the facilitator in technology transfer, then Taiwan would depart from the

31 The Industrial Technology Research Institute (ITRI) is located in the Hsinchu Science Park, Taiwan. It is often credited

for its role in jump starting the Taiwan high-tech industry.

32 Hsu, J.Y. (2004) “The Evolving Institutional Embeddedness of a Late-Industrial District in Taiwan”. Tijdschrift Voor

Economische En Sociale Geografie 95, 218-232

15

developmentalist state hypothesis.33

The role of the Chinese state in the national innovation systems (NIS) does not change

much and thus depends heavily on central funding for mission-oriented basic research of

targeted industries. The Chinese state is acting like a technology seeker that use all means to

secure technologies. (i) protect and promote local firms that have Chinese IP standard ability;

(ii) use “market for technology” to force MNCs to transfer more technology. (iii) continues

mission-oriented R&D funding practice in targeted industries.

However, CAS, which is located mainly in ZGC, is by and largely a main public research

institute based on basic research. It is unlikely to transform into the ITRI model in Taiwan,

which focuses on process and production innovation.

On the other hand, the local governments, such as Beijing and Shanghai, in their eagerness

to build incubators, technological brokers (brokers between capital and technology), venture

capital, productivity centers, technology transfer services, trade conventions, are merely setting

up indirect and passive “soft-infrastructures” that do not necessarily create an environment for

innovation or international technology transfer.

B. Different paths of technological development: Comparing Hsinchu and ZGC

Table 1 below summarizes the major structural differences between Hsinchu in Taiwan

and ZGC in Beijing. In short, the major differences between Hsinchu and ZGC show the

significant limitation of the “brain-circulation” thesis in explaining ZGC. It cannot be applied

to ZGC directly. The exploitation of regional comparative advantage and reciprocal industrial

upgrading through the mediation of transnational technological networks that are prominent in

33 Amsden, A.H. (1989) Asia's Next Giant: South Korea and Late Industralisation. New York: Oxford University Press.

Evans, P. (1995), Embedded Autonomy: States and Industrial Transformation. Princeton: Princeton University Press.

16

Taiwan is not significant in ZGC.

17

Table 1: The two different technological paths: the Hsinchu vs. the ZGC

Taiwan Hsinchu China ZGC Networks

Decentralized networks of specialized producers. (i) Spin-offs of ITRI; (ii) local OEM-ODM manufacturing networks; (iii) firms by Taiwanesereturnees;

Decentralized but fragmented, weak networks that consist of different types of non-manufacturing firms: (i) Institutional-affiliated startups (IAS); (ii) local startups; (iii) MNCs (iv) Chinese returnee startups.

Ownership structure

Private ownership. All firms have clear private ownership.

Mixed ownership. A mixture of wholly-owned/joint-venture MNCs, institutional-affiliated startups (IAS), and small privately owned firms.

Indu

stria

l Stru

ctur

e

Trust and learning

Horizontal Trust-based. Collective learning through participating in the global production networks.

Vertical guanxi-based learning is limited without market-based trust.

Connection with Silicon Valley

Strong “transnationalism,” continuous industrial upgrading. History matters. Taiwanese overseas students have a much longer history (since 1970s).

Weak “transnationalism”, most western trained engineers simply “return home” and not travel between Silicon Valley and home. Chinese overseas students have a shorter history: since the late 1980s, after the Tiananmen Square incident.

Embedded- ness

Taiwanese transnational entrepreneurs are highly embedded in both regions through alumni, professional and industrial networks. Thus, it is very easy for new startups to quickly identify promising new market opportunities.

Weakly embedded in both regions. Due to relative isolation, Chinese startups often do not know how to evaluate their own technological advantage, thus resulting in redundant and repetitive innovation and poorly defined products.

Tran

snat

iona

l tec

hnol

ogic

al n

etw

orks

Relation with Silicon Valley

The Comparative advantage model: Finance and manufacturing strength in Taiwan + engineering and technical skill in the Silicon Valley. Weakness: No domestic market to test their product before going global. Depend on the final product designed and branded by leading multinationals in the U.S.; but this has improved as more OBM firms are showing up. Strength: taking full advantage of the comparative advantage of the two regions.

The Dependency model: ZGC depends on technology transfer from Silicon Valley + depends on the state financially + depends on Taiwanese and Korean firms for manufacturing capabilities. Weakness: highly dependent and risky. Focus mainly on “localizing the product” and compete on cost and not on technology. Large firms rely too much on the strength of the domestic distribution channel, and less on technology. This advantage will soon evaporate with the opening up of distribution channels by MNCs. Strength: focus on own IP or own brand name; have a domestic market to test new products; large firms dominate the distribution in non-metropolitan markets.

Role of state

From technology provider to technology transfer facilitator.

From technology provider to technology seekers that use all means to secure technologies. (i) protect and promote local firms that have Chinese IP standard ability; (ii) use “market for technology” to force MNCs to transfer more technology. (iii) continue mission-oriented R&D funding practice in targeted industries.

Infra

stru

ctur

e fo

r st

artu

ps

Venture capital

Venture capital is very established in the Taiwan High-tech industry.

Venture capital is almost non-existent, partly because of the lack of exit channels in China. Funding source mainly from family/friends; limited state-sponsored venture capital (VC); limited MNCs and local VC.

Firm strategy

Specialized producers targeting global nichemarkets. Bargain with the state to provide more R&D findings, supply more skilled labor, & less restriction on investment in China.

Import substitution strategy to increase domestic market share. Position as “national enterprise” to get attention of limited venture capital and thus funding from state agencies. Keep arms-length relations with state agencies.

Technology focus

Production innovation: OEM manufacturing technology; IC design is beginning to break new ground in product innovation.

Product innovation: focuses on cheaper products to replace the imported products that dominate the existing market. Weak in production technology, but can subcontract to Taiwanese and Korean OEM firms.

Tech

nolo

gy d

evel

opm

ent

Technology upgrading

Upgrading along the commodity chain starting from production technology. e.g. from OEM to ODM to OIM or OBM; Highly sensitive to marketable technologies.

Product innovation and localization strategy. Innovation is often highly repetitive, not sensitive to marketable technology and products.

Source: by author.

18

If the brain-circulation thesis cannot fully explain the industrial upgrading of ZGC, then

what are the actual roles of the Chinese returnees, who are much weaker in transnationalization

as compared to their Taiwanese counterparts? I will answer this question in part 2 of this paper.

II. The Chinese returnees and technological development

While many Chinese engineers and scientists try very hard to stay in the West, they

gradually return to China since the 1990s. This accelerated after 2000 because of four main

reasons. Firstly, the economic recession in the Silicon Valley immediately after the Nasdaq

crash in 2000 creates serious unemployment in the Silicon Valley. Non-resident foreigners

including the Chinese find it difficult to get a job to stay in the US under the H1 working visa.

Secondly, the political xenophobia in the U.S. after the 911 incident pushes many foreigners

including the Chinese to look for jobs outside the U.S.. Moreover, the difficulty in getting a

student-visa after 911 also decreases the entrance of the Chinese postgraduate students into the

U.S.. Thirdly, the attractiveness of the large Chinese market and opportunity is an important

pull factor for the Chinese returnees. Fourthly, the competition of local governments in giving

financial incentives to Chinese returnees to come back and start their businesses in China is

also an attractive offer.

A. Four types of Chinese returnees

There are four identifiable ways in which Chinese returnees engage in ZGC’s knowledge

economy. I will use case studies34 to illustrate how they operate and to indicate their

contribution to the technological development of ZGC.

34 Note: all the names of the companies and the names of the owners/CEOs/directors have been changed to protect the

privacy of these companies.

19

Type 1: Transnationalism with strong local connections

Among the four types, the “transnationalism with strong local connections” type is closer

to the comparative advantage model, which is the archetype of the brain-circulation theory.

This type of Chinese returnees creates strong spatial division of labor between regions:

♦ R&D in the Silicon Valley: because of is experienced engineers and abundant venture capital;

♦ Manufacturing in Taiwan or Taiwanese OEM firms in Shanghai: because of its strength in

process innovation and efficiency; and

♦ Marketing, distribution and public relation functions in ZGC: because of the political

institutional resource or guanxi that are critical in opening up the domestic market, e.g. some

markets need special permission from the state.

Despite the similarity on the surface, there are two critical differences between Taiwanese

and Chinese transnationalism. The first difference is that the Chinese technology transfer is

basically one way – from Silicon Valley to Beijing – and not reciprocal two way flows. The

second difference is that the Chinese local functions are focused on the service side – namely

marketing, distributing and public relations activities – and not manufacturing, as in the case of

Taiwan.

Nevertheless, this is still the most advanced and complicated transnational networking,

which is limited only to very experienced Chinese returnees, because it requires a high level of

coordination.

Mr. Tang and the Tsinghua Tong Silicon Ltd.

Tong Silicon was set up by two Tsinghua alumni – Mr. Tang and Mr. Yong – in California

in 1999. Their main target is to persuade the Chinese Digital TV working committee – a

20

standard setting agency of the Chinese central government – to adopt their Digital TV

transmission standard (terrestrial reception), while China is converting the entire analog

television broadcasting system to a digital one. These Chinese returnees know very well that

technology superiority is not enough without careful guanxi-building.

To realize this goal, they went back to Tsinghua University and got great support from the

CEO of the Tsinghua Science Park – Professor Mei – another Tsinghua alumni, who was a

senior of Mr. Tang in the college days. In fact, many of the experienced Chinese returnees

belong to the same “Tsinghua Entrepreneur Group”35 based in the Silicon Valley.

Mr. Yong was hired as a full time professor and Mr. Tong was hired as an adjunct

professor in Tsinghua University. They set up a semi-public DTV Technology Center directly

under the Tsinghua University to further promote their own standard (or “Tsinghua standard” as

portrayed in the media) in Digital TV terrestrial transmission.

According to Mr. Tang, the major research is still done in their U.S. headquarters because

most of the engineers are there and the venture capital is there as well. He argues however, that

China is where the market is. Thus, their joint-venture effort with Tsinghua is an effort to

compete to be selected as the national standard in DTV transmission. Only by setting up a

joint-venture with a reputable local institution (in this case the Tsinghua University) would Mr.

Tang gain the legitimacy as a Chinese “local enterprise” qualified for the standard competition.

Without this localization and repackaging process as a “nationalist enterprise” (minzu qiye) – a

common practice by many Chinese returnee startups – it would be difficult to win the support

of the state, the media and the public, which all have strong “techno-nationalist” sentiments.

In January 2005, the media suddenly paid attention to the intellectual property rights of 35 The Tsinghua alumni network focuses on the Tsinghua startups in the Silicon Valley focusing on the U.S. market. It was

later moved back to Beijing as Tsinghua Entrepreneur and Executive Club, TEEC in April 2005.

21

the Tsinghua DTV standard. Theoretically, the major IPs should be owned by the Tong

Silicon – an American company that created these technologies – and not Tsinghua University.

Mr. Tang quickly responded publicly that this was a rumor and misunderstanding. He criticized

that the purpose of this rumor was to disqualify them in the standard competition among

“nationalist enterprises.” He emphasized in the newspaper and also during my interview that all

the intellectual properties (IPs) will be applied through Tsinghua University. This imply that

Tsinghua University will retain, or at least co-owned these IP created by Tong Silicon.

However, they were yet to work out a concrete plan36 on how to win the standard by avoiding

the political issue (e.g. an American company owned by Chinese returnees), while still

managing to make profits out of their research effort (via IP ownership) done almost

exclusively in the Silicon Valley.

Mr. Tang’s main strategy is to use Tsinghua University as its major image to compete for

the Chinese DTV standard. On the other hand, Tsinghua University also needs Tong Silicon’s

technologies to qualify for the quest for new DTV standard. Thus, both benefit from this

strategic alliance. Interestingly, their competitor is also using the same strategy. Three Chinese

returnees set up a company in the U.S. to do R&D on an alternative standard, while making a

strategic join-venture with the Jiaotong University in Shanghai. In the media, the competition is

always portrayed as the fight between two national universities for the national standard. This

kind of default media representation fits well into the plans of these two Chinese returnee

teams.

With the baggage of “technological nationalism,” the Chinese returnees need extra

political skills (much more than their Taiwan counterparts) to adapt to the “comparative

36 The concrete plan might be a secret that he will not release to the public at this stage.

22

advantage mode” of the West and China. If the IP issue is raised negatively in public (e.g. a

American company taking advantage of the Chinese manufacturers and consumers), these

Chinese returnee companies might be eliminated from the race. Thus, the truly

technology-savvy Chinese returnees, such as Mr. Tang, often complain that working on the

political issue is overwhelmingly more difficult than working on the technological innovation.

Despite the institutional constraint, this model is still the more promising transnational

model of technological transfer as their commitment to innovation for the Chinese market is

very high. Since they put all their eggs in one basket – i.e. the Chinese market, at least in the

take-off period – this type of Chinese returnee startups tend to have stronger motivations to

further localize their technological innovation. This is due to the deep embeddedness in the

local institutions and the local market-oriented innovation strategy.

Type 2: Transnationalism with weak local connections

For Chinese returnees working in the MNCs in the West, the least risky way of returning

home is to continue to work in the same or similar MNC in China. In fact, the most senior and

experienced Chinese returnees are often appointed or head-hunted by large MNCs to lead their

technological department in China. They are preferred for their unique role in mediating

between the Western technology management culture and the Chinese engineering culture.

Microsoft Research Asia, SUN Micro Engineering Center, Lucent’s Bell Labs in Beijing are all

setup by experienced Chinese returnees in this way. These returnees are often given full

autonomy to choose the location of the R&D center, hire the core R&D team members, and set

up the R&D priorities. It is not surprising that the R&D priorities often reflect the embodied

expertise of these pioneer Chinese returnees.

To exploit the ZGC’s knowledge assets to the MNCs’ global innovation networks, these

23

experienced Chinese returnees tend to locate their centers near the major university districts of

ZGC. Tsinghua Science Park outside Tsinghua University is the most famous university-run

science park. It attracts many MNCs because of its better built infrastructure and stronger

alumni-student-scholar networks, as compared to other universities including the Peking

University.37 One cautionary note is that mere spatial proximity to the university clusters does

not automatically allow MNCs to dominate the human resources from these top universities.

Graduate students from these top universities have plenty other job opportunities within the

ZGC. To take advantage of the spatial proximity to these major campuses, MNCs need to

establish collaborative internship programs, joint-research labs, joint R&D projects, and

sponsor training programs between the MNCs and the universities. The effect of these

programmes in exploiting the local knowledge assets is often determined by how well the

Chinese returnees understand the local engineering culture, and how well they can embed in

the local alumni-student-scholar networks. In this process of establishing the firm-university

linkage, the double identities of some Chinese returnees – the R&D director of MNCs and the

alumni (or adjunct professor) of the universities – provide the institutional linkage between the

embedded local knowledge assets and the MNC’s global innovation networks.

Mr. Li and Star China Research Institute (SCRI)

As the managing director and founder of Star China Research Institute (SCRI), Mr. Li – an

enthusiastic Tsinghua alumni – was one of the early inhabitants of the Tsinghua Science Park

back in 2001. Mr. Li was already an award-winning chief Java architect in the MNC mother

37 For more details on the alumni-student-scholar networks, please refer to Chen, Y-C (2006) “Why do multinational

corporations locate their advanced R&D centers in Beijing?” Journal of Development Studies. Forthcoming.

24

company in Silicon Valley before he was asked to head the SCRI. He was comparing a few

locations to host the SCRI in 2001, and finally decided on the Tsinghua Science Park – a more

expensive and less well designed office space – mainly because of his “connections” with the

Tsinghua alumni-student-scholar networks. 38 Through actively participating in alumni

activities and giving public comments on the problems of software industry in ZGC, he was

able to sell his version of “Silicon Valley culture” to a larger ZGC community.

The major function of SCRI is to focus on software engineering for global products. Most

of the 100 SCRI engineers (as of 2002) are in the software engineering group. Among them,

around 60% have a master degree, 2% have a PhD degree, and 35% have a bachelor degree.

81% of the engineers are below 35 years old, a relatively young workforce, which is a very

common phenomenon among MNC R&D labs in China.

As SCRI’s managing director, Mr. Li has high autonomy in setting up research groups and

research agenda that suit the interest of the local Chinese researchers. With his knowledge of

the Silicon Valley culture and the engineering culture of Tsinghua, he can manage to blend the

two together much better than non-Chinese foreign R&D directors.

As a Tsinghua alumnus, Mr. Li knows how to motivate the local engineers to innovate. He

focuses on cultivating the sense of mission, competition and ownership, which are values

ingrained in the Chinese engineer culture. He explains that his engineers are “productive” and

“enthusiastic” because they have a sense of ownership and engineering responsibility for the

products. Mr. Li thinks that high pay alone does not motivate them to work. Only the “sense of

mission” and “ownership” can motivate them.

Mr. Li also introduces the decentralized decision making structure and fun atmosphere

38 Interview of Mr. Li in Oct 2002.

25

borrowed from the Silicon Valley. To avoid the compartmentalization of various research group,

Friday afternoon teatime and sports time are designed to build friendship across different

research groups. He even imports the so-called “thinking chairs” directly from California to

build the “Silicon Valley office atmosphere.” However, outside the company in ZGC, Mr. Li

laments that there is no interaction among firms. This implies that his “in house Silicon Valley

culture” has very limited values outside the company, thus casting serious doubt on any

effective technological spillover of this model to ZGC.

The main difference between type 1 and type 2 is the constraint of MNCs (type 2) in

forming true technological collaboration with the local institutions in the way Tong Silicon

(type 1) does with Tsinghua University. Tong Silicon is willing to share even their IP with the

Tsinghua University in the standard competition. This is what MNCs will not do. On the

contrary, to protect their IP from leaking to the local institutions, MNCs are quite hesitant in

making deep technological collaboration with local institutions. In short, there is a lack of

formal firm-level collaboration, such as between MNC and university; MNC and IAS or MNC

and local firm. However, on the individual-level, Chinese returnees like Mr. Li continue to

enjoy the advantage of alumni connections in recruiting the best local talents to build his dream

R&D team within the MNC’s global innovation networks.

If there is any potential of technology spillover effect, it is probably the indirect

demonstration effect. Mr. Li has indeed set-up a model for a new decentralized R&D project

management culture, which does not only touched the hearts of his local employees, but will be

emulated by others in the larger ZGC technological community mainly through media

interviews, and eventually, through the mobility of engineers trained at SCRI to the local firms

or local institutions.

26

Type 3: Localism with strong local connections

The next two types (Type 3 and type 4) of Chinese returnees come back to China with little

or no transnational relations with the West. They are the majority among Chinese returnees.

Even if they have an office in the West, these offices do not involve in serious technological

innovation. These two types reflect the true meaning of “Chinese returnees”, as they have

returned from the West and maintain little transnational innovation.

Experienced Chinese returnees are often the target of the local research institutes because

of their eagerness to speed-up technology transfer and commercialization. The public research

labs from the CAS and top universities have undergone serious budget cuts since the 1990s.

The leaders of these public labs are under tremendous pressure to self-finance their labs

through technology licensing, provision of technological service, and spin-off

institutional-affiliated startups. They believe that the Chinese returnees might be able to help

them commercialize their technology rapidly.

To attract Chinese returnees to the universities, many campuses have initiatives offering

adjunct professor positions and providing incentives to set up startups in the incubators or

science parks run by the universities. Many Chinese returnees respond to this incentive of

becoming the adjunct professor not merely because of the prestige, but because of the monetary

value of this unique social capital – the ability to formally tap into the institutional resources of

the top universities such as Tsinghua University.

The competitive and critical resources for adjunct professors include the eligibility for

large public R&D funding, access to the public lab facilities, the right to utilize student

internship programs legitimately, and frequent interactions within the alumni-student-faculty

networks. By tapping into the top research institutions with one foot, the chance of survival of

their startups – the other foot – is much higher.

27

Mr. Tan and Bango Technologies Inc.

Bango Technologies is a new IC design company focusing on SOC (system on chip)

design. It came to the Tsinghua Science Park in 2004. Mr. Tan is a Tsinghua alumni with a

Denmark Ph.D. degree. He worked in Ericsson and taught as a professor in Denmark before he

moved to southern California to work with Globespan for five years. He owns seven patents

and has written four books. With this qualification, he was hired as an adjunct professor at the

Tsinghua University three years ago, before he actually came back in 2005 to create his own

startup in the Tsinghua Science Park.

He argued that the locational advantage of the Tsinghua Science Park is the unique

alumni-student-scholar networks. A very good relation with Tsinghua professors allow him to

consult them easily when he encounters problems. He said that his technological strength is in

IC application, while the strength of the Tsinghua University is in the IC system. Since IC

application relies on the understanding of the IC system, the strength of his startup complement

with the strength of the Tsinghua University. Like Mr. Li, he is determined to bring to ZGC the

Silicon Valley culture of equality, communication and fun.

To cultivate a loyal team, he chooses to hire and train fifteen fresh graduates from the

Tsinghua University, all recommended by other Tsinghua University professors. This assures

their quality and loyalty to the company.

To reduce the risk of startup in ZGC, Chinese returnees try to find formal and informal

ways to tap into the research institutional resources embedded in the top universities, the CAS

or ZGC in general. By making strong connections with the local research institutions, Chinese

returnees can utilize the public resources, such as the research labs, to reduce the cost of

operation. For example, Chinese returnees in pharmaceutical startup need to collaborate with

the hospital to obtain cheap sample (e.g. blood sample) for their clinical research. They also

28

rely on university to do a lot of the testing because they cannot afford to buy all those

expensive equipments as a startup in ZGC. Secondly, they have higher chance of obtaining tacit

knowledge and information for obtaining public grants from the faculties. Lastly, it would be

much easier to access cheap student labor in the form of flexible part-time work or internship.

By sharing cost39 and tacit knowledge through these informal local connections, Chinese

returnees contribute to the informal university-industrial linkages that can help to bridge the

gab between academic research and commercial research.

Type 4: Localism with weak local connections

The majority of the Chinese returnees have neither the transnational connections with the

Silicon Valley nor the local connections with the public institutions in ZGC. The problem with

this type of Chinese returnee startups is that they have a hard time competing with the first

three types for largely institutional-based (institutionalized) resources and governmental

contracts. They also have difficulty competing with local firms that are more familiar with the

ways to access market-based resources.

This type of dis-embedded startups by most Chinese returnees often faces high startup

cost (often hidden) and high failure rate (often locked into low-level, repetitive innovation

activities). The barriers they are facing make their situation not much better than the local

startups, except that they may enjoy a few specialized incentives targeted for Chinese returnees.

B. Comparing the different types of Chinese returnees

By typologzing the Chinese returnees above, this article reveals the great internal diversity

39 The sharing of cost with the university is a tricky and flexible depending on the guanxi or bargaining skills or sometime

kickbacks.

29

among the Chinese returnee community. On the top end of the spectrum (type 1), we have the

most experienced and charismatic leaders with both transnational and local connections. On the

bottom (type 4), we have the least experienced returnees (fresh graduates for example) with

little guanxi-building skills.

30

Figure 2: Comparing the four types of Chinese returnees and their role in technological development Strong local connections Weak local connections

Transnational networks

Type 1 ♦ The most experienced Chinese returnees, who are

often risk-taking entrepreneurs ♦ Embeddedness: Individuals and firms are usually

deeply embedded in the local public institutions to share cost, risk, tacit knowledge, human resources and political resources.

♦ Limitation: higher risk of failure due to multiple uncertainties concerned with new technology, larger capital investment and the sensitivity of techno-nationalism.

♦ Contribution: Engaged in development and transfer of new technology for the Chinese domestic market. Their embeddedness makes their firms more porous and more conductive to technological spillover outside the firm. They have the tendency to be more committed towards localizing their innovation activities in China.

Type 2 ♦ Experienced Chinese returnees, who know

how to integrate local knowledge assets into the MNC global innovation networks.

♦ Embeddedness: Individual Chinese returnees often fully utilize their individual cultural affinity to extract the local knowledge assets. On the firm-level, the MNCs normally keep a distance from the local research institutions to protect their own IP.

♦ Limitation: There are two-way flows of technology within the MNC innovation networks, but the MNCs are conscious of the need for better security against technology leakage to the outside firms.

♦ Contribution: the technology management practice and culture they introduce may have indirect demonstration effects for other local firms to follow.

Local networks

Type 3 ♦ Experienced Chinese returnees, who have no

transnational networks, but manage to form partnership with local institutions in formal and informal way.

♦ Embeddedness: The Chinese returnees either build the local connections themselves or through trustworthy local partners, who have close relations (e.g. former classmates). The degree of embeddedness varies.

♦ Limitation: lack of transnational networks to keep in touch with new technology in the West. Local connections are mainly used for informal sharing of operating cost, access to government controlled venture capital, access to governmental contracts. Some may engage in joint-research with local research institutions.

♦ Contribution: With higher survival rate due to their embeddedness, their informal research networks form informal university-industrial linkages that are useful for the mobility of students and faculties, and for the exchange of tacit knowledge.

Type 4 ♦ Experience or less experienced Chinese

returnees, who have neither transnational networks nor local networks, but are confident enough to startup because they believe they have technological advantage over the local competitors.

♦ Embeddedness: The Chinese returnees either build the local connections themselves or through trustworthy local partners, who have close relations (e.g. former classmates)

♦ Limitation: Lack of transnational networks limits their capability to engage in advanced technological development. Lack of local connections makes their startups highly vulnerable to failure.

♦ Contribution: Compared to local startups, the Chinese returnees bring in new ideas and new technological practices that create higher competition for the local startups. Their contribution is the least among the four types because of their high failure rate.

Source: by author.

For type 1 Chinese returnees, it is possible to practice a model of transnationalism similar

to Hsinchu, which takes advantage of two regional assets in China. Unlike Hsinchu however,

31

this type of transnationalism is only limited to a few Chinese returnees with political skills.

This is because top technological firms must have their R&D centers in global high-tech hubs

like the Silicon Valley, while ZGC is still in its infancy in terms of new technology

development. Many Chinese returnee firms struggle to collaborate (or joint-venture) with

universities and the CAS institutions because they want to gain political support and legitimacy

as local “nationalist enterprises.” The gesture of becoming a “nationalist enterprise” is crucial

in overcoming “technological nationalism” shared by the central government and the public in

general. Transnationalism with political skills seems to be an inevitable requirement for

serious technological startups wanting to access the domestic market, which in many fields, is

still controlled by multiple levels of state institutions.

Type 2 Chinese returnees riding on the MNC wagon may have positive contribution in

technology spillover in the long run. Through the demonstration effect, these contemporary

experienced Chinese returnees bring back the new culture of doing R&D and new ways of

R&D project management, which are seriously lacking in ZGC. These conventions and

practices often accumulate in the research teams within the MNC networks. Through media

exposure, public talks, and working with graduate students, these experienced Chinese

returnees obviously want their influence to go beyond the MNCs’ boundaries to the local

knowledge economy.

For type 3 Chinese returnees, the uniqueness of the ZGC knowledge economy is the

importance of institutional assets in shaping the technology development path. The presence of

thousands of isolated startups may mislead us into thinking that a real Marshallian knowledge

cluster is in the making. However, in reality, only startups with (public) institutional

connections have the resources to sustain uncertainties in the highly competitive domestic

market. Thus, the success of the Chinese returnees will depend on whether they know how to

32

tango with the public institutions by networking the critical assets from both the public

institutions and the market. With stronger embeddedness and being less footloose, they have

more positive influence on the technological development of ZGC at large.

Lastly, type 4 Chinese returnees are the majority. They have only bare-hand enthusiasm

but no institutional guanxi. Most of them begin to realize that they are not competing on the

same level playing field with the other three types of Chinese returnees after they move back to

ZGC. Moreover, their competitors are not only the Chinese returnees but also the local startups,

which may not be technically savvy, but have strong institutional advantages. Nevertheless,

many of this type of Chinese returnees are genuinely building the networks from below. Their

survival is a signal of an emergence of a new market-based, horizontal networked industrial

cluster, which is badly needed for the sustainable technology and industrial development in

ZGC.

Conclusion

By re-contextualizing the often neglected assumptions of the “brain-circulation” theory,

this article attempts to reassess the relevance of the major hypotheses of this theory in the case

of ZGC in Beijing, and in this way, caution against decontextualized application of this theory.

The critical hypotheses of the “brain-circulation” theory are based on the successful role of the

Taiwanese transnational technological communities in exploiting the comparative advantages

of two regional assets, and in return, facilitate technology transfer and industrial upgrading of

the Hsinchu high-tech region. These critical hypotheses include: (i) decentralized industrial

structure with specialized producer networks; (ii) trust-based inter-firm networks that induce

learning; (iii) critical financial infrastructure for IT startups; and (iv) the role of the state in

facilitating technology transfer, rather than producing technologies.

33

The critical reassessment of the relevance of the “brain-circulation” theory is done through

the case study of IT entrepreneurship in ZGC, Beijing. ZGC has a very different historical

starting point from Hsinchu. It departs from all the major assumptions listed above. In many

ways, the technological and industrial development in Taiwan’s high-tech sector is pretty

unique and probably hard to generalize. Thus, students of technology transfer should be more

aware of the implicit hypotheses of the “brain-circulation” thesis, and thus should be more

cautious in applying the “brain circulation” argument to any transnational networks. Policy

makers in China and elsewhere should also be aware that Hsinchu’s model of brain-circulation

is very difficult to reproduce elsewhere. Saxenian herself is always cautious in the direct

application of this thesis to the China and India cases.40 When applied to China, Saxenian

emphasizes on the Taiwanese transnational networks that extended from Taiwan to China

following the relocation of the Taiwanese manufacturing function to the low production

location in China. Thus, the transnational networks are still dominated by the Taiwanese, who

have the first advantage, except that they are now operating in three places (Silicon Valley,

Hsinchu and Shanghai) and not two places (Silicon Valley and Hsinchu).

Since the mid-1990s, Chinese returnees of all the four types analyzed above continue to

return to China especially after the NASDAQ crash in 2000. The contribution of these Chinese

returnees to the local technological development vary widely depending on the two main

variables, which result in four different typologies. Only experienced Chinese returnees that

can manage the transnational networks (R&D in the West, manufacturing in Taiwan or

Shanghai, and marketing and PR in Beijing) and have strong local institutional connections

(informal or formal sharing of cost, access to public research facilities and grants, access to 40 Saxenian, A. (2005) “From Brain Drain to Brain Circulation: Transnational Communities and Regional Upgrading in

India and China.” Studies in Comparative International Development, Summer, 40(2): 35-61.

34

institutionalized venture capital, access to governmental contracts, influence on technological

standard direction ..etc.) will have a stronger impact on the local technological development.

Chinese returnees that have neither transnational nor local networks will most likely repeat the

high failure rate of local startups. Between these two extremes, Chinese returnees in the MNCs

(with weak institutional linkage but strong personal connection to the local knowledge assets)

and Chinese returnees with weak transnational connection but strong local institutional

connections will have some indirect impact though building informal linkages between the

university and industries. This informal and personal linkage helps to mobilize the rich

alumni-student-scholar networks of ZGC.

In sum, there are two fundamental differences between the Chinese and Taiwanese

returnees that make brain-circulation a high-end act among the most experienced Chinese

returnees rather than a common practice.

The first difference is in the length of the transnationalism history. Taiwan has a longer

history by starting 30 years earlier than China in sending students to the US (since the 1960s).

The Taiwanese graduates have a much longer history engaging in the technological revolution

in the Silicon Valley. Thus these pioneers have already acquired stronger and richer

transnational networks when they began to return to Taiwan to build the Hsinchu technopoles

in the late 1970s. With the number of overseas students going to the US peaking only after the

Tiananmen incident in 1989, Chinese overseas students in the U.S. are considered latecomers.

Thus, it is not surprising that they form weaker transnational networks when compared to their

Taiwanese counterparts. Thus, even though the transnational networks of Chinese returnees are

growing, those who can manage a transnational operation (like their Taiwanese counterparts)

are still rather rare and highly valuable.

Second, China is still in the process of transforming from a planned economy to a

35

market-based economy. During these transitions, many important assets (research grants,

venture capital, public research labs) are still by and largely controlled by the various levels of

governmental agencies. On the contrary, the Taiwanese government has been privatizing public

research assets (via spin-offs) and promoting private venture capital since the early 1980s. This

contrast means that in order to be successful, the Chinese returnees have to tap into the

institutionalized resources, which are not yet readily available in the market. Thus, the ability to

embed themselves in the local institutions and exploit the institutionalized assets becomes a

determining factor of technological development.