In 2004, SPRIE launched the three-year Greater China Networks (GCN) research program. Its goal is two-fold. First, the GCN seeks to advance understanding of the systems of innovation and entrepreneurship that drive Greater China's ascendance in high technology. Second, it will study the nature and impacts of the region's integration into the global knowledge economy. The research agenda includes a focus on activities or institutions that underpin systems of innovation and entrepreneurship, especially for the new generation of ascending high tech regions in Greater China. These include university-industry linkages, globalization of R&D, venture capital, new firm formation and development, and flows of technology and business leaders.

On February 25, 2005 a panel of researchers presented preliminary findings from surveys of 176 CASPA members (171 valid responses). The discussion included the interpretation of study findings, key issues and trends, and implications for professionals, corporate managers, and policymakers in the semiconductor industry. Panelists included Hsing-Hsiung (Hubert) Chen (pictured above left), SPRIE visiting scholar and director of Integrated Research Division, ITRI; Jian Hung Chen (pictured above center), SPRIE visiting scholar and ITRI researcher; and Mr. David Wang (pictured above right), vice president, Fibera Inc., CASPA President 2003-2004.

The Rise of Greater China in the Global Semiconductor Industry

Dr. Hsing-Hsiung Chen shared with the participants some striking statistics that clearly highlighted the rise of Asia and Greater China in the semiconductor industry. While in 1985 Asia accounted for only one percent of worldwide semiconductor production and 7 percent of worldwide semiconductor consumption, the numbers are expected to reach 21 percent and 42 percent respectively in 2005. The story for Greater China is particularly salient. Take the year of 2004 as an example. In IC design, the top thirty firms in this sector commanded 86 percent of the worldwide market share and half of them were founded or co-founded by Chinese professionals. In fabrication, eight of the top ten foundry companies were Chinese and together they covered 88 percent of the market. A similar story can be told about packaging and testing companies, with eight of the top ten companies located in Taiwan and mainland China and accounting for half of the global market share.

The Connection of Silicon Valley Professionals to Greater China

Mr. David Wang pointed out the role of the Silicon Valley as a hotbed for entrepreneurs, managers, and engineers that found or join semiconductor companies in Greater China. For example, CASPA, a leading semiconductor professional organization for Chinese Americans, has over 3,500 members in the United States and Asia, with the majority residing in the Silicon Valley. Among 171 CASPA survey respondents, half had worked in the United States and Greater China. Fifteen percent had worked at more than one place in Greater China.

Many aforementioned top Chinese semiconductor companies have strong ties to the Silicon Valley. Companies such as SST and ISSI were founded by Chinese in the Silicon Valley but have strong presence in the Greater China region in production and sales. Other Greater China-based companies like UMC, SMIC, ASMC, HuaHong NEC, and Winbond have drawn an impressive list of senior managers from Silicon Valley companies.

Job Movement Trends

Dr. Jian-Hung Chen presented some preliminary findings of the surveys and identified some interesting trends in job movement. When being asked "Will you consider working permanently in a region different from your current location?" 7 percent of the respondents revealed that they were already considering moving, while 52 percent said they would consider moving within 1-3 years. Young professionals (with fewer than ten years of work experience) were somewhat more likely to move within three years (65 percent) than more experienced professionals (~57 percent). In terms of geographic destinations, only 10 percent favored moving to or relocating to elsewhere in the United States. The rest all preferred relocations within or to Greater China. Shanghai was the most popular site, voted by 50 percent of the respondents as their preferred destination. Taiwan and Beijing followed with 13 percent and 14 percent of the votes. Breaking down the data according to the origin of the respondents into Mainland China, Taiwan and the United States (indicated by the location of undergraduate education), Shanghai remains the topic choice for all three populations. Interestingly, professionals originally from Taiwan were more likely to move to Shanghai (39 percent) than return to Taiwan (33 percent) and none of the professionals originally from mainland China regarded Taiwan as an option.

An examination of key influential factors on movement decisions may shed some light on the observed trends. Growth potential, family matters, and quality of Life were ranked the three most important factors in one's movement decision. others' success/failure examples and company decision received the lowest scores. Senior professionals were more likely to be influenced by company decision (compared to junior professionals), while junior professionals more by others' success/failure examples (compared to senior professionals). Host company size did not seem to be a noteworthy differentiator.

Finally, in terms of the type of company one hopes to work for after relocation -- the choices being startup, multinational, local company and company decision (i.e. internal transfer) -- professionals moving to Taiwan strongly favored local company and internal transfer, while professionals moving to Beijing and Shanghai preferred startup much more than the other three modes.

Greater China's Regional Advantage

Several participants chipped in their insights on the mainland's labor cost advantage. For semiconductor professionals moving from the Silicon Valley to Shanghai, "The rule of thumb is a 1/3 to 1/2 pay cut, although they may be awarded with [the realization of] the growth potential through other forms, such as stock options" observed Mr. Wang. Such expatriate compensation packages are still substantially higher than those for local hires, whose salary level is usually 20-25 percent of that of the Silicon Valley. Overall local income increase is estimated to be about 5-10 percent a year, although most of it is due to the move up the semiconductor industry value chain. Salary increase for the same position is probably less than 5 percent a year. Entry-level salary has also been heavily affected by the recent surge in the local supply of engineers, thanks to a substantial increase in college engineering graduates. Hence, taking into account differences in cost and experience, "for now, the common and effective practice seems to be hiring one expatriate for every 25 local hires," said Dr. Jian-Hung Chen.

Going beyond cost considerations, discussions centered around a provocative question raised by a member of the audience: "Is it just a question of time or is there any other fundamental piece that needs to get in place before Chinese firms can penetrate high value-added nodes in the semiconductor industry value chain [i.e. design]?" SPRIE Director, Professor Henry Rowen responded by pointing out that although to this day, probably only a handful of the 400-500 design companies in Mainland China have real technologies and products, let alone profits, Chinese fabless design companies are growing rapidly and can find lots of applications in the domestic market, especially in telecommunication. Dr. Wang concurred that it is probably more of a matter of time. "If a foundry moves to the next-generation technology, firms up- and down-chain will follow suit and build complementary capacities as well. It's a clustering effect. It just takes time."

Other Issues

Panelists and the audience also engaged in lively discussions about corporate strategy for semiconductor multinational companies in Greater China, family, social and cultural factors affecting the flow of talents, and inter-regional collaboration. Networks and flows of managerial and technical leaders -- particularly their connection to the Silicon Valley and their worldwide reach -- are new priority areas of research for SPRIE.

From a strong pool of applicants, including masters and Ph.D. students with backgrounds from law and sociology to computer science and engineering, SPRIE directors selected Ming Gu and Victoria Wu. They will work with SPRIE faculty and senior researchers at Stanford, conduct field research and data analysis in Beijing and Shanghai during summer 2005, and contribute to SPRIE research forums and publications. As part of an expanding initiative on innovation and entrepreneurship in Greater China, SPRIE is pleased to announce the selection of two outstanding Stanford students as inaugural SPRIE Graduate Research Fellows.

In light of the rise of Asia in research and development (R&D) and the challenge it poses on American supremacy, SPRIE invited industry and academic R&D leaders for a panel discussion entitled "The Globalization of R&D" on February 10, 2005. The panel included Dr. John Seely Brown, visiting scholar, Annenberg Center, USC; Dr. Kris Halvorsen, vice president and director, Solutions and Services Research Center, Hewlett-Packard; and Dr. Yoshio Nishi, director of research at the Center for Integrated Systems, director of Stanford Nanofabrication Facility, National Nanotechnology Infrastructure Network. Participants discussed a wide array of issues, including the economic rationale for new models of R&D, national/regional comparative advantage in R&D, and the coordination of global R&D.

The Economic Rationale for New Models of R&D

Dr. Nishi highlighted the economic rationale behind the quest for new models of R&D. While back in the early 1990s, a $200 million investment in R&D would grant a semiconductor company a one-year lead in technology, by the early 2000s, a one-year lag would transpire with the same investment level. Such an escalation of R&D cost points to the mounting importance of the efficiency of R&D--or as Dr. Nishi put it, the importance of generating "the right technology at the right time for the right cost." The economic forces will not only alter how R&D activities are organized and distributed within and across firms, markets, regions, and countries but also influence the breadth and depth of knowledge searches. For example, R&D alliance might become a viable and lucrative scheme for cost/risk sharing in R&D. The search for non-silicon-based devices might rise in importance as silicon fabrication reaches its limits. By the same token, the division of innovative labor across nations/regions might deepen to further exploit respective comparative advantages.

Regional Comparative Advantage in R&D

One strand of development is the globalization of R&D, which necessitates comparative advantages across regions. Dr. Brown maintained, "I'm moving my analysis from individual firms to [regional] 'niches.' What I see happening is that thousands of [regional] niches are developing all over the place. What's interesting is how dynamic these niches are in building their unique capabilities." The availability of innovative talents, for example, varies significantly across regions. Invoking "the law of large numbers," Dr. Brown pointed out that given its enormous population size, Asia could produce a large number of engineers, even if they are only a tiny fraction of the total population. Currently, the U.S. produces 50,000 engineers every year; the number is 500,000 for Asia--and it is rapidly growing. Meanwhile, more and more immigrant talents choose to return to their home countries after receiving higher education and some work experience in the U.S. Few U.S. companies can afford to ignore such alarming trends. "We need to move with the market for talent," commented Dr. Halvorsen who overseas HP's global R&D activities. Take HP's R&D effort in Bangalore, India as an example. The effort had a humble start in the mid-1980s. Yet, within ten years, the number of local technical staff grew to 3,000. Today, the number is approaching 10,000.

Market-specific demand also pushes R&D to relocate. As Dr. Halvorsen put it, "when success depends on [geographical] closeness, … you need to do design in close loop with the rest of the activities." Furthermore, overseas R&D might well find its way back into the U.S. As explained by Dr. Brown, "The rise of the middle class in China and India at 1/10 of the price point [of the U.S.]" could spur innovations at 1/10 of the price point. Innovations taking place in China or India might be totally unheard of in the U.S. and eventually finds its way into the U.S. market.

The Coordination of Global R&D

While the globalization of R&D brings many promises, it also poses acute challenges to firms that need to coordinate R&D efforts across national boundaries. As Professor William Miller pointed out, "Increase in R&D cost forces specialization. Then you have to put together an assembly of specialists. The problem is that they are everywhere. Therefore, being able to pull them together becomes the differentiator." The story of Li & Fung serves as a perfect example. Li & Fung is a global leader in the apparel business. In 2002, the company contracted with 7,500 factories in 37 countries and generated a revenue of $5 billion. In an industry with thin margins of a few percent, the company continues to uphold a return-on-equity of 30-50%. Yet, Li & Fung owns no factories. Its competitive advantage lies entirely in its expertise in assessing and orchestrating the unique capabilities of each of the 7,500 suppliers. As Dr. Brown summed up, "Making money will depend less on what you own than on what you can mobilize--[i.e. the ability to] orchestrate."

In a parallel argument, Dr. Halvorsen proposed the new model of "meta-national" R&D. Different from the traditional multinational setup, where R&D is orchestrated from the center and diffused to the peripheral, in a meta-national setup, innovation for different parts of the system are consciously placed in different parts of the world. Advances are made in parallel and feedbacks flow bi-directionally.

An even more decentralized model was advanced by Dr. Brown. Dubbed a "swarm ecosystem," such a system is characterized by one (or more) assemblers and hyper-competition among a constellation of component suppliers. The assembler merely provides the focal model with no detailed design, and leaves it to the component suppliers to compete for coming up with the best fit. In this model, the assembler does not orchestrate the development process from top-down; rather, progress is made from the bottom-up. Yet, at the end of the day, only the fittest component suppliers survive and the result is a highly efficient and competitive system that best exploits its own niches.

Other Issues

Panelists and the audience also engaged in lively discussions about intellectual property rights, organizational learning, institutional innovations, the role of public policy, and the impact of culture on innovation. The globalization of R&D--particularly rising competencies in Greater China and their network of relations to Silicon Valley and their worldwide implications--is a new priority area of research for SPRIE.