A recent report from Duke University that critiques the supposed gap between the number of American science and engineering (S&E) graduates and those of merging economies -- especially China's -- has led to false reassurance that the U.S. lead in science and technology is not under threat from China. It would be a grave mistake to drop our concerns about China's competitive challenge.

First, the Duke report simply claimed that China's true number of science and engineering bachelor degrees was 351,000, rather than the widely reported 600,000. Coupling this with an upward adjustment for American graduates still left China producing 214,000 more such degrees than the United States.

Moreover, undergraduates are only part of the concern. China's production of those with doctorates has increased rapidly. By 2003, China's homegrown science and engineering doctorates numbered almost half of the U.S. total.

Chinese were also earning large numbers of doctorates abroad. In 2001, the number of Chinese S&E doctorates earned in Japan, the United Kingdom and the United States equaled 72 percent of the total of S&E doctorates earned by American citizens and permanent residents.

Since 1975, China has increased its global share of S&E doctorates from zero (courtesy of the Cultural Revolution) to 11 percent, not counting doctorates earned overseas. During the same three decades, the U.S. global share has fallen from half to roughly 22 percent.

More worrisome than the aggregate numbers is American universities' reliance on foreigners who earn doctorates. In engineering, foreigners account for over half of America's doctorates, and in computer science just under half.

If foreign-born holders of doctorates continued to stay in the United States, we wouldn't have to worry. Unfortunately, there are many signs that it is becoming much harder to retain them.

One need only look at the flow from Taiwan, one of the former main sources of American S&E doctoral degrees, to see what could happen. Up until 1994, Taiwanese earned more science and engineering doctorates in the United States than members of any other foreign nationality. By 2000, their numbers had plummeted because economic and educational opportunities at home were more appealing.

The Taiwanese didn't just stop coming to America. They also began to leave. As Taiwan's tech sector boomed in the 1990s, huge numbers of Taiwanese technologists (estimates range as high as 100,000) left America for home and took their technical skills with them.

Our two current biggest foreign sources of technologists, China and India, appear to be following Taiwan's path. China has begun to lure back large numbers of technologists. China's central and local governments offer free office space and other benefits to attract technologists home. These inducements are working. A 2005 survey of the Chinese American Semiconductor Professionals Association's members showed that the vast majority regard China as the most likely future work destination, and they rated Shanghai higher than even Silicon Valley on career potential. India's recruitment efforts have also started to bear fruit.

The challenge is not simply keeping up the numbers of technologists in America. China by many measures has improved its technological capabilities. On the Georgia Institute of Technology's Index of Technological Capability, China has more than doubled its index score over the past decade. China now ranks fourth behind the United States, Japan and Germany.

This rapid ascent is not surprising given China's increasing investments. China's research and development spending as a percentage of gross domestic product has tripled to 1.3 percent in the last decade, even while its GDP has ballooned. Few emerging economies spend even 1 percent of their GDP on research.

U.S. patents invented in China are also on the rise. Information-technology patents from corporations' Chinese technologists have risen from 134 in 1997-2001 to 482 during 2002-04. As a first step to meet this challenge, we should increase federal spending on basic and exploratory research. Our R&D spending has been flat at 2.6 percent of GDP for four decades, but the share of federal spending has declined from two-thirds to one-quarter.

Given that corporations now de-emphasize basic scientific research, the federal government should further support the basic research that could maintain our lead at the cutting edge of technology.

Increased federal funding would also address the issue of the falling share of investment in certain disciplines. With spending flat, the rising share commanded by biomedicine has meant a falling share spent on engineering and physics.

Federal support may also play a direct role in increasing interest in pursuing a science education. Since the 1950s, the number of undergraduate S&E majors in America has risen and fallen in line with federal research funding, as Professor Henry Rowen of Stanford University has pointed out.

Before meeting China's challenge, we first must recognize it. Complacency in reaction to "good'' news that China is producing fewer S&E graduates than commonly thought is not the answer.

Multinational corporations (MNCs) have increasingly located research and development (R&D) in developing countries such as China and India since the 1990s. On the one hand, governments in developing countries are eager to attract R&D to their local economies; on the other hand, developed countries are concerned about losing their competitive advantage due to R&D offshoring. At the same time, intellectual property protection is a growing concern.

What are the MNC R&D labs actually doing in China? Quan noted that her 2004 survey of MNC R&D labs in information technology industries in Beijing found that these MNC R&D labs are not just providing technical support, product localization, or product development for the local market; rather, they are developing products for the global market. Her study documents an emerging spatial division of labor in R&D based on the increasing specialization of R&D activities.

Ensuring returns appropriation

Appropriating returns is essential to continuous R&D investment. However, returns appropriation is not necessarily realized through formal IP protection institutions such as the patent system. As the growing trend of globalization of R&D has evolved to this new stage characterized by MNCs locating R&D labs in developing countries, it provides a good test bed to further explore more theoretical mechanisms of IP protection. Considering the weak intellectual property rights regimes these developing countries typically have, it is crucial for MNCs to find an effective way to protect their valuable technologies thus facilitating returns appropriation from their R&D activities in host developing regions. It is in fact the effective means of IP protection that can greatly assist MNCs' location of R&D offshore, in addition to other well-known incentives such as low cost R&D labor and market attraction.

R&D specialization essential

Using evidence from MNC R&D labs in Beijing and Shanghai, Quan's study proposes that R&D is further specialized within MNCs' global R&D network. Furthermore, IP protection and returns appropriation can be realized through such R&D specialization. The key proposition is formulated as below: 'Hierarchical modular R&D structure can be an effective way for MNC R&D labs to protect their intellectual property and thus facilitate returns appropriation in weak IPR regime developing countries'. This 'hierarchy' includes 'core R&D' and 'peripheral R&D', based on two dimensions--technology value-added, desire and ease of IP protection. While 'core R&D' is mostly done in developed countries, 'peripheral R&D' is conducted in developing countries. Dr. Quan's study suggests that this hierarchical modular R&D structure facilitates the global configuration of MNC R&D labs.

Slides from this presentation can be found at the event link below.

SELECT: What is the current size of the outsourcing market? What percent of U.S. software development, call centers, etc., have already moved to developing nations? Is the amount of outsourcing still increasing?

Dossani: To provide some perspective, although the off-shoring of services has been going on for mans Nears, technology led by the widespread use of the Internet has changed things. The resulting new twist in the provision of services is that the required interaction between the seller and the consumer has been substantially limited. The advances in information technology made possible the parsing of the provision of certain services into components requiring different levels of skill and interactivity As a result, certain portions of the serviced activity that might or might not be skill-intensive, but required low levels of face-to-face interaction could be relocated offshore. The sequence of events that enabled this process is the following:

First, the digital age allowed (or, at least, revolutionized) the conversion of service flows into stocks of information, making it possible to store a service. For example, a legal opinion that earlier had to be delivered to the client in person could now be prepared as a computer document and transmitted to the client via e-mail or, better yet, encoded into software. Easy storage and transmission allowed for the physical separation of the client and vendor as well as their separation in time. It also induced the separation of services into components that were standardized and could be prepared in advance (such as a template for a legal opinion) and other components that were customized for the client (such as the opinion itself) or remained non-storable. Taking advantage of the possibility of subdividing tasks and the economies that come with the division of labor, this reduced costs by offering the possibility of preparing the standardized components with lower-cost

labor and, possibly at another location.

The second fundamental impact of digitization was the conversion of non-information service flows into information service flows. For example, the sampling of tangible goods by a buyer through visiting a showroom is increasingly being replaced by virtual samples delivered over the Internet. Once converted to an information flow, the service may also he converted into a stock of information and subjected to the above mentioned forces of cost reduction through standardization of components and remote production.

Third, the low-cost transmission of the digitized material accelerated the off-shoring of services. Services such as writing software programs which were off-shored to India in the early 1970s were enabled by digitized storage and, in the 1980s, by the standardization of programming languages. Still later, as digital transmission costs fell in

the 1990s (just as digital storage costs had fallen earlier), even nonstorable

services, such as customer care, could he handled offshore.

The offshore services outsourcing market (excluding software development) is still small and will probably be approximately $10 billion for 2005. It employs about 500,000 people, two thirds of whom are located in India. The rest are widely distributed, with developing Asia and Ireland accounting for most of the remaining employment. About 60% of the employment is in call centers. The U.S. and U.K. call-center industry together employ about five million people, so the percentage of offshore jobs is still small. It is even smaller for other services.

Offshore software development employs approximately another 500,000 people. This compares with U.S. employment of about two million. This is a larger percentage of the total software development labor market. although most of the outsourced work is programming, while work such as systems integration and design continue to be done in the U.S.

The growth rate is still high and there are concerns about whether or not this rapid growth rate will hurt the quality of work done. However. this rate will still likely he in excess of 30% in 2005 and 2006. The reason for this is the massive wage differential.

Clearly there have been massive failures as well as outstanding successes in outsourcing. What are the critical success factors for making outsourcing work?

The infrastructure (telecom. finance, power) has all been standardized, although the solutions might not he the same as in developed countries. The critical success factors are two: the quality of labor and supervision; and managing growth. Unbelievable there is a growing shortage of labor. The result is that the quality of work is declining. Project supervisory skills are also in short supply. Managing growth, especially keeping attrition

under control, training, developing new vertical skills, moving into back-office work, and offering the client turnkey packages are some of the critical managerial factors for success.

Short of being willing to work for $15,000/year, what can western IT professionals do to provide sufficient value to prevent their functions from being outsourced?

The U.S. educational system still turns out a good product. It is sufficiently ahead of the comparable Indian product so that a recent computer science/computer engineering graduate from the average U.S. university can earn a premium of at least 100% over his Indian counterpart from a good university such as the IlTs, with substantially higher premiums for graduates from schools such as MIT and Stanford. The problem occurs more with mid-career professionals. Those with older skills are unable to compete with freshly trained graduates from India. Therefore, they need to update those skills regularly and take advantage of opportunities to globalize and convert them into managerial skills. This may have to he mandated at some point, as has happened in the financial sector, where stockbrokers need to regularly sit for exams to renew their licenses.

That said, most of the offshore jobs are relatively low-skilled. For example, the single largest category of offshore services is outbound calling for the financial services industry for selling mortgages or collecting overdue receivables. The work is routine, based on scripts that pop-up on the computer screen in response to prompts.

Do these findings suggest that developed countries are likely to be only marginally threatened by the globalization of services? Even if high-end work is stays within developed countries, as has happened in the software industry, the problem is that not everyone in developed countries can readily shift to high-end work. Since the 1960s, the shift in the economies of developed nations towards service-based economies certainly increased the number of highly-skilled service workers, but there was an even greater swelling in the number of other less-skilled service workers. This is partly a consequence of the nature of many services as linked, inseparable sets of activities with different

skill levels, combined with a pyramid of labor requirements, i.e., there is more demand for lower-level work than for high-end work. In manufacturing. the unemployment created by the reduction in demand for blue collar labor in developed countries was offset by the absorption of much of the surplus labor into service industries, often with minimal training. But the shift of low-end service workers to high-end workers will require a longer period of re-education and may have significant interim consequences on unemployment rates.

The threat to developed countries is increased by the fact that, apart from software, the largest growth in off-shoring is happening in business services. These are also the sectors with the largest growth in U.S. employment.

Further, there is evidence that even higher skilled functions can be moved offshore or might evolve on their own. For example, interviews with people at a firm earlier this year revealed that they had initially been contracted by an American firm to call its clients with overdue credit card payments. The offshore company eventually purchased the receivables from its client and assumed the collection risk itself. Another firm, Wipro Spectramind, managed the radiology services of Massachusetts General Hospital for its second and third shifts. Thus, American radiologists, who earn an average of $315,000 a year were replaced by Indian radiologists, who earn $20,000 a year on average.

I understand that there is a whole subculture in Pakistan and India of people who go to work in the late afternoon or evening and then work a full day on U.S. time. What effect has outsourcing had on the cultures of the countries that are recipients of much of the outsourced work? Have labor rates dramatically increased? Is it difficult for local companies in India and Pakistan to get quality IT talent?

Indeed such a subculture now exists. It is viewed as very stressful work and not suitable for a long-term career. Companies that do such work try to ameliorate the stress by hiring psychiatric counselors to provide free counseling to stressed-out employees. They also provide free meals and transportation, sports facilities, etc.

However, labor rates have increased only, a little. This is more than offset by

the rise in productivity of this labor over time.

Outsourcing is clearly a temporary solution. As labor rates equalize, the benefit of outsourcing decreases. In Pakistan and China, there are still huge differences in labor costs, but in Turkey, rates are closer to what they are in the 11.5. and other Western states. Realistically, how long can we gain a significant benefit form outsourcing?

India and China, and to some extent, Pakistan, have large labor pools. That is why, in manufacturing, Chinese wage rates have not changed despite massive employment growth over the past three decades. I think that wage rates in India will actually fall because of increasing supply, which is being drawn into outsourcing. This would mean several more decades of benefit from outsourcing.

One way in which developed countries may retain value is if their firms control the work done, either through providing the risk capital or through subsidiaries. While it is difficult to predict which organizational types will dominate, a number of firm-specific factors that influence the liability of off-shoring and organization structure are summarized here:

The knowledge component of the activity. A higher knowledge component makes the firm more concerned about whether the quality of the service will change due to a location change or the transfer process.

The interactive components of the process.

The ability to modularize the process

Savings from concentrating an activity in one location, leading to

benefits of scale and scope.

Reengineering as part of the transfer process. To transfer a business process, it is necessary to study it intensively and script the transfer. In the process of study, often there will he aspects of the current methodology for discharging the process that do not add value. Very often these aspects are legacies of earlier methodologies that were not eliminated as the production process evolved. During the act of transfer these are easier to abandon than at an existing facility' where they have become a "natural' part of the daily routine. Our interviews identified other unexpected benefits that go beyond the efficiency effects. Simply examining the business processes may reveal previously undetected inefficiencies. During the transfer process, these inefficiencies can be addressed without disrupting work patterns. Workers in the new location then use the reengineered process which is usually more efficient.

The time-sensitive nature of the work.