The Offshoring of Engineering: Facts, Unknowns, and Potential Implications (2008)

Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Offshoring in the U.S. Telecommunications Industry Theodore S. Rappaport I have been asked to discuss network systems, another field. Some of these data are anecdotal, and some are com- way of describing the telecommunications industry. From piled from public records. other speakers at this workshop, you can get a sense of the rapid growth of global telecommunications markets and the EFFECTS OF THE DOT-COM CRASH massive adoption of wireless and telecommunications tech- nologies—with surprisingly little job growth, and sometimes The fallout from the dot-com crash of 2001–2003 persists, job losses, in the United States. which has greatly impacted the behavior of large telecommu- I have spent the last 18 years as a professor in the tele- nications companies, and, I contend, has affected, in turn, the communications field and an entrepreneur who has started behavior of students entering engineering undergraduate and and sold two businesses to publicly traded companies. I just graduate schools. Compared to other countries—particularly returned from a wonderful stint in industry, at Motorola, and nations that are emerging rapidly, either through emerging am pleased to be back in academia. Motorola is headed by markets or through technological innovations in selected a pioneering chairman and world leader in telecommunica- technologies that have become national initiatives—the tions. In his keynote address, Bob exhorted us to remain United States has “lost” its way. hopeful, and I agree. Look at Korea, for example, the most wired country in The massive bubble in telecommunications of the late the world. Korea had a focused, well funded government 1990s, followed by the telecom crash in 2001–2003, were initiative involving universities, major corporations, and car- the most dramatic events I have seen in my professional riers of Korea to bring fiber to the home. As a result, more career. I do not think I can overstate just how dramatic, and than 30 megabits per second of data are available in more devastating, the dot-com implosion has been for the telecom- than 95 percent of Korean homes, and new applications and munications research community in the United States. services are a major part of Korea’s technology future. Of I will show data indicating that the crash of the telecom course, Korea is a compact country where laying fiber is industry, combined with a lack of public policy to bring much more affordable than in the United States. But this together industry and academic institutions, have created a effort was supported by a big push from government, which crisis in the United States that must be addressed if our coun- helped bring together corporate and university leaders to try wishes to maintain its technical superiority and product create new technologies. development/job creation capabilities in the communications That push continues with wireless technology. In fact, Korea has stated, as a national policy, that the country wants to be the exporter of the next-great revolution in telecom,   Theodore S. Rappaport is the William and Bettye Nowlin Chair in Engi- which they believe will be broadband and wireless. Today, neering and Founding Director, Wireless Networking and Communications companies like Samsung and carriers like KT Freetel and SK Group, University of Texas at Austin. Telecom, with government support and subsidies, are build- 213 

214 THE OFFSHORING OF ENGINEERING ing future technologies, which they are exporting already and 85 percent world market share). COST, which has brought will be exporting more aggressively in the future. Korean industry and academic researchers together for the last 20 companies are encouraged to send their brightest researchers years, provides strong government funding and matching to U.S. graduate programs, usually subsidized at least partly corporate funding, which leads to a great deal of cooperation by government. between industry and academia. Research expenditures are The Chinese government has provided significant focus in the hundreds of millions of dollars per year. and funding for the broadband build-out in the rapidly grow- The United States has traditionally relied on the Defense ing Chinese economy through investment and subsidies for Advanced Research Projects Agency (DARPA), the National capital infrastructure projects, start-up ventures, wireless Science Foundation (NSF), and the major telecommunica- spectrum licensing, and technology standards. China is most tions laboratories to fund research in the United States. But likely to insist that its billion-plus population have wireless since the mid-1990s, when DARPA adopted a “problem of technology standards based on Chinese intellectual property. the week” mentality, the agency has moved quickly away New roads, built in very remote parts of China, are being from basic research. Compare that to the late 1980s, when installed with massive fiber-optic capacity, even where the DARPA program managers were instructed to fund the best nearest community is dozens of miles away. The Chinese academic minds in the United States and let them create great government understands that communications connectivity technologies. Today, DARPA requires strict deliverables and will be vital for the developing knowledge-based capabilities quarterly project reviews, as if academic labs were for-profit in rural areas. The government also supports R&D in many contractors. areas of telecommunications and sends tens of thousands of China’s brightest students to attend graduate school in the The Industry-Academia Chasm United States each year. India has a truly fascinating engineering culture. I rec- Although accountability is a good thing, there are ommend Tom Friedman’s book, The World Is Flat (2005), likely to be fewer companies like Broadcom, Atheros, as an excellent read on this subject. One of the interesting SUN, FORE, and others that spun out of universities with things I have learned from talking to businessmen in India DARPA-funded research. DARPA used to empower highly is that the tax structure for outsourcing and for IT and tele- motivated and entrepreneurial faculty. Now it is difficult com companies is so favorable that it’s bad business not to even to find entrepreneurial faculty on U.S. campuses, invest in India. Under the Indian tax code, IT, software, and because industry and academia are separated by a chasm telecom types of companies pay virtually no income taxes. created by the dot-com bust. In fact, just today, I turned on Bloomberg, and Azim Premji, NSF’s proposed funding rates are in the small single dig- the founder of Wipro, was talking about how his tax breaks its, making it very difficult for faculty to obtain funding for will end in 2009 and that for the next decade, there will be a basic research. A “follow-the-herd” mentality has developed, 0.5 percent tax (maybe) on IT companies. and only esoteric, far-removed projects, which are of great The tax structure in India requires that employees pay cerebral academic interest but have little industrial relevance, personal income tax, but companies do not pay corporate are selected for funding by NSF peer reviewers, who are also taxes. This is very, very lucrative for business and another removed from industry’s competitive needs. example of a government focusing on engineering. In In- The chasm between industry and academia in the United dia, government policy is designed to bolster an industry States, which developed in the wake of the dot-com bust, or a capability to increase national competitiveness. The has become worse over time. Another victim has been the professions of “engineer” and “medical doctor” are two of major corporate R&D laboratories that supported applied the highest callings in Indian society. When I ask Indian research and brokered activities between academia and graduate students to describe their culture, they remark how mainstream U.S. corporations. The United States no longer very different it is from the culture of the United States, has Bell Laboratories or Xerox PARC (as they once existed). where engineering is in decline among our own citizens A huge part of the telecom research community has been and where capitation and insurance/drug policies are mak- demolished. ing things difficult for the U.S. medical profession. Not so With the dot-com crash, stock prices dropped by orders of in India where government policies encourage and reward magnitude in the telecommunications field. QUALCOMM people for pursuing careers in these fields and encourage dropped from $200 to $20 a share; Lucent dropped from $80 businesses to establish centers there. to $2 a share. The Telecom Act of 1996 had opened the flood- In the European Union, as many of you may know, the gates of competition, and thus capital, between competitive Cooperative Arrangement with Science and Technology local exchange carriers and incumbents. Huge amounts of (COST), which began in the mid-1980s, has funded telecom money had flooded in, and maybe expectations were too research at much higher levels than anything in the United high. Yes, clearly they were. But the fall of these companies States. COST in Europe created the GSM cellular standard and the layoffs in telecom hit the American psyche harder (the most widely used cellular standard in the world with than anything else I have seen in my professional career. 

OFFSHORING IN THE U.S. TELECOMMUNICATIONS INDUSTRY 215 Renewing U.S. Telecommunications Research, a report minds of American universities and industry and promoted published by the National Research Council in June 2006, interdisciplinary research. documents, in very real terms, some of the issues facing the U.S. telecommunications industry and engineering and STRATEGIES FOR THE FUTURE telecommunications in the United States vis-à-vis the global economy. The study committee for this report, of which I As the study committee of Renewing U.S. Telecommuni- was a member and which included many national leaders, cations Research argued, we need a U.S. initiative, a national had difficulty agreeing on the reasons for the dire situation policy that can repair the industry-academy chasm in the of telecom in the United States. We all agreed we were in United States, particularly to address the huge number of cor- trouble, but we did not agree on how we got there. When we porate research jobs that were lost in the dot-com implosion. got to how we could make up for lost ground, we had trouble The industry-academy chasm will not be closed by market finding a strategy that would enable us to “save” ourselves. forces, which, in reality, encourage telecom companies to One thing we all agreed upon was that there is now a invest in emerging markets outside the United States where tremendous chasm between U.S. industry and U.S. academic government subsidies and billions of potential customers programs. Before the dot-com bust, corporations were active await. If the United States hopes to invent the next Internet on campuses, funding research and scholarships and invest- or remain a leader in telecommunications, if only to meet ing in a dialogue. After the dot-com bust, telecom companies our own security needs, we must have a national policy that went into hiding. They were (and many still are) in survival reverses the decline of U.S. citizens in graduate programs mode. The companies that did survive saw their market capi- and ensures that they can pursue careers and research in talization decrease by factors of 10 or even 100. According communications. to some sources, more than a million people in the telecom Anecdotal evidence based on conversations with depart- engineering sector lost their jobs. One million engineering ment heads, colleagues, and others involved with electrical jobs in one sector is a lot of jobs to lose. and computer engineering departments at various universi- Imagine the dinner-table conversations of grade-school ties reveals these changes dramatically. When you and I were children or high-school children whose moms or dads were in engineering school, a large majority of the students were laid off from Lucent or MCI or WorldCom or Global Cross- U.S. citizens. Today, at the University of Texas, Purdue Uni- ing. Imagine them seeing their parents lose their jobs and versity, and the University of Florida—three schools I picked their parents’ employers going bankrupt or their companies at random—two-thirds to three-fourths of the undergraduate being sold. students are U.S. citizens, and one-quarter to one-third are The dot-com bust dealt a severe psychological blow to from other countries. the telecommunications engineering profession. I have seen In graduate programs today, U.S. students account for 12 this first-hand in students who enroll in undergraduate pro- to 15 percent and are greatly outnumbered. In some schools, grams at the University of Texas. Something must be done the number is as low as 8 percent; in others, it’s, perhaps, to change the situation if we are to have a reservoir of future 20 percent. In short, U.S. students in graduate programs in technical experts who are U.S. citizens in the communica- electrical and computer engineering are a small minority. tions field. In undergraduate programs, they are still a majority, but Since 2002–2003 when U.S. companies went into survival just barely. mode, they have been trying to make quarter-by-quarter re- This isn’t necessarily a bad thing, except that recently I sults for Wall Street. At the same time, U.S. funding agencies have noticed that my own students, and other students I talk have moved farther from the needs of the wounded telecom to, want to go home after they graduate. They see huge eco- industry. In the 1980s, DARPA funded entrepreneurial nomic opportunities in their homelands where multinational faculty and students who created technologies that funda- companies are investing. They see that the U.S. telecom mentally changed the telecom landscape. DARPA’s funding industry is still in “hunker-down” survival mode and invest- supported advances in fiber optics, the Internet, even cellular ing its precious capital in the emerging markets where future technology. However, in the last 10 years, and especially in customers will come from. the last few years, DARPA has moved away from supporting So we have a problem. On the supply side, we have a long-term research on academic campuses. In fact, it is now problem attracting U.S. citizens at the undergraduate level. hard to find students who can qualify for DARPA funding, We must figure out how to attract students whose parents, and because recipients must be U.S. citizens, and they are becom- others in their parents’ generation, lost their jobs. If we don’t ing exceedingly rare in U.S. graduate programs. attract them they will feel more and more like strangers in Although NSF continues to do its part, and is a sponsor their own land, as they become a smaller and smaller minor- of the present study, it has not been able to fund relevant, ity in engineering. As for demand, it is clearly in emerging long-term, industry-captivating R&D with its small grants. markets and not in the United States. And telecom has not been included in NSF’s bold engineer- Another aspect of the slippage in the U.S. position is a ing research center (ERC) program, which has captivated the precipitous drop in the number of conference papers au- 

216 THE OFFSHORING OF ENGINEERING thored by researchers from industry. In fact, industry is now are not limited to telecommunications, but also make refrig- all but missing in action at the IEEE Global Communications erators and power turbines. I included Microsoft as a proxy Conference (Globecom) and other international communi- for software companies. A few companies from China, cations conferences. In 2005, of all the companies in the Huawei, UTStarcom, and ZTE, that have emerged rapidly on world, only eight published more than one conference paper. the telecommunications scene are also on the list. Although In the mid-1980s, Globecom was dominated by presenters their annual revenues are only $3 billion to $6 billion, they from Bell Labs, Xerox PARC, IBM, and Motorola. Today, have grown rapidly in the last few years. however, U.S. industry participation has dropped by more The five major focus areas of R&D, based on all of the than an order of magnitude to 1 in 12 papers presented at company press releases we collected, will give you a sense international conferences. Industry is no longer a participant of the needs being met by the $0.5 trillion of products being in the research dialogue. In effect, professors now present sold around the globe each year: papers to other professors. For the purposes of this talk, I collected some data on how • Subscriber devices or premise equipment—cell phones, multinational companies view funding for R&D. I asked one videophones, voice over IP—purchased by end users is of my research assistants to look up financial records and a major component of telecommunications revenues. press releases at the largest telecommunications companies • Infrastructure equipment and services are the base for the last four years. She scoured thousands upon thousands stations and switching stations, the large infrastructure of Web announcements by large multinational companies that connects carrier-grade telephone, Internet, and to find out where they are making R&D investments. We wireless systems. then correlated the results for location and types of R&D • Switching and routing equipment is the technology/ investments being made and estimated the expenditures. equipment used to connect large hubs with an enter- The results, not surprising, perhaps, underscore that major prise or premise. corporations are not investing in the United States but are • Integrated circuits are essential to the chips in all turning instead to the most promising emerging markets. equipment. Figure 1 shows the companies we studied and their an- • Software and applications, a more and more vital part nual sales for their most recent fiscal years (in U.S. dollars). of the value chain, create the features, adaptability, and These companies, which are household names, have the upgradability of all products. bulk of the worldwide market capital of telecommunications companies—network systems, software, and so on. You The corporate research themes of global telecom com- could say that Texas Instruments should be on the list (Intel panies (listed below) can scarcely be found among research is there as a proxy) or other companies. But keep in mind initiatives at NSF or DARPA: this is just a sample of major companies that have presence around the world. • the development of Internet protocol (IP) to replace As Figure 1 shows, the annual revenues of these 16 com- circuit switching (IP has a double meaning; Internet panies is about $0.5 trillion—a lot of infrastructure, software, protocol and intellectual property [a creator/extractor and handsets. Some companies, like Samsung and Siemens, of value], a key theme in corporate America) • Alcatel: $15 billion • Motorola: $37 billion • Cisco: $28 billion • NEC: $46 billion • Ericsson: $19 billion • Nokia: $40 billion • Huawei: $6 billion • Nortel: $10 billion • Intel: $39 billion • Samsung: $79 billion • LG: $23 billion • Siemens: $91 billion • Lucent: $10 billion • UTStarcom: $3 billion • Microsoft: $44 billion • ZTE: $3 billion FIGURE 1  Companies and annual revenues (USD). Source: Compiled from data on various company web sites. 

OFFSHORING IN THE U.S. TELECOMMUNICATIONS INDUSTRY 217 • the convergence of wired and wireless networks has been active in China, Brazil, and Denmark and, in 2005, throughout an enterprise and throughout the home made a major R&D investment in the United States. • the expansion of intelligence and massive bandwidths NEC and Nokia both opened R&D centers in China. Nor- to the edge of the Internet and into the home or tel is investing in China, France, and India. Samsung made enterprise major R&D investments in China and Korea. Siemens made • the development of multimedia data transfer for mul- an R&D investment in Korea and is partnering with Nokia. tiple providers (cable, telephone, wireless converging, UTStarcom, a Chinese company, is investing heavily in competing, and offering content) India for wireless infrastructure and for IPv6. ZTE, another • the development of low-cost devices and low-cost in- Chinese company, is investing in its home country. frastructure for emerging economies that have different In summary, major telecommunications companies an- price points and applications than in the mature U.S. nounced 57 major R&D initiatives in the past few years. market Of these, only five were in the United States. Thirty-five, • the ongoing development of software and middleware the overwhelming majority, were in Asian countries, where and increasing reusability public policy and regulations are much more welcoming and where markets are experiencing higher growth rates. Twelve The absence of these research themes in U.S. academic R&D major R&D investments were made in Europe, more than shows the complete disconnect between what companies see twice as many as in the United States. in their future (e.g., the corporate vision or road map) and what professors and graduate students in the United States OBSERVATIONS are working on. I fear that professors are becoming increas- ingly isolated from “customers.” What can we learn from these data? First, R&D invest- ments are going to high-growth countries. Second, multi- national companies are investing in countries that have made R&D Initiatives by Major Companies telecom a priority, either through tax incentives, research- The discussion that follows briefly covers some of the expenditure incentives, or other government policies. Busi- ­ ajor R&D initiatives of the corporations listed in Figure 1. m nesses are going where there is less friction, because it makes In 2002, Alcatel made R&D investments in Canada, ­Australia, good business sense. and China; in 2003, the company invested in R&D initiatives We now see many foreign students who come to the in Australia, France, and China. In fact, China appears on United States to get educated in telecom return to their home every company’s R&D investment list. Alcatel did not invest countries. For example, of the Ph.D.s who graduate from our in the United States until 2007, when it purchased Lucent. wireless center at the University of Texas, about half take Only a small number of R&D projects by Cisco have been jobs back in China, Korea, Pakistan, and India—something publicly announced. Cisco uses an open IETF model in de- I haven’t seen in my career before. veloping much of its technology, but the company has opened That trend is not a bad thing in itself, but it points to the centers in Japan, India, and Vietnam. Instead of funding basic fact that the United States has lost its national focus on tele- R&D, Cisco typically buys 10 to 15 small companies per com, which is clearly not on the research agenda of the U.S. year. Ericsson, like most telecom companies, struggled for government or U.S. industry. When I look 10 to 15 years survival during the dot-com bust. Since then, it has managed ahead, I am deeply concerned about what might happen in to open an R&D facility in China. the United States. As the data clearly show, U.S. companies, Huawei has made investments in China (its home coun- rather than investing in the United States, are investing over- try), India, and Malaysia. The Malaysian government offers seas and hiring overseas, for R&D positions. huge incentives to companies to locate jobs there, especially Under these conditions, I wonder if, today, we could in- in telecom and manufacturing. For example, the government vent the Internet or cell phone technology. DARPA funded often pays the salaries of the first 100 or so engineers to help the Internet. NSF took over the build-out of the Internet a company establish a beachhead. to campuses across the country. This required a very long Intel has not publicly announced any investments in R&D investment period, 10 years of funding, before there was centers in the United States. The company has invested in even a hope of creating an Internet. Yet that stay-the-course, China, Spain, and England. LG, a leading Korean company, long-term research was a national policy by a government established a U.S. facility in 2005, but its primary focus is in determined to develop a failsafe communications network China, Korea (its home country), Italy, and France. Micro- that could survive and operate in a national emergency. soft has announced one major U.S. investment, R&D initia- In the aftermath of the dot-com crash, we must rethink tives in England, and a major investment in India. Motorola how U.S. corporations should engage with the U.S. govern- ment and with academia. We must ask ourselves how long   For details see http://users.ece.utexas.edu/~wireless/NAE%20Research. the United States can continue to produce talent at home to htm. build secure networks, defense networks, without the old 

218 THE OFFSHORING OF ENGINEERING DARPA or Bell Labs. And how we will be able to oper- lic policy needs to help pick future technologies. “Road ate secure networks. Perhaps most important, how will the mapping” is probably too strong a term in the case of United States be able to compete globally in telecom against telecommunications, but as you can read in Renewing U.S. countries that are making concerted efforts to develop tech- Telecommunications Research, if we don’t do something, nology for export? Do we need a government policy to prop we are going to continue to lose our lead. As a successful up the telecom industry, just as SEMATECH came to the businessman once told me, if you are not going up, you are rescue of the semiconductor industry two decades ago? going down. Unless the United States moves forward with a The United States does not like to pick winners or los- vision in telecommunications, we are in for a tough time. ers, but considering the dot-com situation and the lack of Now that Bell Labs is gone and DARPA is no longer investment in U.S. R&D, combined with the chasm between doing what it used to, we need a national policy or entity academia and industry, I submit this is one time when the that can bring us together, in telecom. As we just saw, IEEE government should reach in to encourage and assist industry conferences aren’t doing that. Universities are there, but in reinvesting and engaging with U.S. universities to try to industry is absent on campuses. Industry is doing a lot of repair the damage done by market forces. If we care about internal research, but the results lead to filing for patents U.S. citizens having expertise and research acumen in the instead of publishing papers. Microsoft and Motorola have telecommunications field, something must be done to en- huge research organizations, both internal. These companies courage them and to give them a reason to put forth the effort used to engage with the academic community. My concern to attain the necessary skills. U.S. taxpayers should also be is that if we don’t bring industry and academia together, made aware of the falling enrollments of U.S. citizens on our relevance on the global telecom stage will be greatly engineering campuses throughout the United States. diminished. TURNING THE TIDE CONCLUSION In this section I suggest some steps that could be taken I will close with this final thought. Let me ask everyone in to turn the tide for telecommunications R&D in the United the room to please raise your hand if you grew up somewhere States. First, we need a federal policy that encourages and other than the United States. Raise your hand proudly. So, rewards U.S. industry for engaging in a new social contract maybe 10 or 12 percent of the people in this room were born like the one the telecom community had for decades in the outside the United States. How many of you grew up in the United States. That contract meant that companies were United States and went to college in the United States? The actively involved on campuses, and they provided scholar- overwhelming majority, 90 percent. ships. Reviving this social contract may require government Consider that the data I have presented show that in gradu- engagement. ate programs in telecom throughout the United States today, Before the dot-com crash, when I headed the wireless the numbers are diametrically opposed to the numbers in center at Virginia Tech, more than 30 major companies this room. On college campuses today, most of the graduate came to Blacksburg, Virginia, to invest in our research center students in telecommunications and electrical engineering and engage with our students. After the dot-com crash, the did not grow up in the United States; they did not grow up number fell to about 10, and two of those were from China in the culture of the United States. This presents us with a and Korea. That is a real drop-off! If we can’t bring more great opportunity, but also a great challenge. companies back to academia, professors will be working If we want this room to be filled 20 years from now when on problems that are peer-reviewed by other professors in a the United States faces a new crisis, we must make sure that vacuum, without regard to market needs and without long- the students in graduate schools today stay in the United term marketability for the benefit of the U.S. telecommunica- States, can find gainful employment in the United States, and tions industry. have the necessary support for the United States to remain a Second, we need a public-private “big picture,” a big, leader in telecommunications innovation. If we don’t do that, hairy, audacious goal (BHAG), that will lead to hope for the all the creative minds will leave the United States and go to future. Bob Galvin gave us some examples of big-picture other countries, perhaps their home countries, where they projects in his keynote address—new architecture, wire- have the opportunities and incentives to make an impact. less Internet, middleware. The United States has so many opportunities to take the lead, to solve a national problem REFERENCES through technology. But we are not doing that. We need a “man-on-the-moon” kind of mission to rally our industry and Friedman, T.L. 2005. The World Is Flat: A Brief History of the Twenty-First Century. New York: Farrar, Straus, and Giroux. engage creative minds in thinking about how we can improve NRC (National Research Council). 2006. Renewing U.S. Telecommunica- our nation. tions Research. Washington, D.C.: The National Academies Press. I agree with Bob Galvin that, in some instances, pub-