| This page in the original is blank. |
Technology Innovation and American Leadership
William J. Perry
Former U.S. Secretary of Defense Professor, Stanford University Stanford, California
Last year, America's information technology industry earned more than $700 billion—more than the gross domestic product of most nations! The world is truly in the midst of an information technology revolution that is bigger and moving faster than the Industrial Revolution, and the United States is the undisputed leader and major beneficiary of this revolution. As a result, the U.S. economy today is the wonder and envy of the world.
Last year I experienced a powerful example of this economy at work when, after four years in Washington, I returned to Silicon Valley and set out to buy a house. I discovered to my horror that, to buy a house equivalent to the one I had sold four years earlier, I would have to pay 50 percent more. I was thus able once more to practice my time-proven economic principle: "sell low, buy high!"
Earlier this year our local newspaper reported on the economic factors underlying this staggering rise in house prices. During the four years I was gone, Santa Clara County added more than 100,000 new jobs. During the period I was house hunting, jobs were increasing at a rate of 1,000 a week. In the year before I returned, 70 companies in Silicon Valley had their initial public offerings, creating each week four or five new millionaires—each of them wanting to buy a $2 million house! So this is one measure of the extent of technological innovation and its economic impact.
While the entrepreneurs in Silicon Valley were applying their technology to building our nation's wealth, I was back in Washington applying that same technology to building our nation's security. For decades our security strategy has depended on the full use of our technological leadership, especially our leadership in information and aerospace technologies. Today I will talk about the role of technological innovation in two different but complementary tasks: building
our nation's wealth and building our nation's security. Both endeavors depend on America sustaining its world leadership in technology, a condition that many Americans take for granted. But this leadership is not a divine right; indeed, it has not always been true, and it is not certain that it will be true in the future. So it is worth asking: How did we achieve this position of leadership—and how can we sustain it?
During the nineteenth century and the early part of this century, America was an also-ran in the field of science. In the 1930s many of the leading scientists in the world came to the United States to escape the scourge of fascism, and most stayed here after the war. They trained a whole new generation of native-born scientists and engineers, including hundreds of thousands of veterans whose graduate education was provided by the GI bill. Then when the Cold War started, our government, which had provided strong support for technology during World War II, decided to continue that support. So the new engineers and scientists, having just completed their education under the GI bill, were put to work advancing the state of the art in technology.
As a result, the Department of Defense (DOD) developed and was the first to use supercomputers, communications satellites, integrated circuits, high-order software languages, and the Internet (in fact, the predecessor to the Internet was the ARPANet). All of these developments were part of a strategy to maintain technological superiority as a way of offsetting the numerical superiority of Soviet ground forces. Fortunately, we never had to put this offset strategy to a test with the Red Army. But the high-tech weapons systems we developed for that purpose were put to a test in Desert Storm. There we were fighting against a foe equipped with Soviet weapons but in about equal numbers to the allied forces.
American forces were equipped with the new weapons systems developed during the 1970s, which used information technologies to locate enemy targets on the battlefield, embedded computers to guide weapons precisely to those targets, and stealth technology to evade enemy weapons. As a consequence, the allied forces won quickly, decisively, and with remarkably few casualties. That is, when our technological advantage is not needed to offset superior numbers, it can be used to achieve battlefield dominance over a foe with equal numbers.
Our military leaders, having seen the results of battlefield dominance in Desert Storm, decided that they liked it and that they wanted to keep it. And so today our military strategy calls for maintaining battlefield dominance over any regional power with whom we might be engaged in conflict, and to do that through our leadership in technology. This is the same strategy of technological superiority we had during the Cold War but now for a different reason. Not only has the reason changed, but so has the way of applying this technological superiority.
During the 1950s, 1960s, and 1970s, DOD was the principal supporter of research and development for the computer, communications, and semiconductor
industries. In effect, our nation's commercial industry was riding on the shoulders of the DOD. Today that has all changed. The technological explosion in computers, communications, and semiconductors has led to an amazing new set of products for industry, businesses, schools, and homes. Indeed, all of these different users are being tied together today by the World Wide Web in a way few could have predicted a decade ago. Commercial applications of computers are leading military applications, and for computer companies commercial revenues dwarf DOD revenues.
Today DOD is obliged to ride on the shoulders of our commercial information technology industry, and this poses a difficult problem for military planners. Our major weapons systems take 10 to 15 years to develop and then are in the inventory for 20 to 40 years. But the computer technology that most influences their competitive advantage changes every two or three years.
When I became Secretary of Defense, I was determined to do something about this problem. I knew that I needed a systems strategy that would keep major weapons systems in the field for several decades but update them every few years with new information technology. So I approved the creation of a large-scale experimental program to do just that; the Army calls it Force 21—the digitized battlefield. Force 21 holds promise to dramatically improve the way the United States adopts and adapts computer technology to military uses.
The concept behind the experiment was simple. The Army inserted appliques of commercially available digital subsystems into their current weapons systems—tanks, artillery, attack helicopters—thereby giving them a quantum increase in capability. In aggregate these appliques form a "system of systems," an integrated network of powerful computers and high-speed communications—basically, an Internet on the battlefield. This system of systems will transform the way commanders and troops see and communicate on the battlefield.
How does this work in battle? When a tank commander spots enemy forces, he will have a choice: he could engage the enemy with the weapons on his tanks or he could call in nearby attack helicopters, artillery, strike aircraft, or naval gunfire. Because of digital technology and the constant flow of battlefield information to all combatants, these other units will see exactly what the tank commander sees. And any one of them—or any combination of them—will be able to respond with equal precision in attacking the targets. As combat is under way, the supporting logistics unit will be monitoring the ammo usage. So it will be able to conduct resupply at the time and with the amount needed, thereby reducing the huge logistics tail needed to support combat operations. This system of systems is a brilliant application of information technology to achieve battlefield dominance without designing all new weapons platforms. This is the army of the future, and it is not just on viewgraphs. The Army has already outfitted the 4th Division at Fort Hood with this new equipment and is testing it in simulated combat.
This example makes it clear that our national security strategy depends on
U.S. leadership in information technologies. But can we count on our commercial industry to sustain that leadership? To get some insight into this question. consider how the United States achieved world leadership in integrated circuits many decades ago.
After the semiconductor had been discovered, engineers all over the world were rushing to develop applications for this new device. During those heady times, a conference was held of the engineers working on these applications, and a seminal paper was presented at that conference by the British engineer, G. W. A. Dummer. In his paper he said: "With the advent of the transistor (and the work in semiconductors in general), it seems now possible to envisage electronic equipment in a solid block with no layers of insulating, conducting, rectifying, and amplifying materials, the electrical functions being connected directly by cutting out areas of the various layers." Well, you don't have to be an electronics engineer to understand that he was describing what later came to be called the integrated circuit, and he was racing to be its inventor.
But the integrated circuit was not invented by a British engineer named Dummer; it was invented by two American engineers named Noyce and Kilby. At a similar technical seminar held a few years after the integrated circuit was announced, Dummer presented another paper in which he mused about the reasons he and other European engineers came in second. "It is worth remembering," he said, "that the giant American electronics companies were formed since the war by a relatively few enterprising electronics engineers, setting up with either their own capital or risk capital from a bank. Often a government contract would start them off. Hard work was necessary, and the large home market was a great asset, but the climate of innovation was such that any advanced technical product could be sold. Successful businesses are almost always dependent on a few people who are innovative and enthusiastic." That story encapsulates how the United States gained technical leadership first in the semiconductor industry and later in the information technology industry. In essence, Dummer was saying that, although military contracts were helpful, America won primarily because of its innovators.
But the world has changed in many ways since then. In particular, DOD support is much less consequential, and without that support Japan has risen as a world-class competitor in the information technology field and has challenged U.S. leadership. So what is the likelihood that our commercial industry can sustain its world leadership without the major support from DOD that it had during the Cold War? Let me relate "A Tale of Two Countries" because I believe this tale, which is about competitive technology programs in the United States and Japan, illustrates how critically important innovation is to a country's ability to maintain technological leadership.
Just a decade ago the United States was falling behind Japan in its ability to compete in world markets, and all the trends seemed to be negative. The United States had a very high rate of consumption and a low rate of savings, especially
relative to Japan. Japanese auto companies outcompeted American auto companies, even in domestic U.S. markets, with a profound negative effect on our balance of trade. Japanese companies dominated the world market in consumer electronics. And Japan's leading electronics companies were embarked on an intensive and concentrated effort to overtake the lead of American companies in information technology. It appeared that the United States was about to lose one more major market to Japan and this in a field of traditional American market dominance.
How do things look after a decade of this predicted demise of U.S. competitiveness? The U.S. consumption rate is still high and its savings rate is still low, and in my opinion that is not likely to change. Japan's savings rate is even higher than it was, resulting in low consumption, and this is believed to be one of the factors in its current economic slump. That is ironic because a decade ago economists cited Japan's high savings rate as a major driver in its economic growth. U.S. auto companies, by adopting Japanese production techniques and an emphasis on quality, have gained back some of their lost market. While the United States has not gained back any significant part of the market for consumer electronics, Japan has lost market to increasingly difficult competition from the "Tiger" countries. But most importantly, Japan has failed in its bold attempt to seize leadership in the important market for information technologies.
What happened? The Japanese strategy, orchestrated by the Ministry of International Trade and Industry, focused on gaining leadership in three products involving leading-edge technologies: memory chips, fifth-generation (artificial intelligence) computers, and high-density television. They reasoned that these were important products in and of themselves but, more importantly, that leadership in these products would lead to compelling competitive advantages in all other products in the field of information technology. Many industry leaders in the United States agreed with this assessment and became increasingly alarmed as Japanese companies gained an increasingly larger share of the market for memory chips. Today the situation looks very different—indeed, it appears that Japanese companies "bet on the wrong horses." Memory chips have become a commodity product, with increasingly lower margins prevailing after Korea and the other "Tigers" entered the competitive fray. Artificial intelligence applications for computers have not matured, as envisioned a decade ago. And the market for high-density television has been very slow to materialize. In the meantime, information technology companies in the United States have proceeded in very different directions, concentrating their efforts in three areas: microprocessors, which were used in increasingly capable workstations and increasingly versatile desktop computers; telecommunications networks, which created the exploding market for Internet products; and software, which provided a competitive advantage across the board in information technology products. How did these two different countries, each with very capable technologists and managers, come to such different outcomes?
I believe three primary factors affected this outcome: (1) The Japanese strategy was driven to a large degree by its government, while the U.S. strategy was driven by a large number of individual entrepreneurs. When a government puts its support behind a product, it can be a powerful force, but if it guesses wrong the corrective forces found in the marketplace may be rendered ineffective. (2) American entrepreneurs initiating new products or new companies found an abundant supply of risk capital, both from venture capitalists and in the dynamic market for public offerings for high-tech companies. There are no comparable markets for risk capital in Japan. (3) The technological skills required for success in these new markets were in abundant supply at the leading technical universities in the United States. This resulted in new product ideas from university labs and, even more importantly, a fresh wave of scientists and engineers trained in information technology. The training at America's technical universities was relevant and at the cutting edge because of the unique bonding between America's technical universities and its high-tech companies. In sum, the success of American companies stemmed from three great assets: our entrepreneurial spirit, our innovative markets for risk capital, and our great technical universities.
These advantages are quite fundamental and are likely to allow us to sustain our leadership in any product where innovation is the key to success. But if we want to sustain these advantages into the twenty-first century, we must value them and nurture them. I believe that our entrepreneurial spirit will continue to thrive for the indefinite future. It is a basic cultural advantage we have, and we should recognize and cherish it. Similarly, I believe that we will continue to lead the world in our innovative markets for risk capital, but to ensure that we should strive to educate the public and the Congress about the importance that risk capital has to our overall economy. By so doing we should be able to help lawmakers resist the temptation to overregulate these markets, thereby killing the goose that is laying these golden eggs. I am concerned, though, about our ability to maintain the position of world leadership of our great technical universities.
For more than four decades DOD provided the majority of funding for America's technology base, which was underlying all of the remarkable technical products developed during that period. With the end of the Cold War, the DOD budget has decreased, in real terms, by 40 percent; and the funding for the technology base has decreased proportionately. Moreover, DOD's production contracts have decreased about 70 percent, and thus the defense contractor's independent R&D, which is proportional to overall sales, has decreased proportionately. And while there has been some increase in the rest of the federal R&D budget, the increase has not been proportional, and it has mostly been in health-related technologies.
On the positive side, our information technology companies have had dramatic increases in revenues and profits this past decade and have made corre-
sponding increases in their expenditures for R&D. But these expenditures have been almost exclusively for product development and have not served to replenish our technology base. So this is a serious problem for the future and will be solved only by an increase in federal funds for technology-based programs and by industry consortia that pool some of their R&D funds to support R&D base programs at our universities.
I have given some examples here of how technological innovations have played a critical role in our economy and our national security. I would like to conclude by describing how technology is used not just in the making of our weapons systems but in decisionmaking in national security.
Government officials in national security find decisions in this field particularly challenging for three reasons: First, the stakes are so high—the price of a wrong decision can be thousands of lives. Second, the problems are incredibly complex—quantitative decision techniques can generally be applied only to a segment of the problem. Third, the decisionmaker rarely has either sufficient data or sufficient time to make an analysis of all important factors bearing on the decision. This last dilemma was captured brilliantly by C. P. Snow (1962), the British scientist who worked on technically complex defense projects during World War II. In his book Science and Government, he wrote: "One of the most bizarre features of any advanced industrial society in our time is that the cardinal choices have to be made by a handful of men who cannot have a first-hand knowledge of what those choices depend upon or what their results may be .... When I say the 'cardinal choices,' I mean those which determine in the crudest sense whether we live or die."
As Secretary of Defense I was faced every day with cardinal choices—choices that determined whether our soldiers would live or die. And I generally had to make my decision without enough data and without enough time to analyze even the data I had. I understood that I would generally not be able to apply analytical tools to the decisions I had to make. Nevertheless, I found that knowledge of those tools was invaluable. That knowledge allowed me to approach decisionmaking with an objective framework for assessing the validity and relevance of the data I had—and didn't have—and for assessing alternative solutions. This analytical rigor served me well on many of the important decisions that I made.
Let me illustrate some of the important principles of decisionmaking by giving some examples from my tenure as Secretary of Defense. The first principle is that you never have enough data to solve the most important problems analytically, but that does not mean that analytical techniques cannot be used—it means instead that you have to separate the variables in the problem. It may then be possible to apply analytical tools to one or more of the separated parts. For example, the President did not use analytical tools when he decided to deploy U.S. troops to Bosnia in December 1995. But his decision depended on my assuring him of the feasibility and relative safety of the mission, and I based my
assurance to him on some rather detailed analyses. Our simulated war games provided the basis for believing that we could carry out our part of the missions in the Dayton Agreement with 25,000 U.S. troops. Our detailed flow analysis showed that we could move these troops into Bosnia from Germany fast enough that there would not be a period when we would be so weak as to invite attack. And our detailed logistical analysis showed that we could supply these troops from Hungary, where we had been offered a support base. Some of the detailed analyses underlying my assurances to the President required our logisticians to employ rather sophisticated tools from linear programming and queuing theory.
A second important principle is that all important decisions involve uncertainties that cannot be resolved because they involve data not known at the time a decision must be made. Because of the pervasive nature of statistical uncertainty in nearly all of our decisions, our decision tools must be able to deal with random variables.
Let me now give a real example of statistical uncertainty in the defense field that occurred in the 1970s. This was during the height of the Cold War. The United States and the Soviet Union both practiced a doctrine known as mutual assured destruction, aptly nicknamed MAD, which had our two countries locked together in a balance of terror. At one stage we believed that we might be able to break this balance of terror with the deployment of antiballistic missiles (ABMs). But we knew that a deployment would be costly, both in dollars and in the possibility that an ABM system might provoke an attack while it was being built. Therefore, we wanted to be sure that our ABM system could not only shoot down missiles but also that in so doing it would protect the country from the threat we postulated. Calculation of the protection provided by an ABM system involved random variables—some of the relevant data (e.g., the yield of the Soviet warheads) was known with precision but not to us, so we treated warhead yield not as a fixed number but as a probability distribution. Other factors, such as the actual impact point of a missile aimed at a certain target, were unknowable, even to the side firing the missile, except as a probability distribution. Finally, we knew that we were protecting against a large-scale attack directed at diverse targets, but we did not know how specific warheads would be assigned to specific targets; therefore, that also had to be treated as a random variable.
After many false starts and misleading analyses, DOD finally developed the analytical tools that dealt adequately with this complex problem. All relevant intelligence data were assembled about the characteristics of Soviet weapons, the hardness of presumed U.S. targets, and the weapons effects against various targets, as determined by field measurements made on our own weapons. A half dozen of the most significant variables were assigned probability distributions that reflected the best information available about them, including the uncertainty about these data. Then a variety of attack scenarios were planned and run thousands of times on the computer, using Monte Carlo techniques to reflect the probability distributions.
What resulted from this extensive set of calculations was that the results of the missile attack were reflected in a probability distribution that in turn reflected the real uncertainty of the input data. Decisionmakers then were presented with the resulting calculated damage not as a single number but as an average and a deviation about that average. This extensive set of calculations was an important factor in the final decision—namely, not to depend on the deployment of an ABM system to defend the country but instead to negotiate an arms control treaty that limited the deployment of both missiles and ABMs.
This is an example of how a complex operational analysis was used to influence a major decision on a real-life security problem. I do not mean to imply that because such extensive analysis was performed that the answer is right or even that all analysts agreed with the decision—that is far from true. Nor do I mean to suggest that this analysis done on the missile threat in 1970s is relevant to decisions about how to deal with the missile threat in the twenty-first century. Instead, I want to suggest that technology by itself is not enough. It is also important that we make the best decisions about how to use technology to most effectively strengthen our national security.
Let me close with a final quote from C. P. Snow. He wrote: "Technology is a queer thing. It brings you great gifts with one hand, and it stabs you in the back with the other." The time-honored role of engineers is to bring technology to the public in the form of products. The emerging role of the systems engineer is to ensure that the system of systems created by the new technology does not stab us in the back; that technology really does, as we hope, bring us great gifts.
Reference
Snow, C. P. 1962. Science and Government. New York: New American Library.
| This page in the original is blank. |
