Harnessing Science and Technology for America's Economic Future: National and Regional Priorities (1999)

Richard S. Rosenbloom

Professor, Harvard Business School

We meet today in the context of an ascendant American economy. A decade ago, it seemed that key American industries such as automobiles and semiconductors faced mortal threats from Japan Inc. and the Asian Tigers. Recent developments place the so-called Asian miracles in a different light. It is fair to ask, however, just how enduring the American economic success will turn out to be. Are its underpinnings robust enough to sustain its buoyancy for a decade or more?

This era of prosperity stems from multiple sources—prominent among them are sagacious monetary policies, prudent management of the federal budget, and the strong entrepreneurial culture in American society. Also sure to be on any list of causes is the dynamic American capacity for technological innovation. Central to the concerns of today's workshop are the questions: How robust is the current blossoming of innovation? Can it be sustained throughout the next decade? And, to the extent that it is based on new technology, where will the technology of the future come from?³¹

America's innovative successes in the 1990s are concentrated in the science-based industries, and especially in two industrial sectors on the leading leges both of technological opportunity and market growth. The first, broadly speak-

ing, is information technology: semiconductors, computers and software, and communications equipment and services. The other is the complex of industries feeding new technology into health care, mainly biotechnology, pharmaceuticals, and medical instruments. One index of the opportunities—technological and commercial—inherent in these fields is the choices that private firms have made about where to put their R&D money (Table R-1). Among the 50 firms with the largest R&D budgets—firms spending from $300 million to several billion dollars each year on R&D (in aggregate $55.4 billion, more than half of the total for all U.S. industry)—all 20 of the most intensive investors in R&D (i.e., those with the greatest ratio of R&D to sales) are in one of these two fields. Each of these firms spends more than 8 percent of sales on new technology ($18 billion in aggregate). In these two industrial sectors the United States has demonstrated distinctive, and superior, capabilities which have been translated into growth and competitive advantage on a global scale.

It is appropriate that discussion in this hall should focus on science and technology, but we should acknowledge, up front, that productive R&D, by itself, is not enough; new capabilities must change commercial practice before they

create economic value. As the distinguished medieval historian, Lynn White, Jr., once observed, "New technology opens a door ... it does not command one to enter" (White, 1966). What matters for the economy is innovation—the introduction of technology to commercial practice (Smith and Alexander, 1988). How are the results of R&D translated into economic progress?

While it is well established that there is a positive association between economic growth and a nation's abilities in science and technology (Boskin and Lau, 1992) we do not have any clear model of the connections that produce that result. Vannevar Bush shaped public policy and private practice for more than a generation with the notion of a direct relationship—the famous linear model—but we've known for some time that that is an oversimplification. The reality is complex and dominant patterns vary from field to field. The central point is that the process of innovation is at the heart of the linkages between the emergence of new technology and the realization of its economic potential.

What do we mean by innovation? A useful, plain-English definition is: "the processes by which firms master and get into practice product designs and manufacturing processes that are new to them" (Nelson, 1993). We'll take "product designs" to embrace service systems also. This definition embraces a highly diverse set of related but quite different kinds of changes introduced by businesses, ranging along a spectrum stretching from imitation of proven practice to the risky introduction of highly novel technology in radically new applications (Figure R-1).

Let's consider some illustrative examples. Most "innovations," while new to the firm in which they occur, are actually imitative of practices already proven elsewhere. Many of these individually are of small consequence, differing only incrementally from prior practice in the firm. But some imitative innovations, like Microsoft's Windows operating system, can be radically different from their predecessors and produce major economic consequences.

In those cases where the innovation is not imitative, that is, the sponsoring firm is the first to commercialize the innovative practice, the change is usually incremental in character, representing progress along an established technological trajectory. While they may be small individually, incremental changes may cumulate to produce major effects. For example, for the first 60 years following the commercial introduction of insulin as a therapeutic product, technical advance focused on continual improvement in the quality and cost of product made from the pancreases of pigs. Along this trajectory, impurities were reduced to one part per million (ppm) by 1980 from 50,000 ppm in 1930. Incremental change is the principal mechanism behind the phenomenon identified by "Moore's law," which has had such a huge effect on the information industries.

Sometimes, however, novel technology breaks with established trends. For example, in 1978 a new pathway for insulin technology was opened by the introduction of genetically engineered human insulin, a discontinuity that transformed the global insulin industry (Enright, 1989). In other cases, radical change

Figure R-1

A Spectrum of Innovation

creates new categories of products or services, as for example, resulted from Magnetic Resonance Imaging in medicine, or the inkjet desktop printer for computers. Technical change is not a necessary ingredient—important discontinuities can flow from the creative combination of available technologies—the FedEx system, or bank automated teller machines come to mind.

A key comparative strength of the U.S. National System of Innovation—to borrow a phrase from my colleague Richard Nelson—has been its ability to initiate and rapidly to exploit those innovative discontinuities that stimulate economic growth by transforming tired industries or giving birth to entirely new ones. Even when the first appearance of the core innovation is overseas, as for example the computed tomography (CT) scanner, American industry has been able to capture commercial leadership fairly rapidly in many instances.

It would be nice if we could measure the relative importance to economic growth of the different varieties. That sort of analysis is still beyond our grasp. But a few points can be made.

First, sustained economic growth probably requires balanced national capabilities to initiate innovations effectively across the spectrum described above. Imitation of new technology by firms with superior economic capability can be

very productive.³² So can radical change which effectively combines proven technologies.³³

But because our focus today is on science and technology, I will limit my remaining remarks to those innovations in which novel technologies are first put into use. Furthermore, I will refer primarily to the two industrial sectors of greatest current salience, namely, informationand biotechnologies.

One indicator of the character of new technology can be found in the stream of patents granted. While not all innovative technology is patented, and those that are may not necessarily be representative of the whole, patents provide some interesting clues to what's happening on the frontiers of technical advance. Francis Narin's analysis of some 400,000 U.S. patents issued to inventors from all over the world in the mid-1980s and mid-1990s shows a dramatic increase in the extent to which they are based on recent science (i.e., they cite articles published within the preceding decade) (Narin et al., 1997). See Figure R-2 and Figure R-3. Furthermore, while this trend is evident in every industrial country, it is most pronounced in American inventions, and in medical and chemical fields.

The patent data also show that the scientific information originates in many different institutions—75 percent of the scientific references cited by private industry as a basis for new technology comes from universities and government agencies. This is even more pronounced in drugs and medicine, and only about 50 percent for electrical components, but still substantial there as well.

At the core of the national system of innovation is a relatively small set of institutions in government, industry, and universities that produce the scientific and technological developments in the cutting-edge fields. Basic and applied research are creating new options for industrial innovators, not only by creating wholly new opportunities, but also by strengthening the knowledge base that supports incremental progress along established trajectories. Because the nature of industry-university-government relationships is often idiosyncratic to particular fields and industries, no single model provides a useful description across the board (Powell and Owen-Smith, 1998).

It is also clear that we are seeing a reshaping of the relationships among institutions generating new technology. Gibbons and his associates have characterized a new mode of research that spans disciplines, is more commonly organized through networks than through collegial hierarchies, and is characterized by rapid, often non-linear development (Gibbons et al., 1994, p. 20). And we must remember that the working of these wellsprings of knowledge takes place in

the context of a number of larger social forces. What we can say, across the board, however, is that the nature of the relationships and the performance of the institutions has been changing throughout the 1990s.

Within industry, attitudes and practices toward fundamental research and pioneering investments in technology have been transformed (Rosenbloom and Spencer, 1996). The end of the Cold War is reshaping federal spending for science and technology. Universities increasingly are called upon to serve, in the words of one National Science Foundation official as ''creators and retailers of intellectual property'' (Chubin, 1994: quote at p. 125). Despite these changes, the institutional relationships seem well suited to sustain technical advance along established trajectories—to fuel incremental change. But in some respects current trends raise questions about the sustainability of our ability to generate and use radically different technologies. Progress in integrated circuits within the framework of the industry's established road-maps can be achieved, but who will invest in creating the science base for whatever lies beyond that?

Let's start with the private sector. Corporate laboratories dedicated to pioneering in science and technology emerged on the scene at the start of the twentieth century. The institution blossomed most fully in the United States following World War II when numerous corporate laboratories dedicated to fundamental science and long-term development of pioneering technologies emerged in American industry. In later decades a small number of these laboratories were a fertile source of fundamental technologies sparking significant economic growth. Corporate research laboratories flourished most in organizations like AT&T, IBM, and Kodak, whose dominant market positions cushioned budgets from the pressures of narrow margins and facilitated the fullest appropriation of profits ensuing from new technologies. As deregulation and the rise of global competition have forced greater corporate attention to the bottom line, and the richest areas of technological opportunity have shifted to new fields, research budgets in those firms have come under closer scrutiny.

The changes have been most prominent among those firms most notable for their prior accomplishments in creating new technologies. For example:

• Overall employment at IBM's research division was cut by nearly 20 percent in 1993. An atmosphere once characterized as "IBM University" vanished.
• The David Sarnoff Research Center, under RCA ownership a pioneer in electronics technology (inventor of liquid crystals, for example) has become a contract research organization dependent on government funds for its long-term research.

In aggregate, industrial research activity declined by 20 percent in real terms in the early 1990s (Figure R-4).

Of course, there are also counter-currents. Some seasoned companies have continued to support pioneering research. Hewlett-Packard, for example, has intensified its research commitment in the 1990s. New actors on the technology frontiers are beginning to play a role. Prominent among them is Microsoft, which recently established a substantial research organization focused on technologies likely to be significant 5 to 10 years in the future. But many of the new breed of high-technology firms in electronics and information technologies have eschewed traditional research organizations and chosen other strategies. U.S. leaders in the semiconductor industry, including Intel, Motorola, and Texas Instruments, now cooperate to fund research in universities and to develop pre-competitive manufacturing technologies, but none supports a significant central research establishment dedicated to fundamental research on the scale and of the type formerly found at IBM, RCA, and AT&T.

There are multiple forces at work shaping these changes. The end of the Cold War is reshaping the allocation of federal resources for science and advanced technology with undoubted consequences for the laboratories in the private sector. The new competitive environment causes some to question the benefits of private investment in fundamental research. Firms now compete in the global marketplace with rivals that do no fundamental research but are quick to exploit developments made elsewhere.

Where will the technology come from in the next decade? Continued strong progress along established pathways of incremental innovation seems highly probable across the board. Much less certain is whether the nation is investing sufficiently in pioneering research that will establish the foundations of new technologies and new markets in the future.

The direction and intensity of research efforts aimed at novel innovations are shaped by perceptions of opportunity in both the pertinent field of science or technology and in the marketplace. Breakthrough innovations are likely to involve the interplay of universities, government agencies, and firms, but these actors have quite different abilities to discern opportunity in those domains, as well as quite different incentive structures.

Important innovations increasingly have been characterized by being primarily science-based and by requiring a high degree of institutional interaction and flexibility. The full story of major contemporary innovations often displays a complex intermingling of government and university scientific strength, small firm flexibility and initiative, and large firm engineering and marketing capabilities. These phenomena are often most pronounced in the two sectors—information and biotechnology—identified as offering the greatest potential for innovation. In both sectors we have seen a high degree of individual and organizational adaptability and mobility, as evidenced by the high rates of new firm formation and growth in those fields in the United States.

But the pattern of institutional arrangements and the processes of innovation differ significantly between these two sectors. In biotechnology, the boundaries between universities and firms are crumbling, cutting-edge research is now performed by intellectually and institutionally heterogeneous groups, and researchers at for-profit companies play a key role in basic research (Galambos with Sewall [1995] illustrate this in the history of Merck). One thoughtful analysis concludes that "integration between biotech firms and universities is so pronounced that they constitute a common technological community" (Powell and Owen-Smith, 1998:258).

The complex partnerships in biopharmaceuticals seem well suited to sustain continued technological progress and economic growth. In information technology, the role of industry predominates—especially in the hardware underpinnings of technical advance. That works well along established trajectories (Wintel)—but is enough investment going into the search for new technological paradigms? And if so, is that happening in institutions well coupled to the innovative capabilities of the industries that must utilize the results?

Boskin, Michael J. and Lawrence J. Lau. 1992. Capital, Technology, and Economic Growth, in Nathan Rosenberg, Ralph Landau, and David C. Mowery, eds. Technology and the Wealth of Nations. Stanford, Calif.: Stanford University Press.

Chubin, Darryl E. 1994. How Large an R&D Enterprise? in David H. Guston and Kenneth Kenniston, eds., The Fragile Contract: University Science and the Federal Government. Cambridge: MIT Press.

Committee on Science, Engineering, and Public Policy (COSEPUP). 1993. Science, Technology, and the Federal Government: National Goals for a New Era. Washington: National Academy Press.

Enright, Michael. 1989. Novo Industry. Case Study 9-389-148. Boston: HBS Publishing Corporation.

Galambos, Louis with Jane Sewall. 1998. Networks of Innovation: Vaccine development at Merck, Sharpe & Dohme, and Mulford, 1895-1995. New York : Cambridge University Press.

Gibbons, Michael, Camille Limoges, Helga Novotny, Simon Schwartzman, Peter Scott, and Martin Trow. 1994. The New Production of Knowledge. London: Sage

Narin, Francis, Kimberly S. Hamilton, and Dominic Olivastro. 1997. The Increasing Linkage Between U.S. Technology and Public Science. Research Policy 26: 317-330.

Nelson, Richard R. 1993. Technical Innovation and National Systems. in National Innovation Systems: A Comparative Analysis. New York: Oxford University Press.

Powell, Walter W. and Jason Owen-Smith 1998. Universities and the market for intellectual property in the life sciences. Journal Of Policy Analysis And Management 17: (2) 253-277

Rosenbloom, Richard S. and William J. Spencer 1996. "The Transformation of Industrial Research," Issues in Science and Technology. 12(3):68-74.

Smith, Douglas K. and Robert C. Alexander. 1988. Fumbling the Future: How Xerox Invented Then Ignored the Personal Computer. New York: William Morrow.

White, Lynn Townsend, Jr. 1966. Medieval Technology and Social Change. Oxford: Clarendon Press.