Securing America's Industrial Strength (1999)

Through the 1980s a series of studies portrayed the technological leadership and international competitiveness of U.S. manufacturing industries as imperiled and probably on the decline. Only a decade later, trends seem to have been reversed and the prospects for continued strong U.S. economic performance appear bright. Were American industries and firms really doing that poorly and foreign competitors that well a decade ago? Is the apparent reversal an accurate picture and will U.S. technological leadership and competitive strength across a broad range of industries be sustained?

To help answer these questions, the National Research Council's Science, Technology, and Economic Policy (STEP) Board commissioned studies of eleven industries, including so-called ''service'' industries that were overlooked in the competitiveness debate because manufacturing was considered to be the backbone of the economy and more vulnerable. Today, services generate three-quarters of the gross domestic product, employ eighty percent or more of the workforce, and consume much of manufacturing output (e.g., commercial aircraft, medical products, and computers). The industries that the Board examined are steel, chemicals, pharmaceuticals and biotechnology, banking, trucking, food retailing, power metallurgy parts, apparel, computing, semiconductors, and computer disk drives. These studies appear in a companion STEP Board volume, U.S. Industry in 2000: Studies in Competitive Performance , with an introduction by David Mowery, University of California Haas School of Business.

Pessimistic analysts in the 1980s almost certainly mistook adverse macro-

economic trends—in particular, the high valuation of the dollar—for much more fundamental signs of structural deterioration. Nevertheless, the general picture is one of stronger performance in the 1990s on a variety of dimensions, among them investment, export market share, R&D spending, and profitability. Although not universal and not without dislocations to particular firms, groups of workers, and regions, this improvement is true of much of the service sector as well as manufacturing. Foreign competition has been a driver of change in some cases but not in all. Domestic competition, often from new entrants, has played an important role. In trucking, banking, food retailing, and most manufacturing industries, applications of information technology have enabled the introduction of new products and services and recasting of logistics and other processes to be more efficient.

Although precise causal relationships and rankings cannot be determined, in the 1990s the U.S. government followed a supportive mix of macroeconomic and microeconomic policies—deficit reduction, conservative monetary policy, scaling back of economic regulation of transportation, finance, and communications, trade liberalization, relatively permissive antitrust enforcement, and strengthening of intellectual property rights. As it had in previous decades, the federal government continued to support research across a broad range of scientific and engineering fields, although the 1990s saw the beginning of a change in the research portfolio that may not bode well for the future—in particular, a decline in support of several physical science and engineering fields.

Improved U.S. industrial performance also reflects a variety of private sector strategies—repositioning, product specialization, firm consolidation, internationalization of operations, manufacturing process improvement, and cost reduction—that were market driven, not guided by public policies. These benefited some established firms at the expense of others and in many industries opened opportunities for new entrants. As a result, on the eve of 2000 the structure of most industries looks very different than it did even 15 years ago.

Several enduring characteristics of the American political system and economy bode well for the future—the sheer size of the domestic market, encouragement of experimentation, and relatively little protection accorded enterprises resistant to change. Indeed, contrary to recent conventional wisdom about investors' myopia, U.S. capital markets over time do a reasonably good job of favoring firms with high growth prospects.

Nevertheless, in the long run international shifts in comparative advantage are inevitable, as a result of changes in national political, legal, educational, and capital market institutions, wars and other destabilizing events. Today, Americans should be wary of assuming that the 1990s marked an enduring turnaround in U.S. industry performance.

Despite its general satisfaction with the progress of the last decade and guarded optimism about the future, the STEP Board concludes from its investigations that there are four policy concerns that need to be addressed:

• Carefully selected indicators and data collected nationally on a recurring basis are needed to help discern and track changes in industry structure and innovation processes and to help design and evaluate public policies affecting innovation. Current science and technology indicators and data fall woefully short of illuminating changes that are known or believed to have occurred in the 1980s and 1990s. Although questions of feasibility, reporting burden, and cost need to be examined, the STEP Board believes that, in principle, a number of steps should be taken to improve the information base for microeconomic policy design and evaluation. These include collecting data at the business-unit level of the firm, conducting innovation and technology adoption surveys, linking datasets to one another, exploiting data on the training, career paths, and work of technically trained people, and exploring public-private partnerships to produce information useful to both corporate managers and public policymakers.
• Lack of an adequate, well-trained workforce—particularly those skilled in creating, developing, and deploying information technologies—may inhibit the capacity of the United States to remain prosperous and a locus of innovation. Immigration quotas have been raised, some states, educational institutions, and firms are expanding degree and training programs, and companies are paying higher premiums for skilled labor or seeking it abroad. What is not clear is whether, despite these measures, there will remain a growth-limiting shortfall between supply and demand and what additional steps, if any, should be taken to alleviate it.
• Strengthening and extending intellectual property rights (IPRs)—conferred by patents, copyrights, and penalties for misappropriating trade secrets—are appropriate policies for advanced industrial economies where intellectual assets are the principal source of growth. It may be that in some respects these processes should be taken further than the many steps accomplished in the past 25 years. On the other hand, there is growing friction over the assertion and exercise of IPRs and claims that in some circumstances they may be discouraging research, its communication, and use. The question arises whether in some respects IPR strengthening and extension have proceeded too far.
• The Board's case studies and limited national data suggest that the improved competitive performance of at least some of the industries reliant on the physical and information sciences, engineering, and mathematics—electronics, software, networking, and materials processing—has come about in the face of reductions in industry-funded longer-range research. Since 1992, public investment in research in several of these fields has also declined as a result of budget reduction pressures and changing

• federal agency missions. These trends are of sufficient concern to merit a careful assessment of their long-term implications and what steps, if any, should be taken to change them.

In the 1980s a series of studies portrayed the technological leadership and international competitiveness of U.S. manufacturing industries as imperiled and probably on the decline. While acknowledging U.S. strengths in academic research, the education of scientists and engineers, technological development, and venture capital financing of technology-based start-up firms, the studies observed serious weaknesses in the capacity of American corporations, compared with their Japanese competitors, to turn these first-class assets into advanced processes and commercially successful products. One study went so far as to characterize the decline of U.S. manufacturing as "systematic and pervasive" (Dertouzos et al., 1989), another as "a major historical development for this country" (Eckstein et al., 1984). The studies made several diagnoses, not mutually exclusive.

It was widely believed that, despite exceptional cases such as biotechnology, U.S. firms were underinvesting in general and, in particular, shying away from new ventures with longer time horizons but the promise of eventual competitive advantage, market share gains, higher returns, and even the creation of new industries. Different analysts emphasized different sources of this risk-averse behavior—among them, low national saving rates and a higher cost of capital or required return on investments ("hurdle rates") than faced by competitors in other countries. Others decried Wall Street's dictate that corporate managers show quarterly growth in profits to maintain stock prices, a lack of "patient" capital available to Japanese or German competitors through the substantial bank holdings in industrial corporations, and the inability of managers in some U.S. industries such as semiconductors to spread the risk of new technology development across the range of businesses allied in the typical Japanese keiretsu or Korean chaebol.

There was a great deal of concern that barriers to foreign market access and investment, dumping of products in third markets, public subsidies to firms and consortia engaged in technology development, and protection of mature industries from competition put U.S. companies at a serious competitive disadvantage.

A perceived neglect of process research and development (R&D) and product quality by U.S. firms was attributed to poor management education and technical training as well as to investment disincentives. As a result, many observers claimed, quality across a range of products from automobiles to semiconductors was suffering and customers were turning to more reliable non-U.S. sources.

Inferior precollege public education, preparation of school-leavers for work,

and career-long training and simply the lack of culturally ingrained habits of cooperation were cited as long-term U.S. disadvantages.

Only a decade after the last of these reports, the public mood and tone of academic analysis of the American economy are generally upbeat, buoyed by seven years of uninterrupted growth, low inflation, record job creation, low unemployment, and the first federal budget surplus in more than 30 years. Concerns about the competitiveness of American industries have receded even further as a result of the prolonged stagnation of the Japanese economy and the ensuing crisis slowing the economies of Korea, Singapore, Taiwan, and other Asian countries. In the rush to find fault with the Asian countries' economic and political systems, their remarkable growth performance over a generation is subordinated to their current economic problems.

Striking examples of this reversal in thinking since the late 1980s are the industry-by-industry assessments of the Massachusetts Institute of Technology's (MIT) Commission on Industrial Productivity, Data Resources, Inc. (DRI), and the National Academy of Engineering (NAE) contrasted with contemporary studies of the same industries under the auspices of the STEP Board:

Were U.S. industries and firms really doing that badly and foreign competitors performing that well a decade ago? Is the apparent reversal today an accurate picture? Although related, distinguishing macroeconomic fluctuations and generally slower moving microeconomic or structural changes is a difficult but necessary task. The tendency is to read the cyclical set of signals for the structural trends, especially when the macroeconomic environment is especially negative, as it was in the early 1980s, or positive, as in the late 1990s.

As for the future, will American industrial resurgence be sustained? Does the recent performance of a number of industries signify that they have permanently improved their comparative advantage and long-term growth prospects? Almost certainly the answer is that growth will not continue indefinitely in its current configuration or at its current rate. But even if we knew how to sustain resurgence and avoid recession, should we be satisfied with the status quo or should we aspire to a higher growth trajectory? What public policies and private-sector strategies would help achieve it? What current microeconomic trends may undermine it? What areas of ignorance need to be addressed?

To help answer these questions, the National Research Council's Board on Science, Technology, and Economic Policy (STEP) undertook an analysis in 1997 that had an ambitious objective and involved several components. The goal was to understand the role of technological and nontechnological factors in the generally improved performance and competitive status of American industries and the public policies contributing to that improvement on the assumption that such understanding could both help sustain growth and anticipate any extensive deterioration in performance.

The "technological versus nontechnological" distinction is a crude but useful way to underscore that technology and innovation, broadly conceived to include translation of prototypes into manufactured products and services, adoption of technologies from sources external to a firm, diffusion of incremental improvements in products and services, and investment in science and engineering talent and technical skills, as well as R&D, are responsible for perhaps one-quarter of economic growth in the postwar period. At the same time there are many other influences on economic performance—macroeconomic conditions; tax, regula-

tory, and other policies that affect industry sectors differently; education; legal institutions; corporate governance; and industryand firm-level strategies. In the long run, these factors condition one another.

With respect to technological factors, of particular interest is the relationship of apparent changes in public and private investment to current and future economic performance. Among the changes frequently cited are the following:

• There has been a marked change in the public and private shares of research expenditures, from parity in 1980 to a ratio of more than twice as much industry as federal government investment today.
• Largely as a byproduct of the end of the cold war there has been a change in the defense and nondefense shares of the federal government's R&D portfolio, resulting in less support for most fields of the physical sciences and engineering absolutely and relative to the biological and medical sciences.
• At the same time changes in the composition and orientation of private-sector research and innovation are believed to include an apparent shortening of the time horizon of corporate planning, focus on incremental improvements in technology, greater corporate reliance on external sources of technology and collaborative arrangements with public and private institutions, domestic and foreign, and movement of R&D activity off-shore or into regional concentrations of technology-intensive enterprises.
• In some sectors, goods production and service delivery have grown more dependent on technology and highly skilled labor but without an increase in formal R&D investment or much R&D activity at all.

Nontechnological influences on performance, too, have changed in greater or lesser degree in the past decade.

• Firms engaged primarily in the delivery of goods and services have come to dominate the economy, accounting for approximately three-quarters of the GDP and employment.
• Trade has become an integral part of the U.S. economy; exports and imports have increased from approximately 14 percent of GDP in 1980 to 25 percent in 1997.
• The occupational structure of the economy has changed with growth at both ends of the scale—managerial and professional jobs on the one hand and low-wage, low-skilled jobs on the other hand—while the share of mid-level skilled and paid jobs has declined.

The STEP Board has approached the tall order of sorting out these influences and effects from several different angles, including industry-level analysis, an examination of policy influences on corporate behavior, a review of trends in new

venture financing, and an assessment of the adequacy of aggregate data, especially that relating to industrial innovation.

Like the authors of the 1980s studies mentioned earlier, the Board concluded that God as well as the devil is in the details of changes at the industry and, necessarily, firm levels. Under the leadership of STEP member Ralph Landau, the Board identified a number of centers of sectoral expertise that included studies of industry performance and innovation as a principal element. For the most part these are multidisciplinary research projects sponsored by the Alfred P. Sloan Foundation at various universities.

To try to standardize these case studies, the Board asked David Mowery of the Haas School of Business faculty at the University of California at Berkeley to develop a general framework to analyze the determinants of performance over the past 15 to 20 years. An exception was a study of the chemical industry, over 150 years and four countries, by Landau and Ashish Arora.¹ The shorter time period was chosen to frame the change in performance and also for reasons of economy, although we acknowledge, as the chemical industry study illustrates, that the analysis would benefit from a longer horizon. The Board then convened two workshops that included the investigators selected along with industry analysts from the U.S. Departments of Commerce and Energy, the U.S. International Trade Commission, trade associations, and other organizations. The resulting commissioned papers were presented at a conference in December 1997 at the National Academy of Sciences, where they were discussed with representatives of the subject industries, interested government officials, and other scholars. Revised versions of these papers appear in a companion volume to this report, U.S. Industry in 2000: Studies in Competitive Performance.

The group of industries examined (see Table 1) does not represent a sample carefully chosen to be representative of all major sectors of the economy. We were unable to recruit participants with the appropriate range of expertise to assess changes in the resource extraction industries—petroleum and mining—agriculture and forestry, and automobiles, the source of much of the initial concern about declining U.S. competitiveness in the 1970s. Individually and collectively, however, the industries are good candidates for studying transitions in performance among American firms over the past 20 years. Furthermore, they enabled us to capitalize on the cumulative work of analysts informed by close relations with firms in their subject industries and, in some cases, access to proprietary data.

In contrast to the MIT, DRI, and NAE studies, our selection includes three ''service'' industries—banking, trucking, and food retailing. Although the 1990s proved unfounded the fear that U.S. manufacturing as a whole is endangered and in need of rescue, it no longer makes sense to focus exclusively on manufacturing industries, ignoring by far the largest, increasingly international, and knowledge-intensive sector of the economy. Not only are services the dominant sector of the U.S. economy, generating three-quarters of the gross domestic product (GDP) and employing 80 percent of the work force—90 percent if service functions in the manufacturing sector are included (Survey of Current Business, 1998; U.S. Department of Labor, 1998), they are a major customer for manufacturers in general and the sole customers of products such as aircraft, pharmaceuticals, and medical equipment. Conversely, information technology as well as traditional services represent a critical part of the foundation for production of high-value-added manufactured goods. In any case, it is increasingly difficult to classify as manufacturing or services a significant range of economic activities, as many companies perform functions and pursue markets in both sectors. For example, contrast the MIT commission's characterization of the computer industry of the 1980s as composed of "makers" of mainframe computers, minicomputers, and microcomputers (with software as an afterthought) with the contemporary description by Timothy Bresnahan, author of the STEP Board's computing industry case study:

Industry analysts were asked, among other tasks, to identify and discuss the public policies that over the past two decades most strongly affected the structural evolution, technological development, and performance of the industry sectors they studied. Some of their findings are presented below.

Other principal elements of the project included policy analysis, review of the financing of new firms, and evaluation of research and innovation indicators. The STEP Board decided to extend some of its earliest work on tax policy and corporate investment (National Research Council, 1994) to consider how incentives for R&D and U.S. tax rules governing international activities affect the investment behavior of technology-based multinational companies. Former STEP

member James Poterba chaired a steering committee that commissioned papers from leading international tax scholars, practitioners, and policymakers and convened a conference in February 1997 to discuss the results. The papers have been published in Borderline Case: International Tax Policy, Corporate Research and Development, and Investment (Board on Science, Technology, and Economic Policy, 1997).

The Board collaborated with the National Academies' Committee on Science, Engineering, and Public Policy in organizing a workshop to examine the viability of capital, management expertise, and other assistance to technology-oriented entrepreneurs from "angel" financiers and organized venture capital funds. In addition to assessing the prospects for continued growth of venture capital markets, the workshop considered the fluctuations in investment in different technologies, especially information technology, biotechnology, and medical services. Former STEP member Burton J. McMurtry organized and co-chaired the meeting.

Finally, a steering committee chaired by Dale Jorgenson conducted a workshop to assess the adequacy, utility, and policy relevance of the government's data on industrial research and innovation. At the request of the National Science Foundation's Science Resources Studies (SRS) division, the workshop was also designed to generate suggestions for improving this information base. A report of the workshop, including participants' recommendations, Industrial Research and Innovation Indicators was published in 1997 (Cooper and Merrill, 1997).

Before describing the collective lessons of our industry studies, the Board concedes that it is incomplete and may be misleading to compare all but the newest industries' performance over two or three decades rather than across generations and across countries. From the Landau and Arora account of the American, German, British and Japanese chemical industries since the mid-nineteenth century it is apparent that competitive advantage has shifted from time to time from one country to another and short-term strong or weak performance does not necessarily signify a long-term trend unless basic structural conditions remain the same.

First, history matters. Wars and economic upheavals have unpredictable but long-lasting effects. During the decades after the second world war, the U.S. chemical industry grew at twice the growth rate of the economy as a whole. It thus contributed to growth at the macroeconomic level, a role recently assumed by the information industries. This was only partly the result of the destruction of much of the German and Japanese industries and the damage inflicted upon the British. To a greater degree it was attributable to robust innovation, especially in petrochemicals technology.

Second, the size of the national market and the national political and social environment matter. In a peaceful world where natural resources can be shipped around the world, know-how and economies of scale can be decisive factors in maintaining competitive advantages. The "environment" encompasses a complex mix of institutions and policies, not only government macro- and microeconomic policies but also durable national institutions. These include university systems both performing research and training scientists and engineers, legal systems protecting property (including intellectual) and promoting competi-

tion, labor markets, capital markets and institutions transferring capital from savers to investors and allocating it among competing firms and governments, and corporate governance systems providing incentives to managers for investment and risk-taking and dividing profits between owners of companies and other stake-holders. None of these factors is fixed; all of them are changing but at various rates. A factor that may serve a country's industries well at one time may not at another. An example is Japan's low cost of capital, contributing first to an investment boom and then a bust.

Third, technological innovation matters, but in the broad context of human capital development and commercialization and diffusion of know-how, not simply research and invention.

With the benefit of this perspective, it becomes more feasible to summarize the findings of the other 10 commissioned industry studies, limited as they are to a period of only a few decades, primarily from an American point of view. Even though more or less enduring national system differences do not figure prominently in these explanations of industries' improved performance, the two perspectives are quite complementary. Competitive strength in the long run rests on a robust institutional infrastructure and a stable, predictable macroeconomic environment rather than on policies that are targeted on particular industries and firms.

The industries that the Board examined exhibit enormous diversity in structure. Some, such as chemicals, are highly concentrated, with a small number of global firms dominating international trade, capital investment, and R&D spending. Others, such as powder metallurgy and apparel, are populated by small firms with modest technical capabilities. In semiconductors, pharmaceuticals, computer software, and computer peripheral equipment, large and small firms appear to complement each other and sometimes collaborate. Food retailing, trucking, and banking exhibit some regionally based concentrations of small and mediumsized firms but also an increasing number of large national and international enterprises.

The structure of very few industries has remained stable in the past 20 years. In several cases leading firms have been displaced by second-tier firms or even new entrants. In other cases considerable consolidation has occurred. These structural changes are just the tip of the enormous "churning" that has affected regions of the country, customers and suppliers, and, of course, workers, often adversely. This churning has taken a variety of forms—employment downsizing by major companies, shifts in the location of operations both within the United States and abroad, and changing the skill requirements of many jobs, leaving many workers with minimal basic skills and unable to meet required competence levels.²

Although many people have not benefited, restructuring combined with continuous innovation, especially by smaller newer firms, has been accompanied by a steady decline in the unemployment rate in this country to levels not seen in decades. Moreover, the overall picture presented by the industry studies is one of stronger performance, both in mature lines of business and through proliferation of new products, processes, and services. Not all studies used the same measures of performance but frequently cited quantitative and qualitative indicators of improvement include increased market share, usually global market share vis-à-vis foreign-based producers and suppliers; growth in output; growth in productivity, sometimes measured as traditional labor or total factor productivity, but often framed in terms of industry-specific measures of productivity; and opening markets for new products and services.

Some preeminent U.S. industries, among them computers, pharmaceuticals, and chemicals, remain in world leadership positions although challenged by foreign competition. That does not necessarily signify stability and continuity. In most cases it means that U.S.-based firms, some but not all of them industry leaders in the 1980s, have capitalized successfully on innovation processes that have changed radically or are in the process of changing. In contrast to suppliers of other peripheral equipment and electronic components, U.S. hard disk drive producers have steadily increased their world market share by linking U.S.-based technical and design resources to low-cost efficient Asian production.

Another high-technology industry, semiconductors, has recovered from competitive decline, regaining its formerly dominant market share, apparently through a combination of product specialization and manufacturing process improvement. The steel industry has also recovered from a low point in the early 1980s.

Perhaps the most surprising cases, the powder metallurgy parts and apparel industries, have held on, even accommodating new entrants, through a combination of process improvements and responsiveness to customers. The recovery of the automobile industry has contributed to the relative prosperity of steel manufacturers and metal parts suppliers—just one example of a virtuous cycle of growth.

Finally, the banking, trucking, and food retailing industries are being transformed, primarily by the lowering of regulatory barriers to expanding product lines and geographic scope and by the incorporation of information technologies enabling the design and delivery of new services. But in these cases innovation has also been driven by domestic competitors nominally outside the industries—nonbank financial service companies (insurers, brokerages, etc.), railroads and airlines, and the ubiquitous Wal-Mart and warehouse stores.

Landau, Taylor, and Wright (1996) have argued that to explain shifting patterns of industrial performance across nations, it is essential to systematically

examine the context in which firms operate, starting at the most general national characteristics of industrial countries and proceeding through various levels of governmental, institutional and social factors, and macroeconomic and microeconomic policies to the characteristics of particular industries and firms. The synthesis here follows this analytical framework, termed "levels of comparative advantage," which represent higher and lower levels of aggregation (see Table 2).

Figure 1

Consumer price index (percent change from year ago).

Source: Federal Reserve Bank of St. Louis (1998).

A clear difference between the period of apparent decline of U.S. industrial competitiveness in the 1980s and the current period of sustained growth is the macroeconomic policy environment—the collection of actions taken by government to keep inflation low and business cycle fluctuations in output and employment small. Macroeconomic policy includes both monetary policy and fiscal policy. It also includes exchange rate policy and the coordination of national macroeconomic policies that affect international transactions.

Figures 1 through 7 show a strong correlation between a combination of steady and conservative fiscal policy emphasizing federal budget deficit reduction and cautious monetary policy emphasizing stabilization and low domestic inflation, interest, and exchange rates since the late 1980s.

Note that for the past several years not only have inflation and interest rates and the value of the dollar been low compared to their levels for much of the 1980s, but they have also been quite stable, avoiding for the most part the sharp ups and downs that characterized even the 1960s and 1970s. Instability in these market factors tends to discourage investment, which contributes to growth.³

Not surprisingly, this favorable combination of circumstances appears to have been beneficial to output and exports, and to a lesser extent, investment. Bernard and Jensen (1998) examined various explanations for the U.S. export boom in the 1990s and concluded that depreciation of the dollar coupled with increases in foreign income accounts were largely responsible.

Improved productivity—the key to rising real income and increased industrial competitiveness—is less easy to discern. As is well known, in the early 1970s, productivity slowed dramatically across the entire industrial world and has not since achieved the rates recorded in the 1950s and 1960s. The reasons for this

Figure 7

Productivity (yearly percent change in four-quarter average).

Source: Yardeni (1998).

slowdown have been analyzed and debated extensively and still are not a matter of consensus among economists. What is generally accepted is that in the 1990s U.S. productivity growth in the manufacturing sector, where output is easier to measure, accelerated, albeit not to pre-1970 levels, along with improvements in other economic indicators. In the now-dominant service industries, no significant improvement is discernible despite large investments in presumably efficiency-enhancing information technologies, but that may be a function of inadequate measures of output and overreliance on aggregate data (The Conference Board, 1998).

A growing body of empirical evidence at the establishment or firm level, much of it presented to a STEP-sponsored international research conference in 1995⁵ confirms that adoption of new process technologies, especially when preceded or accompanied by worker training and managerial improvements, has spurred productivity growth in many industries. Many of these information technologies have also supported development of new products and the appearance of new services, especially in the service industries. The only thing that aggregate data measure more poorly than improvements are the efficiency and quality of existing services.

Macroeconomic policy is not the only arena in which the U.S. government, frequently criticized for politically expedient economic policy vacillation within as well as between administrations, steered a steady policy course in the 1980s and 1990s with important consequences for innovation and performance in a number of industrial sectors. The following microeconomic policies also exhibit a high degree of consistency.

Beginning with the airlines in 1978, successive administrations and congresses have scaled back government control of exit and entry and services and prices in a series of major industrial sectors: trucking, railroad, energy/natural gas. And the process of deregulation is progressing in telecommunications, electric power, and banking.

Deregulation is enormously disruptive of established industry structures and has contributed significantly to the ''churning'' phenomenon that has characterized the economy over the past two decades, with both positive and negative consequences. Moreover, the effects tend to ripple out well beyond the deregulated sector. For example, trucking deregulation has affected food retailing; and the ongoing restructuring of the U.S. telecommunications industry encourages the entry and growth of firms providing specialized hardware, software, and networking to manufacturing and service industries of all sorts. The benefit of deregulation is to open up new market opportunities and release competitive pressures on established and new firms to exploit them. This appears to be a relatively slow process, however, with deregulated firms growing in efficiency only gradually. The good news for economic performance is that the benefits of deregulation are not a one-shot occurrence but extend over a period of time (Winston, 1998).

In antitrust policy the Reagan, Bush, and Clinton administrations adopted a substantially more lenient enforcement posture than their predecessors, apparently convinced by the argument that vigorous pursuit of limitations on domestic market power could impede U.S. firms in international competition. Justice Department guidelines and review procedures for mergers were relaxed somewhat, and major suits against some high-technology firms were settled or dropped in the early 1980s. In 1984 the White House also supported legislation, the National Cooperative Research Act, limiting antitrust penalties for collaboration among firms in precommercial research. Subsequent amendments extended the same treatment to some cooperative production activities. As a result of the gen-

erally permissive policy, the entries and exits, mergers and acquisitions, and cooperative ventures that characterized the restructuring of many industrial sectors in the 1980s and 1990s were largely unimpeded. Of course, the ongoing antitrust action against Microsoft and the recently settled Federal Trade Commission proceeding against Intel signal strong concerns on the part of antitrust officials about competition in the information technology sector.

Beginning in 1980, a series of legislative actions, judicial decisions, executive branch initiatives, and international agreements spearheaded by the United States ostensibly strengthened the rights of intellectual property owners and extended IPRs into new areas of technology.⁶

• The Bayh-Dole Patent and Trademark Amendments Act of 1980 enabled universities, other nonprofit organizations, and small businesses to acquire exclusive rights to inventions developed with federal support. In 1984 some restrictions on the kinds of inventions that universities could own were removed. Gradually, this policy was extended to federal contractors and research grantees, regardless of size in most circumstances, eventually reversing the previous pattern of federal agency assumption of patent rights and nonexclusive licensing.
• In Diamond v Chakrabarty (1980) the Supreme Court allowed patenting of organisms with artificially engineered genetic characteristics. Subsequently, the Patent and Trademark Office granted innumerable biotechnology product and process patents.
• In 1982 Congress established the Federal Circuit Court of Appeals to handle patent litigation appeals, limiting the wide variation in circuit appeals courts' treatment of patent infringement cases and generally strengthening the position of patent holders.
• The 1984 Hatch-Waxman Act extended the patent terms on regulated pharmaceuticals.
• The Semiconductor Chip Protection Act of 1984 established a sui generis mode of protecting the design or mask work used in semiconductor manufacturing—a form of protection combining elements of patents and copyrights with elements of unfair trade competition and trade secrecy law. Semiconductor firms' reliance on traditional patents and trade secret protection has also increased.

• The 1988 Process Patent Amendments Act enabled U.S. process patent holders to block the import of foreign products produced by methods infringing their patents.
• The U.S.-spearheaded multilateral negotiations under the auspices of the GATT Uruguay Round resulted in the 1994 TRIPS (Trade-Related Aspects of Intellectual Property Rights) Agreement, setting minimum standards of IPR protection and enforceability among World Trade Organization members and requiring protection of certain integrated circuit designs, plants and microorganisms, and trade secrets. At the same time, the U.S. government pursued stronger IPR protection in a series of bilateral venues.
• The 1996 Economic Espionage Act, a law primarily aimed at foreign industrial espionage for the first time subjected domestic trade secret theft to federal civil and criminal penalties. Formerly, trade secrets were protected only by state laws.
• The 1998 State Street Bank decision of the Federal Circuit Court of Appeals upheld the patentability of business application software.

Perhaps with the exception of the biotechnology industry, where new firms with strong patent positions find it easier to attract financing and large established pharmaceutical firms depend on intellectual property protection to protect their enormous up-front drug development investments, the effects of strong IPR protection are far from clear. Overall, there has been an increase in patenting, suggesting that firms are able to appropriate more of their technology investments. At the same time, the costs of protecting IPRs in litigation are high, suggesting that the main beneficiaries of strong IPRs are established firms. Recent research presented at an April 1998 Stanford University workshop on intellectual property and industry competitive standards suggests that patenting motivations, and hence strategies, differ systematically among industries and across countries (Headley, 1998; Cohen et al., 1998).

In two successive multilateral trade negotiations and a host of bilateral settings, some of them focused on particular industries, products, or technologies, the United States has pursued a reduction in tariffs and nontariff barriers and a recognition that a variety of public policies heretofore considered to be of domestic interest only—R&D supports, competition policy, and IPR policies—have discriminatory trade effects and ought to be subject to international rules. Successive administrations have also resisted limitations on foreign investment and avoided the kinds of import limitations previously used on foreign steel, automobiles, and other products.

The federal government's support of academic research in all major scientific and engineering fields as well as its substantial support for training scientists and engineers via fellowships, traineeships, and research project funding is perhaps the postwar microeconomic policy with the longest duration and continuous political support, until recently fluctuating only at the margin and to some extent in its composition because of the dependence of some research fields on mission agencies with changing requirements and fluctuating budgets. Federal support of research by all performers in most research fields continued to increase in real terms through the 1990s in a few research fields. In others, mainly physical science and engineering fields, federal support peaked in 1992 or 1993 and continued to decline until 1997, the last year for which actual funding obligations by research field are known. The implications are discussed below, and the data are presented in Appendix A. Of course, the effects of the reduction in research funding will not be discernable for several years, if then.

Not all public policies have exhibited consistency over the past decade-and-a-half. The taxation of capital has been especially erratic over a long period—for example, with regard to the neutrality of the tax system. With the exception of the Tax Reform Act of 1986, the tax system has strongly favored certain kinds of investments over others, with some biases changing dramatically from one tax bill to another.

The evidence presented at the STEP Board's conference on international tax policy is that U.S. tax rules governing foreign income and expense allocations of U.S.-based corporations and the tax treatment of corporate R&D have important economic consequences through their influence on the levels and location of research and innovation and capital investments of multinational companies that account for most of the R&D performed in and most of the goods and services exported from the United States. Yet these policies have been subject to the vagaries of the federal budget and partisan politics. For example, since its first introduction in 1981, the research experimentation tax credit has expired and been renewed nine times, occasionally after lapsing entirely and from time to time in a slightly different form that the one previously in effect (Nadiri and Mamuneas, 1997).

In one respect there has been greater consistency in federal tax policy. The Tax Reform Act of 1986 lowered the federal statutory corporate income tax rate from 46 to 34 percent, and it has remained near that level since, making the United States, initially at least, a relatively low tax country. Similarly, the statutory tax rate on capital gains has been reduced.

Coordinated, targeted policies supporting particular industrial sectors have

been rare, in contrast to the amount of debate that took place in the 1980s and early 1990s about the wisdom of such policies. There have been several instances of protection against competing imports—steel and apparel, for example—but among the industries examined, the semiconductor industry represents the only case of multiple government interventions. These included support of manufacturing technology development and diffusion, action against foreign dumping, and assistance in foreign market penetration. The authors of the STEP semiconductor study attribute some but not most of the industry's turnaround to these policies (Mowery, ed., 1999). In contrast, a number of state as well as federal programs to support precompetitive technology development have been untargeted but also modestly funded and often short lived. Not surprisingly, their effects have been diffuse.

In explaining U.S. industries' performance, the STEP case studies do not yield firm conclusions about the relative contributions of particular macro- and microeconomic policies. Their influence, undoubtedly, has varied over time and from industry to industry. All in all, however, the U.S. policy environment in this period has been supportive of industrial growth. Indeed, it approximated the most frequent prescriptions for recovery in the 1980s. For example, DRI opined in 1984 that the steps to promote a "healthier development of U.S. manufacturing" should include lowering interest rates, exchange rates, and the cost of capital through budget deficit reduction and capital market liberalization; opening up world markets through more aggressive trade policies; stable monetary and fiscal policies, to avoid the cyclical pattern of the postwar period and to encourage long-term investment; support of basic research and training, including support of cooperative research projects to meet world competition; and regulatory and tax policies that favor industrial development (Eckstein et al., 1984).

The public policy environment, however favorable to innovation and growth, has not dictated the variety of ways—some familiar, others more subtle, and not all of them successful—in which U.S. companies and industries went about responding to domestic and foreign competition, new market opportunities, and technological change. The 11 industries that the Board examined pursued one or a combination of the following strategies, with at least some near-term improvement in performance.

In a number of industries, U.S. firms have restructured their product lines rather than continue competing head to head with Japanese firms in established lines of business. U.S. semiconductor producers, for example, exited from the memory chip market, focusing instead on microprocessors and customized

devices where innovative design capacity conferred an advantage. As expected, these turned out to be the fastest-growing markets and, unlike the memory device market, did not readily attract new country entrants such as Korea.

A high rate of merger and acquisition activity characterized several industries in the 1980s and 1990s, not always with beneficial results. The case study of banking concludes that much of the consolidation as well as much of the industry's investment in information technologies initially diminished rather than increased stock market values.

Nearly all industries, even trucking and food retailing, increased their international activities in one way or another, or by using a combination of strategies—exports, mergers, alliances, and foreign investment—but rarely by large-scale movement of production offshore. An exception was the U.S. hard disk drive industry, which continued to compete head-to-head with Japanese and European producers, successfully increasing its global market share, by locating production of current-generation products in Singapore and older products elsewhere in Asia. Design and R&D functions did not follow, however, but have remained largely in the United States. Indeed superior management of geographically dispersed operations—R&D, production, and distribution—appears to be a comparative advantage of the U.S. Industry.

Several industries focused on reducing costs and improving productivity and quality in the manufacturing process. In the case of semiconductors, this yielded substantial gains, not only for producers but also for the struggling U.S. semiconductor equipment industry, although not surpassing the yield rates of Japanese semiconductor manufacturers.

Probably most important, firms in several U.S. industries have shown a remarkable ability to introduce new products and processes, capitalizing on shifts in demand to create new markets, often through the deployment of technologies new to those industries, as well as to accomplish cost reduction and quality improvement. In many cases this pattern has been associated with new entrants (e.g., specialty chemical firms, "fabless" semiconductor design firms that contract out their manufacturing, package express carriers, and steel minimills) or with intermediary firms supplying information technology and engineering services (e.g., consulting and accounting firms, software producers, systems integrators, logistics suppliers).

The largest trucking companies first purchased logistics services and later established subsidiaries to provide them to other transportation firms.

In computers, pharmaceuticals, and perhaps chemicals, innovation processes have changed radically, and U.S.-based firms, not necessarily the industry leaders of a decade or two ago, have capitalized on the shift to achieve a strong competitive advantage. Bresnahan describes how in the computer industry a "Silicon Valley" system of organizing innovations—multiple innovative companies excelling in components, hardware, software, networking, and other specialized parts of the industry—replaced the integrated, hierarchical, more self-contained "IBM" system of innovation.

In pharmaceuticals, the revolution in molecular biology was exploited as a production tool by small biotechnology start-up firms and as a research and drug discovery tool by the large established pharmaceutical producers.

A similar biotechnology-based shift may be occurring in the chemicals industry as Dupont and Monsanto focus their product portfolios on agricultural and other life science products. Their manufacturing know-how differs and is not easily shared, supporting the competitive advantage of an individual firm.

These fundamental changes in information technologies and biotechnology remain linked to long-range research in electrical engineering, computer sciences, and molecular biology, but over the past decade or longer U.S. industries have evolved different ways of accessing research.⁷ The large corporate research facilities of such companies as IBM, AT&T, Dupont, and Xerox were sharply reduced in size and, apparently, refocused on shorter-term product development. New corporate linkages to university research have been created, through direct funding by a single company of a particular university center, institute, or laboratory, through consortia such as the Semiconductor Research Corporation, or through faculty involvement in launching start-up companies.

Although the incidence and value to firms of outsourcing R&D are unclear, the increase in their frequency extends to the proliferation of joint research ventures, strategic alliances with foreign and other U.S. firms (many of them focused more on joint marketing than on R&D), and cooperative arrangements with federal laboratories through cooperative research and development agreements (CRADAs). The downsizing of some central corporate laboratories appears to have gone hand-in-hand with decentralizing the R&D function, perhaps linking it more tightly to business units and therefore to profit-making incentives.

Increasingly, however, innovation in many industries is not traceable directly to any source, inside or outside a firm, with formal research as a major activity but is introduced from

• intermediary firms such as specialized engineering firms (SEFs) in chemical processing, consulting and accounting firms in information technologies, and logistic firms providing Global Positioning System (GPS)-based vehicle location systems to truckers and time-sensitive delivery systems to food retailers and apparel firms;

The Board's case studies, especially of the service industries—trucking, food retailing, and banking—but also apparel, powder metal parts, and other manufacturers, underscore that efficient absorption and deployment of technology from external sources such as software, systems, consulting, and accounting firms, as well as suppliers, customers, and competitors, are themselves risky endeavors that, in addition to capital, require knowledge, skill experimentation, and analysis that for the most part are not classified as R&D.

The STEP Board believes that the goal of public policy should be (1) to sustain a high rate of growth of the economy over the long-term by removing sources of inefficiency and (2) to achieve a wider distribution of its benefits. The corporate strategies and apparently supportive public policies of the 1990s, although successful, almost certainly will need to be changed in the future to fit different circumstances. This race has no single clock and no finish line.

Moreover, several developments have undermined the stability associated with a world of vertically integrated firms with in-house R&D and proprietary technology. One is the emergence of specialized providers of R&D and technical services purveying expertise to all comers throughout the world and moving industrially relevant technologies more rapidly across national boundaries.

Several enduring characteristics of the American polity and economy probably bode well for the future. One is the sheer size of the domestic American market and the scale of its resources. A second is the remarkable flexibility of the political and economic systems, encouraging experimentation in the development and commercialization of new technology and providing relatively little protection for enterprises committed to established ways of doing business. A related factor is the culture's tolerance for failure. Finally, contrary to recent conventional wisdom about managers' and investors' myopia, markets over time, although they fluctuate, do a reasonably good job of favoring firms with high growth prospects.⁸

Despite the Board's general satisfaction with the progress of the past decade (and, taking a longer perspective, the postwar period) and our guarded optimism about America's economic future, the position of American industries in world markets requires further study and continual monitoring. In the meantime, our collective investigations raise four policy concerns that should be addressed: (1) the adequacy of measures and statistical data on research and innovation broadly defined; (2) the employment, income, and labor market effects of industrial resurgence and the adequacy of human capital to sustain it; (3) the implications for research, innovation, and technology diffusion of some aspects of the

continued extension of intellectual property rights protection; and (4) divergent trends in public and private investment in R&D infrastructure.

Revealing as the Board believes the industry case studies are, the findings are anecdotal. Carefully selected statistical data, collected nationally on a recurring but not necessarily very frequent basis, also are needed to help discern and track changes in innovation processes as well as to help design and evaluate public policy measures affecting innovation. Unfortunately, the data gathered by the federal government and some private investigators shed little light on the structural changes in innovation processes described above. For example, industrial R&D spending data collected at the enterprise level rather than the business-unit level cannot be linked to particular products and services or locations and reflect shifts in competition, orientation, and organization only on the basis of broad industry categories of dubious value. Current science and technology indicators and data fall woefully short of illuminating a major part of the story of American industry in the 1990s—the origins of a variety of information technology products and services and their implementation in a cross-section of industries. This is because data on innovation-related activities and investments other than formal R&D and patenting are extremely limited. In particular,

• technology adoption is captured only in occasional surveys and only in the manufacturing sector;
• specialized technology providers (consulting, engineering, and systems firms, etc.) are not surveyed regularly;
• intersectoral flows of information are captured poorly, in part because data on the mobility and activities of technically trained people, the principal agents of technology transfer, are not adequately developed or exploited; and
• measures of the value of intellectual capital and innovation are lacking.

The STEP Board believes that in principle the following steps would greatly improve the information base for microeconomic policy design and evaluation, although questions of feasibility, burden and compliance, administration, and cost need to be examined:⁹

• R&D spending data should be collected at the business-unit level.

• The government should conduct periodic innovation and technology adoption surveys in service as well as manufacturing industries.
• Emerging industries and intermediary organizations that play a key role in technology transfer and implementation should be included in appropriate surveys.
• Statistical agencies, scholars, universities, and professional associations should consider how human resources data (on training, career paths, and work patterns of technically trained people) can be improved and used to assess knowledge flows and innovation trends.
• Where possible, relevant currently collected datasets (e.g., R&D, patents, publications, employment) should be linked to each other and to geographic location by identifying information.
• Federal statistical agencies should explore whether public-private partnerships could produce information useful to both corporate managers and public policymakers at less cost and effort and with less burden on respondents.

The employment and income implications of technological and industrial change have been among the most contentious economic issues in the 1990s. One debate involves the extent of job displacement by downsizing, movement of operations offshore, or other factors and the effectiveness of public policies to assist workers' adjustment and retraining. A related issue is the apparent increase in wage differences between workers at the bottom and those at the top of the income distribution. This phenomenon is probably due in part to the fact that technological changes place a premium on skilled workers and put workers with minimal basic skills at further disadvantage. This income dispersion is moderated when other measures of welfare—total compensation and household consumption—are substituted for individual wages.

A third concern is that lack of an adequate, well-trained workforce may inhibit the capacity of the United States to remain prosperous and a locus of innovation. There is no doubt, in particular, about the great demand across most sectors of the U.S. economy and elsewhere for workers skilled in creating and developing information technologies. Innovation in and deployment of information technologies are straining the capacity of educational institutions and training programs to produce people with the necessary knowledge and skills to sustain this momentum. Immigration quotas have been raised, states and some firms have hastily expanded degree and training programs, and companies are paying higher premiums or accessing foreign skilled labor through foreign direct investment or telecommunications. What is not clear is whether, despite these measures, there will remain a critical shortfall between supply and demand and what if any steps should be taken to alleviate it.

In advanced industrial economies where, increasingly, intellectual assets are the principal source of value, productivity, and growth, strong intellectual property rights—conferred by patents, copyrights, and penalties for misappropriation of trade secrets—are an important inducement to invention and investment. For this reason, the extension and strengthening of IPRs in the United States and elsewhere in the past 25 years were appropriate and probably necessary. It may be that in some respects those processes should proceed further. On the other hand, there is growing friction over the assertion and exercise of some IPRs and claims that in some circumstances they may be discouraging research, its communication, and use. The question arises whether in some respects IPR strengthening and extension have proceeded too far.

Many enhancements of underlying IPR regimes reflect a greater professed appreciation of the incentive effects of protection on investment in R&D and use of intellectual property and appear to have had tangible results in a number of sectors, such as biotechnology and software. In recent years there has been an unprecedented surge in the overall number of U.S. patents applied for and granted to U.S. firms each year; and several major corporations, such as IBM, have made a great effort to exploit intellectual property through licensing. But these trends contrast with survey evidence suggesting that U.S. manufacturing firms in industries other than pharmaceuticals and chemicals rely more heavily on trade secrecy and lead time to recoup their R&D investments than they do on legal mechanisms such as patents and that, if anything, the effectiveness of patents as a means of appropriating R&D returns has declined since the early 1980s (Cohen et al., 1998).

In short, apart from pharmaceuticals and biotechnology, the effects of IPR changes on innovation and technical advance are highly uncertain—with respect to either the incentive provided to the innovator to capture the benefits of his invention, investment, and effort and therefore invest more resources and effort in innovation or the encouragement to the inventor to provide the information to others who might improve upon it. At the same time there are concerns about the manner in which IPRs are being asserted and exercised in some circumstances. These concerns can be categorized by their potential effects:

• on the performance and communication of academic research
• • concern that an international agreement favored by the European Union and the U.S. Patent and Trademark Office to extend copyrights to scientific databases will inhibit research;
  • concern that expressed gene sequence and other biological material patents will make it prohibitively complicated and expensive to conduct research using these tools or, alternatively, expose research investigators to infringement suits;
  • concern that allowing federal grantees to obtain patents has altered their incentives to conduct basic versus applied research;

• concern that universities', researchers', and sponsoring companies' financial interests in exploiting academic results (by IPRs and otherwise) are inhibiting open, timely scientific communication; and
• concern that universities' and potential industry research sponsors' inability to resolve differences over IPRs will discourage corporate support of academic research.
• • on personnel mobility and informal technical communication between rival companies
• • concern that enforcement of new federal trade secrecy laws, providing civil and criminal penalties for misappropriation, will have a chilling effect on mobility and informal know-how trading among firms (von Hippel, 1987).
• • on industry investment in R&D and innovation, both radical and incremental, initial and subsequent innovation
• • concern about the uncertainty of the scope of IPRs;
  • concern that slow and secret patent administration processes reduce R&D incentives;
  • concern about high litigation uncertainties and costs, both financially and in terms of the time of scientists, engineers, and managers; and
  • concern about licensing terms barring probing the intellectual content of software or genomic material and making modifications and improvements (so-called ''decompilation'')
• • on industry competition and structure
• • concern about the use of patent portfolios to block competitors' entry or discourage related research; and
  • concern about the penalties for initial innovators (e.g., business software developers) when IPR protection shifts from trade secrecy to patents.

The STEP Board believes that broad reassessment of IPR policies is therefore timely. What have been the costs and benefits of the actions taken in the last several years? The unintended as well as intended consequences? What should be the direction of IPR policies in the next decade or two decades? Should there be different approaches to intellectual property protection depending on the subject matter?

The improved competitive performance of many of the industries examined by the STEP Board has come about in the face of reductions in industry-funded longer-range research in some sectors. The industry case studies and limited

national data suggest that this frequent observation applies to at least some of the industries reliant on the physical and information sciences, engineering, and mathematics—electronics, software, networking, and materials processing (steel, chemicals, and metal parts). Leading corporate performers of industrial research in the 1970s—AT&T, IBM, Kodak, DuPont, and Xerox—had by the mid-1990s all downsized, redirected, and restructured their activities, particularly those concentrated in central research facilities, out of economic necessity and, in all likelihood, with near-term benefits for corporate balance sheets. The fastest-growing firms in information technology—Intel, Sun Microsystems, and Microsoft—for the most part eschewed traditional large-scale research organizations. But this pattern has not extended to pharmaceutical companies, the new biotechnology enterprises that have become profitable, or the chemical firms that have shifted emphasis to life science products.

Since 1992, public investment in research as well as development has declined as a result of budget reduction pressures, but also unevenly. As a function of their dependence on agencies with changing missions and declining budgets overall—the Department of Defense (DOD), National Aeronautics and Space Administration (NASA), and the Department of Energy (DOE)—certain fields of research have borne the brunt. They include most engineering and physical science fields, especially electrical and mechanical engineering, physics, and chemistry, although apparently not computer science and materials engineering. Together, the federal government's electrical engineering research support for all performers declined 36 percent between 1993 and 1997; university research support dropped 32 percent. There is little evidence that agencies' research portfolios (e.g. the National Science Foundation's) have been adjusted to compensate for the reductions in mission agencies' spending in these fields. At the same time, research in the biological and especially the medical sciences has benefited from steady growth in the budget of the National Institutes of Health, their principal source of support. Budget projections by agency through the year 2003 show a continuation of the same trends.¹⁰

NSF data on federal spending by field of research are available only through fiscal year 1997. In most cases, the reductions began to occur in fiscal year 1993. Five years is simply too short a period to be sure that these are long-term trends. Furthermore, agency research portfolios even in the same field differ markedly in character, so that a small reduction in one agency's budget might have a qualitatively more important impact on research in the field than a larger reduction in some other agency's spending. Determining how changes in spending by an

industry relate to changes in federal support of fields contributing to innovations in the industry would require a careful and detailed assessment.

Despite these caveats, the downward trend in public and private investment in certain fields of research is of concern because

• with a lag time, overall private R&D spending tends to follow the pattern of public spending, suggesting that a downward trend is difficult to reverse;
• the majority of federal investment in most research fields is, appropriately, a function of particular government missions and their political support; but the productivity of a field and its long-range prospects for contributing to successful applications may be neglected in the process of allocating resources to different programs; and
• although there is no reason that currently constituted research fields should continue to be supported at the same or increasing levels, there is apparently no mechanism for assessing support of fields of research related to industrial activity across agencies and for making adjustments in one agency's budget to compensate for another agency's spending reductions dictated by changes in the latter's mission.

The trends in several engineering and physical science disciplines are of sufficient concern to justify a selective effort to assess whether they are adverse and, if so, what steps should be taken to change them. Among the questions that need to be addressed in each assessment are the following:

• What kinds of research in what subfields are being negatively affected?
• Are investigators able to shift research sponsorship from federal agencies with declining budgets to agencies with increasing budgets?
• Is industry or another nonfederal source compensating for the decrease in public spending?¹¹
• To what extent do changes in support levels reflect changes in research and technological opportunities?
• What are the sources of support for graduate education in the field and is there a direct relationship between research funds and graduate training support?

If there is judged to be a deficiency in investment, what is the solution? The obvious answer—increased spending, greater efficiency—beg the question "how?", especially when it may not be feasible or appropriate to increase a mission agency's budget to support a particular research field. Other approaches are:

• Encourage inter-agency coordination. Among other forms this might take, agencies could arrange to share research facilities, avoiding duplication of infrastructure spending and maximizing limited programmatic funds.
• Encourage government-industry coordination. The semiconductor industry's MARCO program is an example of a private sector effort to compensate for government retrenchment.
• Institute a balance wheel. This might entail identifying an agency as a focal point for monitoring one or more major fields of research, assessing the need to pick up slack resulting from other agencies' mission-driven decisions, and adjust its own research portfolio accordingly. ¹²
• Undertake high-level priority setting. OMB and the Office of Science, Technology and Economic Policy might more directly take the health of key research fields into account in issuing budget preparation instructions, conducting budget cross-cut analyses, and negotiating agencies' budget requests.

The STEP Board's inquiry about U.S. industrial performance was prompted by the contrast between the diagnosis in the 1980s of secular economic decline and permanent loss of competitiveness and the experience in the late 1990s of growth, profitability, and stock market acceleration. In part the earlier pessimism was a function of the narrow focus on manufacturing industries and overestimation of their foreign competition. But it is also true that underlying U.S. strengths in innovation were masked by adverse macroeconomic conditions, especially high interest rates and the high valuation of the dollar. The resurgence is therefore partly macroeconomic—the combination of steady conservative fiscal policy producing low domestic inflation, interest, and exchange rates—and partly microeconomic—the combination of diverse regulatory, trade, and research policies and the responses of U.S. firms to domestic and foreign competition, new market opportunities, and technological change.

Hindsight yields cautionary lessons, however. Satisfaction with the resurgence and confidence in its sustainability run the risks of discounting the vulnerability of the macroeconomic environment and ignoring microeconomic trends

that may seriously undermine performance in the future. In the Board's judgment, four issues that merit attention are 1) the adequacy of measures and statistical data on research and innovation broadly defined; 2) the adequacy of human capital to sustain the resurgence; 3) the implications for research, innovation, and technology diffusion of the continued expansion of intellectual property rights protection; and 4) divergent trends in public and private investment in R&D and infrastructure. Short-term strong performance does not necessarily signify a long-term trend unless supporting institutions and policies are both strengthened and adapted.

Arora, A., R. Landau, and N. Rosenberg, eds. (1998). Chemicals and Long-Term Growth: Insights from the Chemical Industry. New York: John Wiley & Sons.

Bernard, A., and J. Jensen. (1998). Understanding the U.S. Export Boom. Working paper number 6438, (March). Cambridge, MA: National Bureau of Economic Research.

Board on Science, Technology, and Economic Policy. (1994). Investing for Productivity and Prosperity. Washington, D.C.: National Academy Press.

Board on Science, Technology, and Economic Policy. (1997). Borderline Case: International Tax Policy, Corporate Research and Development, and Investment. Washington, D.C.: National Academy Press.

Board on Science, Technology, and Economic Policy. (1999). U.S. Industry in 2000: Studies in Competitive Performance. Washington, D.C.: National Academy Press.

Bos, T. (1998). U.S. Index of Leading Indicators; Monthly SA, 1987=100. http://bos.business.uab.edu/charts/cgi-bin/charter.exe/CGI?feddal.

Bresnahan, T. (1998). Computing. In U.S. Industry in 2000: Studies in Competitive Performance, D. Mowery, ed. Washington, D.C.: National Academy Press.

Cohen, W., A. Gao, A. Nagata, R. Nelson, and J. Walsh. (1998). R&D spillovers: Patents and incentives to innovate in Japan and the United States. Unpublished manuscript.

Committee on Science, Engineering, and Public Policy. (1999). Evaluating Federal Research Programs: Research and the Government Performance Act. Washington, D.C.: National Academy Press.

Committee on Science, Engineering, and Public Policy. (1999). Observations on the President's Fiscal year 2000 Federal Science and Technology Budget. Washington, D.C.: National Academy Press.

Committee on Science, Engineering, and Public Policy. (1998). Observations on the President's Fiscal year 1999 Federal Science and Technology Budget. Washington, D.C.: National Academy Press.

Cooper, R., and S. Merrill, eds. (1997). Industrial Research and Innovation Indicators. Washington, D.C.: National Academy Press.

Diamond v. Chakrabarty, 447 U.S. 303, 206 U.S.P.Q. 193 (1980).

Dertouzos, M.L., R. Lester, and R. Solow. (1989). Made in America: The MIT Commission on Industry Productivity. Cambridge, MA: MIT Press.

Eckstein, O., C. Caton, R. Brinner, and P. Duprey. (1984). DRI Report on U.S. Manufacturing Industries. New York: McGraw-Hill.

Federal Reserve Bank of St. Louis. (1998). National Economic Trends. Available at http://www.stls.frb.org/publications/net.

Headley, W. (1998). The Stanford Workshop on Intellectual Property and Industry-Competitive Standards: Rapporteur's report. Unpublished manuscript.

Landau, R. and A. Arora. (1998). The dynamics of long-term growth: Gaining and losing advantage in the chemical industry. In U.S. Industry in 2000: Studies in Competitive Performance. D. Mowery, ed. Washington, D.C.: National Academy Press.

Landau, R., T. Taylor, and G. Wright. (1996). Mosaic of Economic Growth. Stanford, CA: Stanford University Press.

National Academy of Engineering. (1983). Competitive Status of the U.S. Pharmaceuticals Industry. Washington, D.C.: National Academy Press.

National Science Board. (1988). Science & Engineering Indicators--1998. Washington, D.C.: U.S. Government Printing Office.

Nadiri, I., and R. Mamuneas. (1997). R&D tax incentives and manufacturing-sector R&D expenditures. In Borderline Case: International Tax Policy, Corporate Research and Development, and Investment, J. Poterba, ed. Washington, D.C.: National Academy Press.

Taylor, J. (1998). The mosaic of economic growth. In Chemicals and Long-Term Growth: Insights from the Chemical Industry, A. Arora, R. Landau, and N. Rosenberg, eds. New York: John Wiley & Sons.

The Conference Board. (1998). Computers, Productivity, and Growth: Explaining the Computer Productivity Paradox. Research report 1213-98-RR.

von Hippel, E. (1987). Cooperation between rivals: informal know-how trading. Research Policy 16:291-302.

Winston, C. (1998). U.S. industry adjustment to economic deregulation. Journal of Economic Perspectives, 12(3):89-110.

Yardeni, E. (1998). Best Charts of the Month. http://www.yardeni.com/public/hand.pdf.