In This Issue
The Changing Nature of Engineering
June 1, 1997 Volume 27 Issue 2
There seems to be a powerful confluence of technological and economic trends that, taken together, could produce dramatic shifts in engineering.

Science Policy v. Technology Policy: Resolving the Ideological Confusion

Sunday, June 1, 1997

Author: Lewis M. Branscomb

Government should focus on its role as sponsor of basic scientific and technological research, while responsibility for narrow problem-solving should be left to the institution with the problem.

In its annual Views and Estimates report to the House Budget Committee, the House Committee on Science, chaired by F. James Sensenbrenner (R-Wis.), tried to make the distinction between the government programs it supports and those it opposes. Here is the language the chairman used in describing the criteria he thought appropriate for federal funding of specific research programs:

Federal R&D should focus on essential programs that are long term, high risk, non-commercial, cutting edge, well-managed and have great potential for scientific discovery. Funding for programs that do not meet this standard should be eliminated or decreased to enable new initiatives. (U.S. Congress, House Committee on Science, 1997)

To make clear what the Science Committee majority does not like, Rep. Sensenbrenner added: "Beyond the demonstration of technical feasibility, activities associated with evolutionary advances or incremental improvements to a product or a process, or the marketing or commercialization of a product or process should be left to the private sector."

Chairman Sensenbrenner's use of six qualifying adjectives to describe the kind of research the government should support is an indication of the care thoughtful political leaders must take in avoiding the pitfalls of "industrial policy" and "corporate welfare." The issue is further complicated when, at the end of the sentence, the Congressman seems to restrict his approval to scientific discovery, implying that government should not be equally enthusiastic about technological discovery, if it also were "long term, high risk, non-commercial, cutting edge, [and] well-managed."

And, where does this definition leave research in academic engineering departments and national laboratories? How would the committee square federal funding for new Internet protocols or a massively parallel computer operating system, if it took Mr. Sensenbrenner's restriction to "scientific discovery" literally? Yet, he seems comfortable with "demonstration of technical feasibility," which sounds a lot more like technological research than "scientific discovery."

Two Issues of Confusion
This example of a sincere effort to clarify the role of government support for R&D attempts to deal with two issues that have been a source of confusion and conflict in S&T policy for years. The first is that the federal role in R&D should be defined by where the work falls in the spectrum from "pure" to "applied." That is, one should assume that because a research result may have early and substantial utility, markets can be relied on to fund it. The second is that the federal role in R&D should be defined by who performs the work, in the spectrum from academic to industrial. That is, one should assume that the market will reward firms for all the activities that it is appropriate for them to do.

I believe, and hope to show, that both perspectives are inadequate as a basis for sound decision making in U.S. science and technology policy. There is a better, simpler approach.

Ever since the great debate between Vannevar Bush and Sen. Harley Kilgore in the 1940s, there has been a robust consensus of both liberals and conservatives on two points:

  • Government must fund the most abstract, conceptual, and speculative scientific research--"basic" or "fundamental" are the adjectives usually employed. History shows enormous returns to society and usually weak appropriability to individual investors from basic science. Empirical evidence shows that only a few of the very largest firms invest in it, and they are reducing that investment as they turn to new strategies for commercial innovation.
  • Government should not fund commercial product development (except where the government intends to be the customer for the product, as in defense, space, and some other areas). (1) The market reward from successful commercial innovation should be sufficient incentive for both product and process development, in most cases.

This consensus worked fine during the cold war, when defense and other federal missions dominated the budget and economic benefits were expected to follow, automatically and without additional cost to government, from spinoffs from defense and space R&D and start-ups from basic research.

Today, the same consensus no longer solves the policy debate. Although federal investments in commercial technology are still very small compared with government mission-related R&D--probably less than 1 percent--these new civilian technology programs have become the focus of the policy debate (Goodwin, 1997). (2)

The Clinton administration, in a technology policy declaration (Clinton and Gore, 1993) made within a month of taking office, focused its attention on the 80 percent of the budget devoted to applied R&D and outlined a wide-ranging program of R&D activities intended to shift attention (and some $8 billions, ultimately) from defense R&D to civilian programs aimed primarily at supporting a more competitive, high-tech economy. The Advanced Technology Program (ATP), through which the National Institute of Standards and Technology funds "high-risk, generic" research in commercial firms, was the flagship of the administration's new attention to civilian technology in the 103rd Congress (1993-1994). It became the lightning rod for conservatives in the 104th Congress (1995-1996), who objected to such expenditures on "corporate welfare."

There are two ways to resolve the disputes about the role of the federal government in civilian, commercial innovation. The first is to hide behind the shield of the post-war consensus (i.e., don't fund it unless it is "basic scientific research"). The second is to resolve the complicated arguments about applied research, corporate welfare, and the role of the federal government in support of the nation's capability to innovate.

If one cannot make progress on the latter, one has no choice but to rely on the old consensus. The strategy of labeling those research areas appropriate for government support as "basic science" has quite a lot of appeal, not only to neoclassical economists and libertarians, but also to academic researchers. Even engineers in research universities are willing to let their work be called basic science if it assures the protection of their autonomy in the performance of their work. (3)

My concern is that failure to address the language and the policy ideas needed to talk about the federal role in private innovation may marginalize the nation's scientific and technical enterprise. The stakes are too high to allow this to happen.

A new attempt to resolve the confusion and identify policy principles that might attract bipartisan support is now underway. Sponsored by the Competitiveness Policy Council (4) and strongly encouraged by Mary Good, undersecretary for technology in the Department of Commerce, the project (5) urges a shift in the boundaries of the debate. Instead of searching for the line between science policy and technology policy, it advocates an integrated "research policy and innovation policy."

A report from the project (Branscomb et al., 1997) presents a nonpartisan evaluation of the federal government's science and technology policies and activities. The document puts forth proposals for parsing the debate, establishes six policy principles, and draws inferences about how federal S&T programs might be managed in the future.

There are two key ideas in the report. The first seeks a redefinition of the categories of research into which both government data and the political debate are divided. The second addresses modes of diffusion of government funded work to its ultimate beneficiaries.

With respect to the former, government should focus on its role as sponsor of research, including both scientific and technological research. Basic technology research deserves as much federal attention as does basic science research; indeed, the two are often indistinguishable. The National Science Foundation (NSF) and other government agencies should abandon the use of the highly ambiguous category of "applied research." Instead, the report suggests that types of technical activities should be designed to help distinguish public and private roles in the nation's science and technology enterprise.

With respect to the latter, the report suggests that public-private technology partnerships be structured so that it is clear how the results will reach a broad range of users and benefit the public at large. This implies using consortia of industry, universities, and national laboratories in almost every case where the government is not the customer for the end product (e.g., defense). It also recognizes an appropriate role for state governments, especially in those federal programs intended to foster economic development through industrial innovation and new job creation. Let me consider each of these ideas in turn.

The Division of Federal Research Activities
Why is the division of federal activities into "science" and "technology" troublesome? Too many people attribute to the word "science" a purity and distance from commercial value that demeans it and to "technology" the notion that commercial markets will create all of it that the economy needs. This is an old problem. George Brown (D-Calif.), when he was chairman of the House Science, Space and Technology Committee (now simply the House Science Committee), tried to convince NSF to give more priority to technology research by threatening to create a competing National Technology Foundation. He finally agreed to compromise by adding "and engineering" to "science" each time it occurred in the NSF enabling statute.

Researchers understand that the boundary between scientific and technological research is very fuzzy indeed. The gap between research and development, on the other hand, is a chasm easily seen and hard to bridge. One must invent technology to make progress in science. It takes scientific thinking to make progress in technology. I am pleading for the acceptance of the idea that support for basic technology research should be at least as strong as public support for basic scientific research. We should mean both when we say "research." I look forward to the time when it would be sufficient for the House Science Committee to say, "Government should fund research, but responsibility for narrow problem-solving should be left to the institution with the problem, whether it be a private firm or a government agency."

Who Should Pay?
What distinguishes research from other activities? It is the risk of failure and the uncertainty about who the greatest beneficiaries might turn out to be. What activities do not fit that description? More narrow or targeted problem-solving, where the risks are lower and the beneficiary clearly identified, should, like product and process development, be funded by the intended beneficiary. This might be a private firm or, in the case of technologies for use by the government, by the using agency. Who pays? Only the collective society can be expected to pay for research as I have defined it. But the institution that has a specific problem or wants to develop a product should pay for the work, whether that be a commercial firm or the department of defense.

Government should not shy away from supporting research simply because it may be useful, just as it would not shy away from exciting science simply because no one can say just how it may become useful. Seen in this light, a number of government initiatives (e.g., ATP, the Program for New Generation of Vehicles, Dual-Use Applications Program, and the Environmental Technology Initiative) can--or at least should be--seen as basic technology research, or, as I prefer, research. Industry understands what research is; firms have little need for the words "basic" and "applied." These words may be useful when discussing the motives of the scientist or engineer, but for debating the rationale for investment, I prefer Harvey Brooks' language: "opportunity-driven" and "need-driven," which convey the motives of the investor.

How about the conditions under which publicly funded research should be performed? Whether opportunity driven (from curiosity about nature) or need driven (to support new capabilities), government-funded research should be conducted with a high degree of investigator autonomy and in highly creative institutions that are effective not only in research but in diffusing the results to potential users.

The importance of providing widespread access to research results is based on the reality that private firms are looking increasingly outside their own organizations for innovative ideas, products, and processes. As big firms downsize their corporate laboratories, "Not Invented Here" is much less of an impediment than it used to be. Just as firms are changing the ways they look for ideas and are emphasizing an array of new collaborative relationships, so too the government should pay more attention to how technical ideas are made accessible to potential users. Therefore, in deciding whose research to fund, government should not only select the most creative, competent, and productive performer, but it must also take into account the likelihood that the results of the work will find their way in one form or another to many other users.

Technology transfer, or diffusion, is a contact sport. Even in the most abstract fields of science, information flows primarily through personal contact. Two kinds of institutions have particularly effective ways of making their technical knowledge useful and available to others. Universities do it when their graduates carry their new ideas with them to private employment. National laboratories try to do the same thing through cooperative research and development agreements (CRADAs) with private firms.

Firms do it internally, when the research is conducted there, but diffusion throughout the industry is normally slow, since it is carried primarily through the products themselves. If the firm's government-funded research is conducted through a consortium of firms and other research institutions, the sharing by the industry can also accelerate the diffusion of new knowledge, while reducing political concerns that one firm may capture an unfair fraction of the benefits.

State governments are spending over $1 billion in technology-based economic development and can help to assemble firms and other institutions to cooperate with federal research programs addressing the competitive needs of regional industries. I have long felt that this was an opportunity the ATP program has overlooked. The companion Manufacturing Extension Program (MEP) has worked collaboratively with states from the beginning and with considerable success.

As an example of the application of these ideas, the reports suggests that the ATP should be run directly by the Technology Administration; should fund primarily consortia or associations of firms, universities, and government-funded laboratories; should not provide funding to the largest corporations; and should focus government funding on basic technology research, while consortia members and states address commercialization, finance, labor, and other issues.

Adoption of the idea of a federal research policy, in which basic science and technology research enjoy the necessary autonomy for highly creative work, with the performer selected on the basis of merit--subject to evidence of effective diffusion processes for results--is the place to start. Next comes the question of how much to invest in research, both scientific and technological. Here two streams of priority converge on the same research community. Need-driven research derives its priorities from national socioeconomic and security priorities. Trade-offs are made against alternative ways of addressing the same national need. Opportunity-driven research derives it priorities from horizontal competition (i.e., from the research community itself).

The investment must be sufficient to meet demands on the system of higher education, create opportunities and capabilities to respond to need-driven demands, and respond to intellectual opportunities for rapid conceptual progress. But one must understand that the distinction between need-driven and opportunity-driven research is not found in the laboratory as much as it is found in the investor's resource-allocation decisions.

Given the apparently harmonious relationships between Reps. Sensenbrenner and Brown, the new policy taskforce led by Congressman (and physics Ph.D.) Vernon Ehlers (R-Mich.), and the productive beginnings of the Senate Science and Technology Caucus, there appears to be an opportunity to make progress toward a more broadly understood and supported research and innovation policy. In any case, a better way to explain the federal role in S&T and to resolve the bitter political conflicts over "corporate welfare" in the science and technology budget must be found, if these sophisticated leaders in the House and Senate are to attract the support of a majority of their colleagues over a long period of time.

References
Branscomb, L.M. 1997. From technology politics to technology policy. Issues in Science and Technology 13(3):41-48.

Branscomb, L.M., R. Florida, D. Hart, J. Keller, and D. Boville. 1997. Investing in Innovation: Toward A Consensus Strategy for Federal Technology Policy. Sponsored by the Competitiveness Policy Council. Cambridge, Mass.: Science, Technology and Public Policy Program, Harvard University.

Clinton, W.J., and A.G. Gore Jr. 1993. Technology for America's Economic Growth, A New Direction to Build Economic Strength. Statement on February 22. Washington, D.C.: The White House.

Goodwin, I. 1997. Washington reports. Physics Today 50(4):47-52.

U.S. Congress, House Committee on Science. 1997. Views and Estimates of the Committee on Science for Fiscal Year 1998. Available on the committee's web page, http://www.house.gov/science/v&e_1998.html. May 19, 1997.


Notes
1. Of course, the government does subsidize commercial product development through an indirect subsidy: the R&D tax credit. This provision has often lapsed and has never been made permanent. Many believe Congress will not renew the credit when it expires at the end of May 1997.

2. While some 55.7 percent of the budget is defined as "technology development," mostly defense-related, it would be hard to find any civil agency prepared to admit that it spent any of its R&D budget subsidizing commercial product development, other than the Congressionally-mandated Small Business Innovation Research (SBIR) program. Even here, the government funds are to be used for development of commercial ideas; commercialization is to occur in phase III, where federal funds are not to be used. SBIR was expanded to 2.5 percent of agency R&D procurements for FY 1997, for a total of about $1 billion. The original intent of Congress was that SBIR was a mandated set-aside of a fraction of agency R&D procurements made in pursuit of government missions, not private markets, although the goals were broadened in the 1992 reauthorization act.

3. On March 4, 1997, a coalition of 23 concerned science and engineering societies called on Congress to enact a 7-percent growth rate for the federal R&D budget. A number of distinguished engineering societies, including the Institute of Electrical and Electronics Engineers, joined with science societies in asking for more money for "basic scientific research." Apparently, the engineers felt the phrase "basic science" better conveyed the conditions of support they need, even when they seek to justify a broad range of research in mathematics, science, engineering, and mixed fields such as oceanography, energy, and materials research.

4. The council is a bipartisan body created by the Congress in the 1988 Omnibus Trade and Competitiveness Act.

5. This summer, the project will be releasing a monograph published by MIT Press. Edited by Lewis M. Branscomb and James Keller, it will include 17 chapters on current technology policies and programs, authored by members of the project steering committee and other experts from across the country.

About the Author:Lewis M. Branscomb, a member of the National Academy of Engineering, is Aetna Professor of Public Policy and Corporate Management, emeritus, Harvard University and chair of the Harvard Technology Policy Assessment Project.