What Price Science?
This year, the leaders of nearly two dozen scientific societies and key members of Congress are issuing an urgent call to arms. Federal funding for research, at $75 billion, has dropped more than 3% in real dollars since 1994. And in President Clinton's proposed budget, the next five years look even grimmer, with a 14% decline scheduled. What's more, science luminaries charge, industry has cut back on long-range fundamental research. The implication: America is failing to lay the research groundwork for tomorrow's new products and industries. Without increases for basic science, warns Senator Phil Gramm (R-Tex.), "it will be impossible to maintain the U.S.'s position as the technological leader of the world."
Strong words. But are they true? Many economic studies point to big payoffs from science. But a careful analysis reveals that things aren't quite so simple. When it comes to today's critical policy question--will science funding cutbacks threaten America's future?--the economic studies shed no light. The only honest answer: No one knows for sure. The U.S. is "running a giant, uncontrolled experiment," says Brookings Institution economist Kenneth Flamm. "If there are effects, we won't see them until the next century."
But there are clues that the doomsayers are wrong. How research dollars are spent is clearly as important as the actual levels of funding--and recent trends are encouraging. In industry, companies from B.F. Goodrich Co. to Xerox Corp. have figured out how to increase the payoff from their own research and development. And after several years of stagnation, industry support for R&D is expected to rise 6% to a record $120 billion in 1997, according to a survey by Battelle Memorial Institute and R&D Magazine (box, page 170).
Moreover, government-funded academics are forging new links with other researchers and with industry, speeding innovation. "I don't have any sympathy for the people crying wolf about funding levels," says David J. Roessner, professor of public policy at Georgia Institute of Technology. "These new interactions may be making up for declines in federal funding."
More controversially, new studies suggest that the role of fundamental science--research that pushes the boundaries of knowledge--has changed. Conventional wisdom has it that scientists working in the lab make breakthroughs which are exploited in new products. But studies by Don E. Kash, professor of public policy at George Mason University, show that complex devices such as advanced chips spring not from new discoveries but from better engineering and new combinations of technologies. When Kash asked companies like Intel and GE for examples of how fundamental breakthroughs paid off in chips and jet engines, they came up blank. His conclusion: "Basic science has nothing to do with what is necessary for technological success."
Few experts are as heretical as Kash. But these new trends in implementing research do have radical implications for policy. For example, the U.S. might get a lot more bang for its R&D buck if it did more to encourage companies to forge research alliances and consortiums.
That's not to say that boosting federal R&D spending is a bad idea. After all, many U.S. execs agree with Microsoft Corp. CEO William H. Gates III when he says: "There is a systematic underinvestment in research" in both industry and government. But an increase just may not be possible. Today's fiscal realities require tough choices among the many programs vying for taxpayer support. Do we spend on Headstart or new particle accelerators? For decades, science won these trade-offs because it was thought vital to beating back the Evil Empire. With the end of the cold war, scientists now argue that continued support is just as vital to our economic future as it once was to our security.
CREDIBILITY GAP. As evidence, scientists point to four decades of economic studies that try to measure the role and payoff of research. But the studies fail to make a compelling case that cutbacks would be disastrous. In one type of study, economists add up all the usual inputs to the economy, such as labor and capital, and then assume that growth unaccounted for by these inputs comes from new knowledge. Some studies have suggested that science accounts for more than half of all growth. Trouble is, economists admit, the economy is so complicated that these results simply aren't believable.
In another approach, economists look backward to calculate the payoffs from individual research efforts. A new Commerce Dept. study concludes that a $431,000 project to produce error-correcting software for factory tools brought a $93 million gain in manufacturing efficiency.
Companies that have put their own R&D efforts to the test have gotten similar results. Xerox readily admits that it failed to capitalize on legendary advances in computing made at its Palo Alto Research Center (PARC). Yet, "PARC has in fact paid off many times for Xerox," argues Harvard business school professor Richard S. Rosenbloom, because studies show that investments in lasers, software, and new materials made it possible for Xerox to remain a leader in high-speed printers.
Yet such studies and anecdotes of research have two flaws. First, they can't tell us what the benefit is from the nation's entire R&D investment. And they can't answer the crucial policy question of whether cutting or adding a few billion dollars in funding would have much of an impact. For basic science, "the marginal rate of return is unknown--and may be unknowable," says Congressional Budget Office Director June O'Neill.
However, there is growing evidence that budgetary pressures can act on R&D operations like Darwinian natural selection to improve the breed. Consider the profound changes in industrial R&D. "Ten or 15 years ago, the idea was: Get the best scientists you can, turn them loose, and good things will come," says Joe A. Miller, senior vice-president and chief technology officer at DuPont Co. It worked for products such as nylon and Kevlar. But research chieftains point out that today's complex technologies, intense competition, and information overload have required new approaches. At B.F. Goodrich, for instance, research managers have not only linked R&D more closely to business units, they also have put in place a rigorous measurement system that balances risk and potential payoff in long-term projects such as new optical and electronic materials. With decreases in R&D funding, "better decisions are being made about what to spend the money on," says Victoria Haynes, Goodrich's vice-president for R&D.
Other companies have discovered that their once vaunted central research labs were dinosaurs--too slow, too insulated, and too expensive. So instead of trying to do all the research in-house, many are using their labs as windows to the larger world of new research. MCI Communications Corp. takes this strategy to its logical extreme. The company has no central research facility. Instead, MCI's 12,000 scientists and engineers work on developing new networking technologies procured from the world's best labs. "Why rely on one central R&D organization?" asks Chief Engineering Officer Fred M. Briggs. "We have a new and faster market-driven model of R&D."
PATHBREAKING. To many in science, this is a worrisome trend. "With corporate labs becoming users rather than creators of new frontiers, where will the new frontiers come from?" worries California Institute of Technology physicist Michael L. Roukes. Indeed, Briggs admits that MCI will never invent the next transistor. But company execs vigorously deny they're neglecting the future. "The perception that support of fundamental science is waning is really a misunderstanding," says Melvin I. Cohen, vice-president for research effectiveness at Lucent Technologies Inc.'s Bell Laboratories. Instead, he says, the epicenters of basic research have shifted from the physical sciences to pathbreaking work in software and networking.
Industry's R&D transformation offers hope that federally funded science could also do better. Some steps could be unpopular: Shutting down some of the 700 federal labs (total budget more than $10 billion) whose research is no longer top-notch, for example. A bit of budgetary pain, it seems, could provide the incentive to both weed out wasteful science spending and develop sounder policies.
Hints of what those policies should be come from recent studies on the role of federal funding in past innovations. A research team led by Georgia Tech's Roessner found that in three cases--reaction injection molding (used in plastics manufacture), magnetic resonance imaging (MRI), and the Internet--new technology did not flow directly from breakthroughs in basic science. In developing the Internet, for instance, the National Science Foundation played the role of matchmaker, putting together diverse collections of people and technologies without spending much on actual research.
These case studies offer important lessons for federal policy. One of America's great strengths is the ease with which scientists and innovators can forge links with, and move among, academia, industry, and government. Science and technology agencies should use more of their budgets to foster cooperation, some policy experts argue. "For 20 years, I thought that the most important thing government could do was fund R&D," says George Mason's Kash. "But now I think it's getting companies to collude."
Another U.S. asset is a university system that lures and trains the best scientists and students from around the world. Better policies would make sure this training continues, instead of focusing only on dollar figures for research and development.
CRUCIAL TIME. There is also a clear need for funding stability. Because research may take decades to flower, lower but constant funding is more productive than a roller-coaster budget that might average far more, says Alvin W. Trivelpiece, director of the Oak Ridge National Laboratory.
This is a crucial time for science and technology in America. The major justifications for spending billions in taxpayer dollars have crumbled with the Berlin Wall. Claims by the mandarins of science that funding cutbacks threaten the nation's economic future aren't supported by the evidence. These arguments also neglect the fact that some areas of science--like exploring distant galaxies and deciphering the origin of life--deserve support even if they don't have an obvious practical value.
Congress and the scientific community should spend less time fighting about levels of R&D spending and more time devising new policies for making better use of the money they have. If both sides get smarter about our R&D dollars, America's remarkable flowering of science and technology will bloom far into the next century.