(FORTUNE Magazine) – IS AMERICA investing enough in research and development, a cornerstone of future economic success? Total U.S. outlays for government and civilian R&D peaked at $157 billion in 1989 and have since slipped by a little over $1 billion. Government spending -- about half the total -- slumped because of cuts in the Pentagon budget; corporate outlays sagged along with the economy. Meanwhile, rival economic superpowers -- Germany and Japan -- are investing more than ever. While they are still some distance from matching the U.S. in total outlays, both have surpassed it in the percentage of GDP they dedicate to R&D: Japan spends 3% and Germany 2.9%, vs. 2.7% for the U.S. Moreover, by committing a smaller share to defense R&D than the U.S. does, Germany and Japan concentrate their efforts more intensively on R&D that packs a probable commercial payoff. Does all this imperil American competitiveness? And if so, where are the handholds that will keep the U.S. from sliding down the mountain? Should it follow the European Community, which stimulates R&D with cooperative ventures funded equally by governments and industry? The largest, Eureka, begun in 1985, has invested over $8 billion in robotics, microelectronics, and other % promising high-tech projects. Democratic presidential candidate Bill Clinton wants a civilian agency modeled on the Defense Advanced Research Projects Agency (Darpa), which has had astounding success in underwriting obscure ideas that turn into formidable weapons. A commercial Darpa would begin with an appropriation of $5 billion, say, and invest in businesses willing to develop technologies where the U.S. lags behind -- for example, flat panel displays for laptop computers, on which the Japanese today hold a near monopoly. President Bush has not spoken out on the subject, but the White House has long been chilly to proposals that would try to sort out ''winning'' industries from losers, as a civilian Darpa inevitably would. Could such an agency succeed? Critics fear bad bets and boondoggles. George Heilmeier, president of Bellcore, which does R&D for the seven Bell operating companies, headed Darpa in the mid-1970s. ''The beauty of the agency was that we knew who the customer was and what he wanted,'' says Heilmeier. ''Who is the customer for the civilian version?'' The $5 billion might be spread thinly and ineffectively over 435 congressional districts. Even Eureka seems weakened by the need to stroke political constituencies; it has funded 385 projects in 19 countries. The controversy over a potential civilian Darpa may obscure the fact that, in many respects, the American R&D system doesn't require major fixing. It's a complex, one-of-a-kind piece of machinery assembled largely in the aftermath of World War II. Fundamentally it's still sound, driven by three powerful engines -- corporations, universities, and federal laboratories. The first engine is sputtering a bit; the second needs to have some parts replaced; the third is oversized and underused. With moderate tinkering and shifting around of funds, especially from engine No. 3 to No. 2, the U.S. could get more thrust out of its spending. Here's a more detailed diagnosis and a prescription for the form a tuneup or an overhaul should take: Corporate R&D outlays grew at a lusty 7.5% a year through the early 1980s, topping $80 billion in 1989 and falling off slightly since. But when you look at the numbers industry by industry, what really is -- and isn't -- happening at the lab bench looks less alarming. Many companies are indeed spending less, a few more. But more is not necessarily better, if probable return to shareholders is the measure. In electronics, everybody, foreign and domestic manufacturers alike, seems to be spending too much. Computer makers continue to put 6% to 12% of revenues into R&D as if the heady days of the 1970s, when profits multiplied with each new generation of boxes, had never ended. ''They've had damn little to show for that investment recently,'' says Barry Bosak, a Smith Barney analyst who has followed the industry for 26 years. Trying to find its way in a changed world, even IBM will likely shrink its magnificent research network, which cost the company $6.6 billion, 10% of sales, last year. At the other end of the satisfaction scale is the pharmaceutical industry, where R&D investment is swelling and paying off. Investment banker Felix Rohatyn, a Pfizer director, says that company has upped its annual R&D budget to about $1 billion, a stunning fivefold increase over a decade. ''The technology for screening new compounds has advanced so rapidly that Pfizer has ten to 50 times as much research power as a result,'' Rohatyn adds. How then to make sense of the general trend, the decline in absolute R&D dollars? The mergers of the 1980s explain it in part. Corporations consolidated departments, including research, and used the savings to service the debt that accompanied acquisitions. But there are almost as many additional explanations as there are industries, mostly related to market forces. For example, because Middle Eastern oil is still cheap and plentiful, the need to invent alternative fuels seems less urgent than it did a few years back. Exxon's research on other energy sources peaked in 1982. Companies are also pooling some of their money with that of competitors to investigate industrial mysteries together. Sematech in Austin, Texas, collects $100 million a year from semiconductor manufacturers and others -- and gets a further $100 million annually from Darpa -- to improve chipmaking equipment. The Big Three auto companies (no foreigners need apply) have organized nine consortiums for precompetitive research. Many major companies seem to be shifting their R&D money downstream, away from basic research -- science with no particular commercial end in mind -- toward applied research and development, the workaday studies that lead to better, more competitive products. For example, at General Electric's celebrated laboratory in Schenectady, New York, it's tougher now for a scientist to pursue a line of inquiry if he cannot persuade one of GE's operating businesses to pick up the check. American chemical companies spend about as much on research as their German rivals, the other world leaders, says Wertheim Schroder security analyst David Silver. ''But everyone is getting away from pie-in-the-sky kind of stuff,'' he says, to concentrate on such down-to-earth matters as lowering the cost of plastic food packaging.

Okay, but those custards in the clouds -- great leaps of imagination that end who knows where -- are vital as well. Sometimes they create whole new industries, as the transistor and the laser did when they emerged from Bell Labs decades ago. These days Bell Labs' $3 billion budget -- slightly more in real terms than before the 1984 breakup -- focuses more tightly on improving the existing telecommunications system. That pleases AT&T shareholders and phone users, but the U.S. as a whole may have lost something in the process. Argues Frank Press, a geophysicist who is president of the National Academy of Sciences: ''Basic research is our comparative advantage in the world. In time, a lot of countries will be able to manufacture as well as the Japanese. We're different in being able to create wealth with science.'' Washington therefore would do well to increase support for the 150 or so research universities that have made American science superlative. They range from specialty savants -- the University of Akron is tops in rubber science -- to versatile megabrains like MIT and Stanford. Biotechnology was born in the laboratories of universities, the complexities of advanced computing resolved in their math departments. By contrast, Japanese universities are second rate. Europeans concentrate their research in a handful of centers like the Max Planck Institutes of Germany. Great though they are, they lack the diversity and creative tumult of American campuses. The Bush Administration has been generous with the universities. While other discretionary domestic spending has remained flat, budgets for the National Institutes of Health, the National Science Foundation, the Office of Naval Research -- chief patron of oceanography -- and other agencies have risen 22% since 1990 to about $12.5 billion. Most of the money goes to individual professors, who turn over part of the grants to their universities to pay for overhead. Yes, universities have misused some of those dollars, most notoriously for a pleasure yacht that Stanford says was charged to the government by mistake. But the great majority have been well spent. The schools need more federal help. ''Our enterprise is getting a bit ragged,'' says University of Chicago President Hanna Gray, a campus of inspiring but aging Gothic architecture framed in the window behind her. It was in a squash court at Chicago that Enrico Fermi assembled the first successful nuclear reactor in 1942, but labs built 40 or more years ago don't have sufficient ventilation and wiring for modern equipment. Should the taxpayers pay for the updating? Absolutely, argues Cornell University President Frank Rhodes: ''Research laboratories are really for the good of the whole society. Besides, it's impossible to load those costs into tuition, already $17,000 a year here, with 70% of our students on financial aid.'' Walter E. Massey, director of the National Science Foundation, estimates it will cost $3 billion to modernize the labs of the 100 leading research universities and an additional $2 billion to upgrade their electron microscopes, computers, and such. That would require a federal investment of $1 billion a year for five years. It's worth the price.

THE MONEY would be readily available if Congress would stop funding pork barrel science, projects of dubious scientific value, at a rate approaching $1 billion a year. For example, West Virginia Senator Robert Byrd, chairman of the Appropriations Committee, has plucked about $40 million out of the Treasury to pay for science and technology projects at his state's Wheeling Jesuit College, a pleasant but undistinguished school on a bucolic byway. Why does Wheeling merit the funds? Explains the genial rector, the Reverend Joseph A. Burke: ''It's good to spread the joy around.'' The Bush Administration wants $650 million from Congress for continued work on a superconducting supercollider in Waxahachie, Texas. In the enormous machine, expected to cost about $8 billion in all, tiny particles would be smashed into one another; the resulting smithereens could be the fundamental bits of matter of which the universe is made. At this point, the U.S. has no facility comparable with the one at the CERN European research consortium in Geneva. Even so, the exotic supercollider should probably have a lower priority than some of the worthy small research projects at universities. Says Ralph Gomoroy, who ran IBM's research department for 20 years: ''Historically, the individual investigator has been far more important dollar for dollar than the megaproject.'' Most of the other scientists FORTUNE interviewed agreed. The reasonable way to finance more basic research without increasing the U.S. deficit is to reduce the $22-billion-a-year budget of the federal laboratories, the third engine of the American R&D system. There were 726 such labs at last count, from lonely agricultural stations where a couple of botanists puzzle over a stubborn fungus to a dozen or so prestigious national labs that are largely responsible for American prowess in high-energy physics and nuclear weapons -- Argonne, Brookhaven, Fermilab, Lawrence Livermore, Los Alamos, Oak Ridge, and Sandia among them. Together the 726 employ an astounding 100,000 researchers -- about 20% of all the science and engineering doctorates in the U.S. ''These are the smartest guys in America,'' says Donald Runkle, General Motors' vice president for advanced engineering. ''Let's use the resource.'' He is: Los Alamos scientists are helping GM develop a fuel cell car. Runkle is right, but first the resource ought to be trimmed. Many federal labs are mediocre and could be shut down. Others are brilliant but redundant. Livermore, east of San Francisco, was created in 1952 to rival Los Alamos, the pioneer, in the design of atomic weapons. The competition may have made sense during the Cold War, but today it is hard to justify two labs -- each with a budget of $1 billion a year -- at a time when the U.S. isn't making new bombs. PUTTING THE GENIUS of the great labs at the disposal of business could be done in phases. Los Alamos, for example, has much to offer industry. A blood analyzer invented there to track the effects of radiation can also be used to detect cancer. The drilling technique the lab developed to plant test weapons deep underground can be adapted to tap geothermal energy in the earth's crust. Much trickier is the second phase of cooperation between industry and the labs: creating something new together. Foreign as such government-industry collaboration may be to both sides, it can be made to work. The Argonne lab, near Chicago, has undertaken an enormous project that could serve as a model. In the early 1980s the Department of Energy, which operates Argonne, surveyed academic, corporate, and government scientists and found that their wish list of useful tools was headed by a super X-ray that could peer into the atomic structure of materials. So Argonne is building an $800 million accelerator, the Advanced Photon Source. When it starts up in 1995, it will be an X-ray machine a billion times more powerful than the one in your dentist's office. Fifteen mixed consortiums of corporations, government agencies, and universities have agreed to pay $5 million to $15 million each for access to this remarkable viewer. Pharmaceutical manufacturers will be able to examine drugs down to the atom and design them faster. Oil companies could trace what happens to catalysts when they crack petroleum molecules. The great appeal of the Argonne machine: It has a clear research function useful to a number of industries crucial to American competitiveness. That makes it exactly the kind of undertaking the U.S. should be betting its R&D money on.

BOX: INSIGHTS

-- In many respects, U.S. R&D doesn't need major fixing. -- More corporate R&D spending isn't necessarily better. -- Washington should increase funds for basic university research. -- The national labs should be trimmed and tailored to work effectively with business.

CHART: NOT AVAILABLE CREDIT: FORTUNE CHART/SOURCE: NATIONAL SCIENCE FOUNDATION CAPTION: R&D SPENDING