Research and development are supposed to be the fun parts of business--the free-thinking sectors where ideas are born and nurtured inside wind tunnels, autoclaves, supercomputers, and test tubes. Lately, though, some of the joy has gone out of R&D. Just ask Charles F. Larson, executive director of the Industrial Research Institute Inc. in Washington. "I am frequently in touch with literally hundreds of research directors around the country," Larson said in an Apr. 6 speech to the American Association for the Advancement of Science, "and I can assure you that not many of them are enjoying their jobs very much these days."

Worldwide, R&D managers are under pressure because their budgets are.

According to Standard & Poor's Compustat, a division of McGraw-Hill Inc., last year's increase in R&D spending by 900 big U.S. companies was just 1.9% after inflation, the smallest rise since 1986 (chart). Hard-pressed IBM was a big factor: The increase would have been more than a percentage point higher if IBM, which cut R&D 13%, had not been included. Abroad, companies in Britain, Switzerland, and Sweden had strong growth in R&D. But the composite of non-U.S. companies--before inflation adjustment--actually underperformed the U.S. (table). Japan was especially weak, with a 1% decrease in R&D outlays in the latest period for which results are available.

The good news, though, is that R&D managers are learning to make do in their straitened circumstances. Continuing a trend that began in the late 1980s, they're ditching marginal projects; decentralizing R&D efforts; creating teams to squire product ideas from lab to market; and figuring out how to collaborate fruitfully with outside experts, be they in other companies, consortiums, universities, or government labs.

The results of these efficiency drives show up around the world, from Italy's Fiat, which has replaced academically oriented "scientific committees" with pragmatic "innovation committees," to Canon Inc., which is setting aside Japanese pride and eagerly seeking collaborations with U.S. companies that are more in tune with the Information Revolution. For R&D, today's bottom line is the bottom line. Although blue-sky corporate R&D may benefit society as a whole, few companies will shoulder investments without clear payoffs for themselves. And more is not necessarily better. Just look at General Motors Corp. It ranks No.1 on BUSINESS WEEK's latest global R&D scoreboard. But it has steadily lost U.S. market share despite massive investment in research and development.

AWESOME THREESOME. In the quest for more efficient R&D, the hottest trend is collaboration. With technology getting ever more complex, no single company can do it all. Not IBM, Siemens, or Toshiba--three of the world's largest electronics companies--which are teaming up to develop 256-megabit memory chips. Not even hugely profitable Intel Corp., the world's biggest chipmaker. Intel's microprocessor dominance is being challenged by the PowerPC chip,

developed by the threesome of IBM, Motorola, and Apple Computer. On June 8, Intel struck back, announcing a venture with Hewlett-Packard Co. to develop a chip that will run a wide variety of machines, including personal computers, workstations, and file servers.

Collaboration is sweeping every field, in fact, from autos to aircraft to biotechnology. In Detroit, GM, Ford Motor, and Chrysler have formed 12 consortiums on such topics as electric-vehicle batteries, parts recycling, and better crash dummies. The idea is to avoid duplication by jointly performing basic, "precompetitive" R&D. One partnership's goal is to reduce today's more than 100 different wiring connectors to fewer than 10, which would be standard throughout the industry, lowering parts costs, reducing the difficulty of building cars, and making them easier to repair.

In biotech, ideas hatched in small companies often are nurtured by big partners. Case in point: An AIDS therapy from a tiny Quebec company, BioChem Pharma, that's being shepherded through clinical trials by Britain's Glaxo Holdings, will be marketed by yet another company, Burroughs Wellcome. The goal of such deals is to defray costs, spread risk, and promote the cross-fertilization of ideas.

One drawback is that collaborations are notoriously hard to manage. Turf battles and miscommunication, common enough within companies, can be magnified when separate companies try to cooperate. Mistrust can be fatal: It's why VM Motori, Europe's largest diesel-engine manufacturer, has steered clear of alliances. Indeed, VM Motori refuses even to patent its products for fear that its competitors will copy its knowhow.

Even where mistrust isn't an issue, R&D partnerships can fail if members decide linking up serves no vital need. The Advanced Computing Environment agreement, to create machines built on MIPS Technologies Inc.'s microprocessor design, quickly sank beneath the waves despite the nominal backing of everyone from Compaq Computer to Digital Equipment and Silicon Graphics. And Apple and Toshiba killed off their CD-ROM project--code-named Sweet Pea--before even producing a prototype.

The key is commitment. Says Yasutsugu Takeda, senior executive managing director at Hitachi Ltd.: "Get rid of the idea of `strength in numbers.' You can't create something just because you sign up a lot of companies that are barely committed and don't have anything to bring." Takeda cites Hitachi's alliance with Texas Instruments Inc. on 256-megabit memory chips as a successful "meeting of equals."

The same goes for the partnership of Hewlett-Packard and Canon, which has a giant 70% of the world market for laser printers. Canon is in charge of the "engines," which spit ink onto the page; Hewlett-Packard handles software, microcontrollers, customer research, and marketing. The alliance has lasted even though Canon and HP compete fiercely in lower-priced ink-jet printers. HP's ability to handle such a relationship bodes well for its chances of working with hard-nosed Intel on new chips. To insulate their engineers from corporate politics, Intel and HP are likely to put them in a new company with its own board of directors--assuming, of course, that the deal clears antitrust review.

Collaborations don't have to be between companies. Some of the most successful are internal--between business units of large companies that are scarcely acquainted. In Germany, Daimler-Benz Chief Executive Edzard Reuter is hell-bent on turning his company into "an integrated high-tech" concern. In pursuit of synergy, Daimler's Mercedes unit is developing a fuel-cell-powered vehicle with the aid of space jocks from Daimler's Dornier and MBB aerospace units, as well as an outsider, Canada's Ballard Power Systems.

SUPERSCRUB. Similarly, 3M Corp. uses internal cooperation as the linchpin of one of the world's most admired innovation programs. Two years ago, 3M was getting 25% of its revenue from products developed within the previous five years. This year, it expects to improve that to 30% of revenue from products developed in the past four years. How? One example is lowly soap pads--a market 3M was never in. Consumer research showed that people disliked rusting steel-wool pads. So 3M brought together experts from its adhesives, abrasives, coatings, and nonwoven technologies divisions to create its Never Rust plastic soap pads. Over the past 18 months, Never Rust and its spin-off, Never Scratch, have captured some 30% ef the U.S. soap-pad market.

As companies focus more on `D' and less on `R,' many are looking to universities to perform research that once might have been done in-house. IBM, for example, has donated or discounted millions of dollars worth of computers and lab equipment to such schools as Cornell, Iowa State, and the University of North Carolina in return for access to work done by professors and graduate students. According to the Battelle Memorial Institute, a Columbus (Ohio) contract-research organization, U.S. companies will fund $1.6 billion in research at universities this year. And universities will account for 16% of all U.S. R&D spending, up from about 11% in 1985.

Corporations in other countries can't tap their local universities for ideas nearly as easily. Japanese universities are starved for research funding, for instance. And the great European universities, many of them centuries old, tend to be more insular than U.S. schools. In 1989, students at Italy's tradition-bound public universities staged a massive nationwide protest against a proposal for partial American-style private funding, fearing it was the beginning of a corporate takeover.

Haltingly, U.S. federal labs are trying to get a piece of industrial R&D, too. Bomb makers, such as Lawrence Livermore National Laboratory, are working with such machine-tool companies as Cincinnati Milacron Inc. Even the Air Force's Ballistic Missile Defense Organization--the group that almost brought you Star Wars--is trying to get in on the act by licensing high-tech spectroradiometers for checking tailpipe and smokestack emissions. But the national labs still tend to gold-plate their designs. Says Jules J. Duga, a senior policy analyst at Battelle: "NASA wants to send somebody to the moon. I just want to send somebody to Toledo."

Don't want to deal with the stresses and strains of R&D partnerships? Sometimes the best way to get a new idea is to buy it. Not glamorous, perhaps, but cost-effective. Japanese companies have been doing it for years. "Most technology that's developed is not used. It sits on the shelf," says Kenneth Kovaly, president of Technical Insights Inc., an Englewood (N.J.) consulting firm. "As much as 10% of your R&D budget should be focused on finding technology outside your own company." If people such as Kovaly are heeded, the day may come when "not invented here" becomes a matter of budget-minding pride, not embarrassment.

BUSY, BUSY. As always, the hardest part of R&D isn't coming up with the ideas--it's pushing them through the organization. At Siemens Corporate Research in Princeton, N.J.--the U.S. lab for Germany's Siemens--George A. Drastal, group leader for knowledge-based systems, is slowly persuading managers in charge of industrial and medical-instrument units to try a new kind of "expert-system" software that helps technicians diagnose broken machines. It's a slog, he says, because amateurish systems have given the concept a bad name. And winning converts could get harder as business-unit leaders are more and more pressed by corporate streamlining. "Everybody's busier than they've ever been," says Drastal. "It can be hard to get somebody just to return a phone call."

That's true in the labs, as well. With their unstructured free time shrinking and the pressure for results mounting, it's quite possible that R&D staffs will one day suffer a creativity short-circuit. On the other hand, R&D never was synonymous with R&R. After all, Thomas A. Edison, the inventor who won more than 1,000 patents, used to work all through the night and kept a cot in the corner of his lab for catnaps. And he's the one who first said that genius is 1% inspiration and 99% perspiration.

                      Nominal percent change  1993        As percent      R&D per
                                              R&D spending
                                     1992-93  Millions      of sales  employee
      SERVICE INDUSTRIES              15%    $205.8           1.0%    $1,141.1
      HEALTH CARE                     11      13,886.9        10.6    18,247.8
      AUTOMOTIVE                      8       13,441.4        4.0     8,975.9
      ELECTRICAL & ELECTRONICS        8       8,504.5         5.5     7,060.7
      TELECOMMUNICATIONS              7       4,240.6         3.7     7,149.1
      CONSUMER PRODUCTS               7       2,322.9         1.5     2,294.7
      CONTAINERS & PACKAGING          7       154.0           0.9     1,347.2
      FOOD                            6       640.8           0.8     1,390.8
      MANUFACTURING                   5       3,931.1         3.0     4,314.7
      HOUSING                         2       494.7           1.8     2,743.6
      OFFICE EQUIPMENT & SERVICES     1       17,212.6        8.0     15,739.5
      CONGLOMERATES                   0       3,432.3         2.5     4,776.9
      PAPER & FOREST PRODUCTS         0       440.6           1.1     2,051.7
      CHEMICALS                      -2       4,993.0         4.1     10,265.1
      FUEL                           -3       2,660.9         0.8     4,466.7
      LEISURE TIME PRODUCTS          -3       1,891.5         5.4     8,087.1
      AEROSPACE & DEFENSE            -7       4,210.4         4.2     6,470.8
      METALS & MINING                -21      360.6           0.9     1,541.5
      ALL-INDUSTRY COMPOSITE 1993     4       83,023.6        3.8     7,476.4
      ALL-INDUSTRY COMPOSITE 1992     7       79,438.6        3.7     7,106.3
                      R&D'S BIGGEST U.S. SPENDERS
                                              1993 R&D expenses
              1       GENERAL MOTORS          $6,030
              2       FORD MOTOR               5,021
              3       IBM                      4,431
              4       AT&T                     3,428
              5       HEWLETT-PACKARD          1,761
              6       BOEING                   1,661
              7       DIGITAL EQUIPMENT        1,530
              8       MOTOROLA                 1,521
              9       EASTMAN KODAK            1,301
              10      GENERAL ELECTRIC         1,297
                                              1993 R&D expenses
                                              Per employee
              1       IMMUNEX                 $535,607
              2       BIOGEN                   208,724
              3       GENENTECH                117,594
              4       GENETICS INSTITUTE       107,657
              5       AMGEN                    83,302
              6       S3                       80,132
              7       CHIPS & TECHNOLOGIES     79,975
              8       CYRIX                    68,694
              9       ADOBE SYSTEMS            64,932
              10      CHIRON                   64,263
                                              1993 R&D expenses
                                              As percent of sales
              1       IMMUNEX                  340.9%
              2       GENETICS INSTITUTE       97.6
              3       CHIRON                   58.4
              4       BIOGEN                   58.1
              5       GENENTECH                48.5
              6       GENZYME                  36.0
              7       CONTINUUM                35.3
              8       MacNEAL-SCHWENDLER       28.1
              9       ENCORE COMPUTER          26.8
              10      KNOWLEDGEWARE            25.3
                    HOW THE U.S. STACKS UP
                         IN CORPORATE R&D GROWTH
      Top R&D spenders    Nominal % change  1993 spending   As percent     Number of
      among 200 companies 1992-93           Millions       Sales   Profits companies
      BRITAIN                 12.6%          $8,945         2.9%    33.5%   29
      SWITZERLAND             12.2            5,672         6.5     80.9    10
      SWEDEN                  10.8            1,961         5.7     560.3   7
      FRANCE                  3.2             8,749         3.8     74.7    23
      GERMANY                 1.9             25,942        5.9     347.0   25
      ITALY                   1.7             3,663         4.2    -178.5   9
      CANADA                  0.7             2,297         5.4    -454.0   6
      JAPAN                  -1.0             37,574        5.3     152.1   75
      THE NETHERLANDS        -3.1             3,805         2.8     45.4    6
      NON-U.S. COMPOSITE      2.1             99,850        4.7     115.7   200
      U.S. COMPOSITE          4.2             83,024        3.8     64.1    900
Before it's here, it's on the Bloomberg Terminal. LEARN MORE