"Companies our size cannot afford to do basic research, yet we need it," says Littlefuse CEO Howard Witt, shown here in his company's factory.

With $340 million in annual revenue, Littlefuse doesn't have the financial heft to invest in scientific inquiry that can take decades to percolate into products. The company, which makes fuses and circuit breakers for everything from cell phones to satellites, invests 3 percent of its revenue in R&D, but CEO Howard Witt says that the money goes mostly to D—developing products—rather than to R—basic research. Unless someone is doing enough R, the day may come when Littlefuse won't be able to create competitive products.

"Companies our size cannot afford to do basic research," says Witt. "We can't afford it, yet we need it. If basic research isn't done, we won't stay in business over the long haul."

Witt is among a growing number of electronics executives concerned that America faces a widening basic research deficit, with tectonic implications for the industry, the U.S. economy and the nation's defense. Basic research, conducted mainly in universities and national laboratories, is aimed at advancing fundamental scientific knowledge. Applied research for commercial product development and solving specific problems both flow from the wellspring of basic research.

America still leads the world in research, but the primary patron of basic research—the federal government—has steadily decreased the amount it invests. Mirroring the broader trend in overall R&D spending, federal investment in the physical sciences, math and engineering fell from 0.25 percent of gross domestic product (GDP) in 1970 to 0.16 percent in 2003, according to the Alliance for Science & Technology Research in America (ASTRA)—see the chart "How Much R&D in the GDP?" below.

"We've been living off research done years ago and have not replenished it to make sure the next round of innovation is forthcoming," says Paula Collins, director of government relations for Texas Instruments. "There isn't sufficient funding going into basic research for physical sciences, math and engineering. Federal spending has been flat for 30 years and has gone down as a percentage of GDP."

The most pressing need is to find a replacement for that building block of the Information Age, the CMOS transistor, which will reach its physical limits in about 15 years (see the sidebar, "When the CMOS Workhorse Retires, What Will Replace It?" below).

"We've been living off research done years ago and have not replenished it. Federal spending has been flat for 30 years and has gone down as a percentage of gross domestic product."
—Paula Collins, Texas Instruments

The basic research deficit has grown insidiously, despite some half-hearted measures by policy makers. Industry leaders, including Intel's Craig Barrett, TI's Tom Engibous and Hewlett Packard's Carly Fiorina, have begun to speak out as part of the offshoring debate (see "The Perfect Storm Brews Offshore," March 2004). One antidote to job loss due to offshoring, they argue, is to invest enough in research to propel the electronics industry forward and create new American jobs.

Apart from these voluble executives, "industry leaders have not been galvanized yet," says Ronald Turner, CEO of Ceridian and chairman of the board of governors of the Electronics Industries Alliance (EIA), which launched a campaign to do just that. "We have to come together and focus policy makers on how important innovation is to this country."

The challenge will demand more persistent pressure on Washington than the industry has ever mustered, according to interviews with nearly two dozen experts. It isn't clear that most electronics executives grasp this fact or have even figured out the right tactics.

Between 1980 and the present, GDP doubled, to $12 trillion, whereas federal investment in research for physical sciences, math and engineering declined by about one third (as a percentage of GDP). The last several presidents, aided and abetted by Congress, share responsibility for this trend. Now, facing record budget deficits and the costly war in Iraq, President Bush vows to reduce federal spending across the board, to the likely detriment of nondefense R&D, especially basic research.

The proposed budget was a mixed bag for electronics-related basic research. For example, the Pentagon budget for science and technology R&D, a source of research grants to universities and support for graduate students in electronics-related disciplines, was slashed by 15.5 percent from the 2004 budget, to $10.6 million. However, the proposed budget increased funding for the National Science Foundation (NSF), another source for electronics-related basic research, by 3.7 percent, to $4.2 billion, for the agency and by 2.1 percent, to $1.7 billion, specifically for physical sciences, math and engineering. As it worked through its appropriation bills, Congress' tinkering did little to reverse the long-term trend.

In its final Pentagon appropriation, Congress increased the R&D budget over Bush's request to $70.3 billion, a 7.1 percent increase over the 2004 budget. It increased science and technology R&D to $13.6 billion, 27.7 percent more than Bush had requested and 8.1 percent higher than in 2004. About $4.8 billion is for applied research and $7.3 billion for advanced technology development—primarily D. Only $1.5 billion is for basic research, but that is 12 percent higher than what Bush proposed and 6.1 percent higher than in the 2004 budget.

"We're living off the transistor, a 55-year-old technology. How much milk is left in this cow?"
—William Spencer, National Academies

The final appropriation bill that includes the NSF is not expected until year's end, but the House version trims the NSF's budget: The total is 4.4 percent less than what Bush requested and 0.9 percent less than in 2004. The amount for physical sciences, math and engineering is 2.7 percent less than in the Bush proposal and 2.5 percent less than in 2004. This in spite of the fact that Congress approved a bill in 2002 to double the NSF budget over five years.

In an analysis, the American Association for the Advancement of Science (AAAS) concludes that the 2005 budget is a repeat of recent years' budgets, which saw record-breaking totals for R&D due mainly to big increases for weapons development, homeland security and the doubling of the NIH budget. However, it says, "funding for other areas of federal R&D has remained stagnant or declined." If the trend continues, it concludes, the Pentagon science and technology R&D budget will be down by 2009 more than 17 percent from 2004 and NSF funding will be down 4.7 percent.

"The budget picture is extremely bleak for the next several years," says Vasundara Varadan, director for electrical and communications systems at the NSF, which funds 40 percent of all university research in engineering. She says that 30 of every 100 research proposals are worth funding but that there's money for just 15.

Even the Defense Advanced Research Projects Agency (DARPA), which created the infrastructure for the Internet, has had budget increases but with decreasing amounts earmarked for microelectronics since 1999.

There are some bright spots. The National Nanotechnology Initiative, an interagency, multiyear program in physics, chemistry and biology, has increased its investment from $116 million in 1997 to $961 million last year and a requested $982 million in 2005. The 2005 proposal includes about $300 million to be distributed by the NSF. And state governments are spending more on basic research (for more on state government efforts, see "R&D has a New York State of Mind").

Even with initiatives such as these, the U.S. is losing ground scientifically. The number of industrial patents filed for in the U.S. by Americans has fallen steadily, to 52 percent. The rest are held by interests outside the U.S. In 1983 Americans contributed nearly two thirds of the articles in Physical Review, a series of physics journals; today, they contribute less than one third. And America graduates fewer engineering undergraduates than China, Japan or Europe, and advanced degrees are down (see "Where Did All the EEs Go? " August 2004).

A recent study by the President's Council of Advisors on Science and Technology concludes that the electronics industry accounts for 10 percent of GDP but drives 30 percent of economic growth. "We don't have a crisis yet," says George Scalise, president of the Semiconductor Industry Association (SIA) and the council member who led the study. But with growing global competition, the government needs to double or triple its research investment in electronics if America is to maintain its economic strength, he argues.

Here's the gravest concern: CMOS, which has delivered continuous price/performance improvements for 30 years, will hit its physical limit by 2020. R&D efforts to replace it are underfunded by $1.5 billion a year, the SIA estimates.

The research deficit for CMOS is unconscionable, given the semiconductor's importance to the nation's economy and defense. The semiconductor improves productivity across all industries. It is the engine for the Information Age the way steel was for the Industrial Age. But America is approaching a crisis, because it is not investing enough to sustain its dominance in this technology. How did this happen?

It's been 20 years since electronics research policy made headlines. In the 1980s, Japan took over the DRAM business, raising the possibility that America might lose its semiconductor dominance. This led to the 1987 founding of Sematech, an industry consortium with matching federal funds, whose mission was to advance America's chip manufacturing technology.

"I don't believe that Sematech alone turned the U.S. chip industry around, but it was one factor," says William Spencer, who served as Sematech's chairman from 1990 to 1999. Spencer is retired but still active in policy debates as vice chairman of the science, technology and economic policy board of the National Academies.

In 1994, during the Clinton administration, Spencer and others unsuccessfully proposed a research initiative to address life after CMOS, similar to a recent proposal by the SIA. Except for a small uptick in the late Clinton years, federal funding of basic research for electronics declined. "We're living off the transistor, a 55-year-old technology," Spencer warns. "How much milk is left in this cow?"

The 1990s boom probably gave politicians a false sense of security. With the stock market shattering records, job growth exploding and tremendous wealth being created, how could research be underfunded? And with their short-term focus, politicians aren't likely to tackle this problem until the 11th hour.

CEOs share the blame, argues Robert Sutton, a management professor at Stanford University. He believes that executives, under intense pressure from Wall Street to drive up their stock prices, also focus on the short term, with few incentives to consider the long-term consequences of corporate strategy or federal policies.

Actually, the semiconductor industry raised R&D investment from $2.6 billion in 1990 to $14 billion in 2003, according to the SIA (see the chart "Steady Growth in Industry R&D"). Intel alone invests 10 to 15 percent of its revenue—more than $4 billion in 2004—on R&D. According to ASTRA, 70 percent of corporate R&D money goes to product development, 25 percent to applied research and 5 percent to basic research. With corporate emphasis on D, the federal government needs to bolster R more than ever.

Some argue that the mind-set in Washington is the biggest hurdle. "Despite the lip service they give it, most economists and policy makers don't really think technological innovation is that important to economic growth," says Robert Atkinson, vice president of the Progressive Policy Institute, a Washington, D.C., think tank. "Economists think capital is the most important thing for economic growth, so you just have to get the tax code right, have enough savings and cut taxes enough and then companies will do wonderful things."

The White House, says Doug Comer, Intel's director for legal affairs, views basic research as short-term spending, not as long-term investing. "We're trying to convince them that this is an investment you cannot afford to delay. But with Bush trying to hold down discretionary spending, we're just trying to avoid losing ground this year."

The AAAS's analysis notes that Bush promises to halve the nearly $500 billion budget deficit within five years, adding, "The President proposes to hold domestic discretionary spending at just a 0.5 percent increase while continuing to lavish resources on defense and homeland security discretionary programs, proposing $1 trillion in tax cuts over the next decade, and signing into law a new Medicare prescription drug benefit costing $539 billion over the next 10 years."

It would be a cruel instance of the law of unintended consequences if America were to lose its supremacy in electronics because the Bush tax cuts—designed to stimulate the economy—left too little money to invest in research to replace CMOS.

And that's why business groups are mobilizing. The EIA is campaigning for new policies for education, taxes and research. "We need a national vision that fosters an environment for winners in technology to emerge," says Ceridian's Turner. In the months ahead, Turner and others plan to meet any policy makers who will listen.

In April the Task Force on the Future of American Innovation, a group of businesses and universities led by Barrett, launched a campaign intended to educate policy makers and the public. Comer says that the goal is to double federal investment in research for physical sciences, math and engineering. "That's not going to happen in one year. We're telling Congress to raise it a little each year."

"Most economists and policy makers don't really think technological innovation is that important to economic growth."
—Robert Atkinson, Progressive Policy Institute

There's no dearth of ideas. The council of science advisers' study led by Scalise offers many recommendations, including the following: increase federal investment in basic research, involve state governments more, make the R&D tax credit permanent and create something to replace the great corporate research centers—such as Bell Labs—of the past.

Atkinson thinks electronics executives don't understand the challenge. He says technology is off the radar for most public officials: Republicans want to talk about nothing but tax cuts, and Democrats want to discuss protecting jobs. "Tech leaders don't understand this environment. Because they live with technology, they see the promise. They just assume that most people in Washington see it that way too. But most people don't."

One does wonder how persuasive industry leaders are. They've lobbied Congress—unsuccessfully—for years to make the R&D tax credit permanent.

Spencer says it's time to pull out all the stops. "Flocks of lobbyists from pharmaceuticals and biotech went to Washington in the 1990s. They met with key congressional people and twisted the arm of the president, first [the elder] Bush, then Clinton. They pushed Congress into increasing the NIH budget." Electronics executives need that intensity and a willingness to make personal sacrifices, he argues.

He believes that Sematech came to pass only because two industry luminaries set aside their personal careers to make it happen. Charlie Sporck, CEO of National Semiconductor, spent virtually all his time for more than a year lobbying for Sematech and then launching it. Robert Noyce, then CEO at Intel, gave up his job to be the first chairman. Not everyone thought these were the smartest career moves, Spencer recalls.

The campaign to trim America's basic research deficit will take a personal commitment Spencer doesn't see yet in today's executives. "I don't know if you've got people today like Sporck or Noyce. You'll never get a Steve Jobs or a Bill Gates or a Larry Ellison to think beyond being the richest person in the world."

Are we innovating on borrowing time? Send your thoughts to feedback@eb.reedbusiness.com.

Bill Roberts, a Silicon Valley-based EB contributing writer, urges top executives to donate the money they get from the Bush tax cuts to basic electronics research.