US innovation: On the skids

Technologists look to a new White House to reverse decade-long slide in R&D

It would be hard to exaggerate the angst that has gripped the US in recent months as the election nears, markets churn and assets melt. But the headlines that have made us dread picking up the newspaper mask a long-term problem that may shape the future of America more than John McCain's plan for Iraq, Barack Obama's health care ideas or Uncle Sam's heroic efforts to rescue the economy.

By most measures, the US is in a decade-long decline in global technological competitiveness. The reasons are many and complex, but central among them is the country's retreat from long-term basic research in science and technology, coupled with a surge in R&D by countries such as China.

R&D has two parts, of course, and published figures showing a rise in "research and development" hide a troubling trend. Companies still spend billions annually on development, typically aimed at the next product cycle or two. But the kind of pure research that led to the invention of the transistor and the Internet has steadily declined as companies bow to the pressure for quarterly and annual results.

To take but one example: Bell Laboratories was founded in 1925 and went on to "help weave the technological fabric of modern society," as its Web site today rightly claims. Its "top 10 innovations," according to parent Lucent-Alcatel, include the transistor, data networking, cellular telephony, digital switching, communications satellites and the Unix operating system. Although Bell Labs continues to innovate in most of those areas, all of the top 10 had their origins in the 1970s or earlier.

In January 1982, Time magazine reported: "With 22,500 people on its payroll (3,000 of them PhDs), 19,000 patents and an annual budget of [US]$1.6 billion, Bell Laboratories is a mighty engine of research and development. It is possibly the finest, and certainly the largest, private operation of its kind anywhere."

But since then, Bell Labs, beginning with its breakup in 1984 in the AT&T divestiture and continuing through subsequent sales and restructurings by its parent companies, has become steadily more focused on advanced development rather than pure research. On September 4 this year, The Star-Ledger of New Jersey reported that Bell Labs was disbanding a group of scientists doing basic research in areas such as material science and device physics. The paper reported that research director Gee Rittenhouse had explained that "the team was going to have a hard time integrating its research into product development."

Not only has industry cut back on research, it has taken much of it offshore, says David Farber, a computer science professor at Carnegie Mellon University. That deprives US scientists -- as well as non-US scientists who were educated here and want to stay in the country -- of some of the best jobs, he says.

Page Break

And the jobs of university researchers aren't so hot these days either, as professors and graduate students scramble for federal funds. "Faculty spend their careers writing proposals now. They don't get funded. The hit rates are low. People put in 20 proposals in a year," Farber says.

"Once you reduce university research, you are really mortgaging your future, because the way you train new scientists is by apprenticeships at graduate schools," he adds.

Where will the apprentices turn? "Eventually, we could all be hamburger flippers, or Wall Street brokers, if there are any left," Farber says.

Washington watch

The refocus from long-term research to shorter-term development in industry -- and Bell Labs is by no means the only example -- has been mirrored by a similar trend among the Washington agencies that fund science and technology, such as the Departments of Defense and Energy, the National Institutes of Health and the National Science Foundation. Federal funding for R&D has not declined overall -- it has, in fact, increased. But since the early 1990s, funding has been more and more focused on the short-term needs of government.

In particular, critics say, under the George W. Bush administration, the Defense Advanced Research Projects Agency -- which gave birth to the Internet, computer timesharing, computer graphics, LANs and much more -- has concentrated its research on short-term needs for warfare and homeland security. DARPA funding now tends to go to those who can promise measurable results in a year or two.

"DARPA funding has become short term, applications-oriented, highly competitive -- with small amounts of money and lots of reporting requirements," says Leonard Kleinrock, a professor of computer science at the University of California, Los Angeles, and an Internet pioneer in the 1960s. "That does not engender quality research."

Kleinrock recalls an earlier DARPA that would bring in extremely bright program managers and give them generous funding and carte blanche to pursue basic research in projects that could go on for many years, often with no promise of tangible results.

The American Association for the Advancement of Science, in a recent bulletin to its members about the federal budget, said, "Although high-priority investments in physical sciences research, weapons development and human space exploration help to keep the federal R&D outlook brighter than the bleak outlook for domestic programs overall, the FY 2009 budget continues the recent trends of declining federal support for research."

Page Break

The AAAS said the federal investment in basic and applied research would fall in real terms for the fifth year in a row under the fiscal year 2009 budget proposal. Meanwhile, it said, other countries, including China and Korea, are boosting government research by 10 percent or more annually.

The AAAS also presented data that shows that despite a big surge in health research funding for the National Institutes of Health between 1998 and 2003, total federal R&D spending as a percentage of gross domestic product has been in decline since 1976.

"Federal research investments are shrinking as a share of the US economy just as other nations are increasing their investments," the AAAS observes.

The Technology Policy and Assessment Center at the Georgia Institute of Technology recently completed a study that compares the technological progress of 33 countries between 1993 and 2007. It concluded that China has progressed more -- and more rapidly than any of the other countries -- while the US and Japan have slowly declined.

The Georgia Tech study computes the relative "technological standing" for the countries based on myriad social, economic and technological indicators, some statistical and some based on expert opinion. Combining the indicators provides a numeric measure showing that China has moved from a position far behind the U.S. to a point roughly even with it.

"The pattern is inexorable," says Alan Porter, one of the authors of the study. "China is coming up strongly, and it's in high-tech areas, not just cheap consumer goods."

China's rise is aided by an authoritarian government, low wages and a good manufacturing base, he says, but that isn't all. "You see tremendous effort in research in China," Porter says. "The US and China are neck and neck in basic science."

Gone astray

"We have kind of lost our way in some respects," says Vinton Cerf, chief Internet evangelist at Google and another Internet pioneer. "We have a significant diminution of industrial long-term research in IT, and we have seen one of the major federal sources of IT research -- DARPA -- essentially withdraw from a lot of that. Historically, DARPA would accept that it might take five to 10 years for an idea to yield anything."

Cerf says we should get used to the idea that countries like China will catch up with us in technology, simply because they have far more people. In fact, he says he doesn't like the word competitiveness because it suggests an adversarial relationship. He says he'd prefer that scientists and engineers work across borders to collaborate openly and publish their results.

Page Break

Cerf suggests that the new administration encourage immigration by the most talented science and engineering students. "I've always been puzzled by the H1-B issue, because for the past decade or two, the major fraction of researchers at universities have been from outside," he says.

"They are the crème de la crème, because they can't get in otherwise," he says. "Maybe they go home and maybe they stay, but they contribute mightily to the health of research and add a great deal of value to US research initiatives."

Henry Chesbrough shares that goal. "We are losing our ability to attract the best and brightest at the graduate level to come to the US," says the executive director of the Center for Open Innovation at the University of California, Berkeley. "There are two reasons," he explains. "Our concerns about security and immigration have caused us to be perceived as less welcoming. And the options back in the home countries are better than they have ever been. So at precisely the time we need to be more competitive to attract and keep these people, we are pushing them away."

But attracting bright PhD students is one challenge, funding their work is another. According to the AAAS, total federal funding for R&D at universities has risen slightly recently but, adjusted for inflation, has declined in each of the past two years.

Kleinrock says he is troubled by how campus researchers are changing the way they approach research in order to win the kind of short-term funding that comes from DARPA. "A lot of people are resorting to simulation, and that's fine, except they don't stop to ask what's behind the results they get," he says. "They are not being pushed to get a fundamental understanding; they are looking for the answers now, for this system, for today."

He says he worries that this short-term view of science will propagate from professor to student in a way that weakens subsequent generations of researchers.

Of course, given the current economic turmoil, most everyone sees other critical needs pushing R&D even further down the list of federal priorities. "Yes," Chesbrough admits, "that's going to make this deferred gratification even more difficult to accomplish."

Cerf says that while he's "deeply concerned" about the state of long-term research in IT in the US, he advises IT people to put it in perspective. "IT research is probably less utterly critical than, say, energy and the environment or, for that mater, figuring out what to do about health care and entitlement programs that are destroying our ability to have discretionary funding. You can't do research without discretionary funding."