Xerox PARC's Ed Chi wears a head mounted eye tracking device in Palo Alto, Calif. in 2001. (AP Photo/Paul Sakuma)

Once upon a time, some major U.S. corporations functioned almost like universities. Many of the country's brightest scientists worked in the research departments of firms such as DuPont, AT&T and Merck, and they were largely free to pursue whatever paths their experiments opened for them.

"That was a pretty special period in American history, indeed in world history," said Ashish Arora, an economist at Duke University and one of the authors of a new working paper titled "Killing the Golden Goose? The Decline of Science in Corporate R&D." He and his colleagues found that these days, large firms are less willing to maintain laboratories in house or to buy small companies with promising ideas, and that investors on Wall Street place less value on scientific capability.

The authors, who also include Duke's Sharon Belenzon and Andrea Patacconi of the University of East Anglia in England, argue that the change may be due to increasing competition from overseas firms. With narrower margins and less spare cash, firms lacked the resources to make risky, long-term bets on science. Yet the authors did not find evidence that science has become less valuable.

Indeed, firms are investing as heavily as ever in the proverbial goose's golden eggs -- applied research and patents. That is, companies remain dependent on the fruits of scientific research. If anything, shareholders value patents more highly than before, and large firms will pay a premium to buy out smaller competitors with many patents.

"There's a pessimism out there that somehow all the important discoveries have been made and the remaining ones are going to be harder," Arora said. "Our data casts some doubt on that pessimism."

Research isn't only useful to firms because it can yield new, profitable technologies. Since corporate researchers are often in touch with colleagues at universities and other companies, a staff of qualified scientists can provide all kinds of useful intelligence to managers at large firms. They can help identify patents or small businesses worth buying up. And in a large corporation that manufactures a range of products, a measure of intellectual freedom can help the company's bottom line. Whatever a company scientist happens to discover, an engineer in some other department has a chance of putting it to use.

At the same time, scientists in large firms inevitably have to deal with supervisors with business degrees who don't necessarily understand their work. The bureaucracy of a blue-chip company can be stifling. And it can be difficult for firms to keep control of their most valuable inventions. Xerox had developed a personal computer complete with a mouse and menus on the screen, but Steve Jobs was the one who figured out how to make the machine a commercial success.

It could be that corporate laboratories were ultimately a waste of money for many shareholders. They seem to think so, anyway. The value that the stock market imputes to scientific knowledge -- as estimated by the relationship among a corporation's market capitalization and its research expenditures -- has declined among the sample of 1,014 firms that Arora, Belenzon and Patacconi analyzed between 1975 and 2007. Corporate scientists are also publishing fewer papers on basic research, though their work in applied research hasn't diminished.

Directly measuring corporate investment in research and development is difficult, because the costs of employing scientists and buying laboratory equipment may change from year to year. Yet funding from sources outside the federal government -- which roughly indicates how much corporations are spending  -- has declined as a share of the total national research budget from roughly a third during the Clinton administration to about a quarter today, as shown in the chart below from the paper.


 

The three economists suggest that as U.S. firms confronted increasing competition from manufacturers abroad, they were forced to look again at their budgets and find ways of saving money. Laboratories were closed down or sold, and so were less lucrative units, narrowing the broad range of companies' activities that made research profitable for them.

Yet that leaves the question of where funding for basic research will come from -- a pressing one as Congress and President Obama prepare to negotiate a budget.

Arora and his colleagues argue that research remains necessary for society as a whole, if not for particular firms. They found that newly issued patents cite recent papers, indicating that inventors and engineers still depend on new science to devise useful machines and materials. Even if Xerox's shareholders didn't benefit from spending on Jobs's Macintosh, the rest of us did.

These days, many scientists try set up companies based on research that they've performed at universities with federal grants. They then hope they'll be acquired by an established firm.

As a share of the economy, however, federal funding for scientific research has been declining for decades, as shown in the chart below from the American Association for the Advancement of Science. Obama's proposed budget, released Monday, would increase federal funding for scientific research by about 6 percent.

Meanwhile, large concerns are paying less and less to buy small firms, unless those firms have demonstrated that their ideas are commercially applicable. "In order for scientists to get paid, they have to find a way to translate their science into products," Belenzon said. "They need to become businessmen." It's a waste of their talents, and they're likely to fail.

With established companies unwilling to invest in unproven technologies, the market for science is "broken," in Belenzon's words.

More federal research grants wouldn't fix it, but they'd be a good start. "It surely would not hurt," Arora said.