Stanford University

News Service



CONTACT: Stanford University News Service (650) 723-2558

COMMENT: Timothy Bresnahan, Economics Department, (415) 723-9471;

Why computers have been slow to change large workplaces

STANFORD -- Computers have yet to show up as a large benefit to society in national income accounts, but that may be good news for our grandchildren, says Stanford economist Timothy Bresnahan. New information technology has developed so rapidly, he says, that we are still confused about how to use it well and that suggests potentially large benefits are yet to be reaped.

Bresnahan, who previously conducted an in-depth study of the auto industry, co-directs the Stanford Computer Industry Project, a large research initiative that looks at many aspects of information technology development and use. He leads the sub-team investigating ways that large businesses, non-profit institutions and governments use computers. These white-collar bureaucracies are believed to be the best places to get a close-up view of the "productivity paradox" ­ the much-discussed observation that computers have been around 50 years but have yet to show up in national productivity statistics as a major spur to economic growth.

Stanford economist Paul Krugman translated the paradox recently into more personal terms when he noted in a New York Times essay that earlier technologies reduced a cross-country trip from several days to five hours after World War II and switched shoppers from mom-and-pop stores to efficient, cheaper supermarkets. In comparison, Krugman pointed out that "computerized ticketing is a great thing, but a cross-country flight still takes five hours; bar codes and laser scanners are nifty, but there's still time to read about celebrities and space aliens in the checkout line."

Through interviews and other research inside bureaucracies, Bresnahan and his graduate students have uncovered several reasons why intelligent, effective use of information technology lags well behind the inventions of the raw technology itself. He recently cited these reasons at a gathering of Silicon Valley business leaders and economists that was sponsored by Stanford's Center for Economic Policy Research:

  • Information technology is not one invention but a long series that includes thousands of localized uses. White-collar work varies dramatically within and across companies, Bresnahan said, and since computers are believed to have their greatest impact on white-collar work, people have to invent thousands of ways to use the technology. "Progress is slow and incremental because brilliant ideas are always scarce," he said.
  • "Information technology isn't a cost-lowering technology in most commercial uses, but more of a tool for new product development or improvement in quality, especially for service industries," Bresnahan said. Quality improvements, such as automatic teller machines that allow people to bank around the clock, often don't show up directly as greater profits in a company's bottom line. In fact, one of the arguments made recently by a congressional commission for lowering the government's consumer price index is to make official statistics on the purchasing power of U.S. households reflect such quality improvements. (That commission is chaired by another Stanford economist, Michael Boskin.) "I don't think there is any serious question that society as a whole is much better off as a result of information technology," Bresnahan said. "But there is still a challenge to understanding why those benefits to society as a whole have come so slow."
  • Most new technologies get cheaper over time so that more companies can afford to reap gains from them. Information technology is getting cheaper per unit, but there are also large qualitative changes in capability, such as networking. "These quality changes cause excitement about the potential future uses but difficulty in foreseeing future directions," Bresnahan said. In a previous study, he showed that the advent of networked computers delayed decision-making about the purchase of computer equipment.
  • Sellers of the technology have yet to figure out the best ways to sell it, partly because of the quality changes. Commercial customers are "cranky and angry" about how the technology is marketed to them, which suggests, he said, there are still great opportunities for those who can figure out the best ways to meet buyers' needs.

Bresnahan's insights are based not just on statistical analysis and economic theory but on interviews that he and graduate students conducted with middle-level managers in 80 large companies, mostly in Northern California. They have documented two types of change processes within these bureaucracies as a result of computers.

"Big, radical change involves installing a whole lot of information technology including networks and changing the company's business model ­ the company's relationships with its suppliers and customers ­ at the same time," he said. He cited the example of Procter and Gamble , a company that used computer technology to respond to large retailers such as Wal-Mart.

"A Wal-Mart purchase order is one of the most frightening documents you ever hope to see," he explained in an interview. "If you get a shipment there in the wrong half day, the second time you are docked $50,000. If you don't want their business, you don't have to sign the contract."

Procter and Gamble invested in information technology to build shipping and inventory control systems that "would let them keep the level of inventory essentially just in time, which is very difficult, " he said. "The return on this type of investment is probably 30 percent and up, but this type of change is also really scarce."

Only 5 to 10 percent of information technology use in large business and government organizations has been that big or radical, Bresnahan estimates. More typical, he said, is the case of the Texaco department, where someone responsible for paying the company's electric bills bought a personal computer and built a data base that would pay the bills ­ replacing a fax machine, telephone and clerk for that function.

People involved in such incremental changes, he said, tell the interviewers that "I've been reading in the trade press how you ought to be able to this better, and I just can't see how to do it here in my company or here in my relationship with this group of customers, so I'm doing something that might be going in that direction and hoping that somebody else in the company or customer base can enhance it."

As a result, Bresnahan said, "you get iterative processes of invention which take a long time to show up as an overall gain in worker productivity."

Iterative processes of invention are not unique to computer technology. In 1989, Bresnahan's Stanford colleague Paul David pointed out in a now-famous paper that the invention of the electric dynamo also did not show up quickly in productivity statistics. David pointed out that it took a while for manufacturers to figure out how to deploy its advantage over the steam engine to lower their costs.

"Everybody knew pretty early on that the big advantage of electricity over steam power was that you could fractionalize it," Bresnahan said, "but people had to think of ways to take advantage of fractionalizing power. Most of what worked out was that instead of having one big power source running through the center of the building and turned by a steam engine, you scattered little engines all over the building and that meant the building could be redesigned so that you had a logical flow of work through an assembly line."

Bresnahan believes it will take even longer to figure out the most productive uses of computers because of the greater variety in white-collar work. "Making airplanes and making bicycles is more alike than, say, Wal-Mart and Stanford are alike," he said. "And when you talk about network computing, the span of a typical application gets bigger and bigger, so that the amount of knowledge that people need [to develop it] gets wider and wider."

The employees who devised Federal Express's successful supply chain management system for tracking packages not only possessed deep knowledge of computers but of how to get delivery people to hustle and how to please legal secretaries who were their key customers, he said. "As networks spread out, the knowledge to do what you want to do gets harder and harder."

Information technology "holds out this very general possibility for making sellers more responsive to buyers' desires, to make them responsive more quickly and to avoid producing the wrong things. But it is a possibility in this wide variety of institutional contexts, so it's not so surprising that it would take a long time and that people have very different ideas about what the technology can do."

For information available on the World Wide Web, see the Stanford Computer Industry Project at



Download this release and its related files.

The release is provided in Adobe Acrobat format. Any images shown in the release are provided at publishing quality. Additional images also may be provided. Complete credit and caption information is included.

© Stanford University. All Rights Reserved. Stanford, CA 94305. (650) 723-2300. Terms of Use | Copyright Complaints

© Stanford University. All Rights Reserved. Stanford, CA 94305. (650) 723-2300. Terms of Use | Copyright Complaints