5/6/98

CONTACT: Kathleen O'Toole, News Service (650) 725-1939;
e-mail kathleen.otoole@stanford.edu

Assessing the costs of the millennium computer bug

Many small businesses could be in for a shock soon when they start losing contracts because they have not upgraded their computer software to deal with the century date change.

That warning came at a gathering of Stanford professors and Silicon Valley business people at Stanford's Center for Economic Policy Research on Thursday, April 30.

Stanford economist Michael Boskin, who sits on several corporate boards, estimated that small businesses will begin to feel pressure from larger companies in six to nine months. Companies that are addressing the problem have been told they can still expect disruptions if their suppliers and customers are not, and so they will pressure the smaller companies to take action, Boskin and others said.

Only about half of American small businesses have begun to look for infestations of the so-called millennium bug, Michael Burns, vice president of the Year 2000 sales project of IBM, told the Stanford audience, and the rest of the world is well behind the United States.

The millennium bug refers to the fact that most computers were programmed using only the last two digits to indicate the year. They therefore will read the year 2000 as 1900, throwing off financial, security, manufacturing and other systems. In one survey of companies, about 20 percent reported already experiencing problems because of the bug, Burns said. Some people, for example, have not been able to use credit cards with expiration dates after the year 2000, a pharmacy company has had problems producing medicine labels with the proper expiration dates, and some prisoners were reportedly released by mistake in Arizona because the record system couldn't properly register sentences into the next decade.

Stanford engineer James Gibbons, special counsel to the university president who also sits on several corporate boards, urged American companies to assist foreign companies with their century date change problems. It is in the best interest of many companies who are heavily reliant on foreign businesses for supplies or customers, he said.

Great Britain is behind the United States in addressing the computer bug problem but ahead of the rest of Europe, which is ahead of Asia, said Boskin, a former chairman of the President's Council of Economic Advisers who consults with leaders of foreign central banks.

Gordon Werkema, executive vice president of the Federal Reserve Bank of San Francisco, said bankers are concerned about countries such as Thailand and Indonesia that have recently automated their financial systems, and about mid-size banks elsewhere that could fail in such a way as to have a ripple effect. The Fed, he said, has "hundreds of people focused" on ensuring the reliability and efficiency of its payment system. Several weekends this year have been reserved for banks and other financial institutions to test their computer systems' interactions with the Fed's computers in simulations of the Year 2000. The effort is distracting the banking industry from other efforts, he said, but there is no choice because not fixing the problems would cost the industry an estimated $2 billion a day after 1999.

Could the millennium bug cause a worldwide recession? Economist Edward Yardeni of Deutsche, Morgan Grenfell created a stir recently when he placed the odds at 60 percent that the bug will cause a year-long recession to match the 1973-1974 downturn provoked by oil price increases. Speakers at the Stanford forum didn't seem to think the bug would cause a recession, but they wouldn't rule out the possibility.

The costs of fixing the millennium bug, estimated now to be between $600 billion and $1.6 trillion globally, are spread over several years in a world economy that is about $30 trillion to $40 trillion annually, Boskin said.

Professor Emeritus William Miller, a co-founder of Stanford's computer science department, noted that the estimated total cost of fixing the bug in the United States is greater than the cost of the Vietnam War, which was a drag on the economy.

Some of those costs are not really "money down a rathole," in the sense of being a cost to economic productivity as a whole, noted Mark Wolfson, a professor of accounting and finance in the Graduate School of Business. Like building owners making repairs after an earthquake, he said, many companies are probably upgrading their information technology systems as they are fixing them, and some of the costs of fixing problems are really a transfer of high salaries from several types of professionals to software programmers and engineers. Software engineers now make 40 percent more than other similarly educated engineers because of the labor shortage, according to Miller.

The bug will affect the quarterly earnings reports of companies, economists said, and several wondered if U.S. accounting principles should be changed to treat software as a capital investment, subject to depreciation, rather than as a shorter-term investment that is "expensed" in the year it is bought. "One of the technological surprises has been how long-lived software has been, which we expense rather than depreciate like hardware," said Timothy Bresnahan, a Stanford industrial organization economist.

Gibbons also suggested changing liability laws so that the bug problem does not become "a merry Christmas for lawyers" in the next century. A bill to limit computer-bug-related liability in California courts was rejected recently by a legislative committee.

Many of the fixes that companies are making are temporary fixes that will require a longer-term solution later, Burns said. Miller cautioned business owners that buying new software tested for 2000 compatibility won't solve their problem if they plan to use it with older systems that aren't 2000 compatible.

"Over 40 years we've put layers of software on layers of software or wrapped new software around old legacy systems," said Miller, who wrote his first computer code in 1953. Fixing and testing large systems can take about two years, he said. With a little more than 600 days left, he urged businesses to focus on fixing their most crucial systems and on informing clients and suppliers of their back-up plans for remaining problem areas.

Miller also advised companies against "trying to hire your way out of it." One large bank, he noted, reported that it had assigned 400 people to the task. "I doubt they have 400 good people. Sometimes adding people will make the problem worse if they are enough slower."

The good news, he added with a grin, is that "if we do it right, we don't have to do anything more about it for 8,000 years."

-30-

By Kathleen O'Toole