500 days from Y2K, potential troubles loom

	Sign In Sign-Up

Wednesday, August 19, 1998

By Ted Bridis

The Associated Press

WASHINGTON -- Inside a suburban home darkened by a power outage from a summer thunderstorm, the top bureaucrat hired by President Bill Clinton to make sure the nation's computers survive past 2000 was just trying to survive the evening's muggy weather.

Lightning had knocked out electricity to the neighborhood, leaving John Koskinen and his wife without air conditioning. Inside, with flickering candles casting the only light, the temperature and humidity climbed.

"You know," Koskinen recalls telling his wife, "this could be what it's like on January 1st, 2000."

With just 500 days remaining, 2000 looms almost mythically, largely because of a decision made decades ago by computer programmers.

Left uncorrected, the Year 2000 computer flaw, commonly known as Y2K, could threaten the world's electrical grids, its financial markets, its water supplies and its air-traffic control systems. It probably won't affect the computer chips in your television or VCR, and your car or pickup probably will start, but the traffic lights or the subway might be fouled on the way to work.

"Over the last 20 years, we've grown more and more reliant on information technology," said Koskinen, chairman of the President's Council on Year 2000 Conversion. "We're able to produce all sorts of stuff for less than we used to, but it comes with a risk. Everything is interconnected."

The roots of the problem date back to the earliest days of the computer industry, when every byte of computer code counted because of the limited storage and memory then available. Early programmers represented each year by its last two digits rather than by all four - 1972 became 72. The trouble begins when computers using such programs reach the Year 2000, or 00.

Newer programs and devices don't have the problem. But many businesses have not updated some programs on large, mainframe computers, and those computers play vital roles.

What could go wrong

A few examples of what could go wrong if computers are not fixed in time:

* A telephone call that begins just before midnight Dec. 31, 1999, and ends minutes later on Jan. 1, 2000, could be billed as a 99-year conversation.

* ATM machines could refuse to issue money.

* A credit-card bill owed on Jan. 7, 2000, could be mistaken as 99 years past due.

* Social Security computers might refuse to issue a check to a woman born in 1912 because it might appear she hasn't been born yet.

* Computers that regulate the flow of electricity through power grids could become confused, producing widespread blackouts, while similar problems could disrupt telephone service and shut off water supplies.

Tests of computers that do all sorts of jobs show the problem is real enough. For example, when Chrysler rolled its clocks forward to simulate 2000 at one of its Michigan assembly plants, the computerized security gates wouldn't open.

But just what will go wrong when the calendar flips to the year 2000, and whether society will face a disaster or a mere inconvenience, is anybody's guess. What happens will depend on how many computers are fixed in the next 500 days and what is controlled by those computers that aren't fixed in time.

The size of the job is staggering. In the United States alone, there are 157 billion software functions that need to be checked, according to Capers Jones of Software Productivity Research, a software consulting firm in Burlington, Mass. And the 30 most industrialized countries of the world have an estimated 700 billion.

The larger, old mainframe computers still used by government and big corporations for many vital functions are particularly vulnerable. But the problem isn't limited to them. The electronics industry has found that some microprocessors - tiny chips used on circuit boards in everything from coffee makers to oil tankers - are susceptible and must physically be replaced.

The government figures that only 1 percent to 2 percent of all those chips might be faulty, but there are billions of them, and you can't tell which ones will fail without checking every one of them.

"You don't know which of those chips, and you don't know where they are," said Sen. Robert Bennett, R-Utah, chairman of the Senate's Year 2000 committee.

Estimates of the cost of fixing the Y2K problem in America alone range from $40 billion to $200 billion, and worldwide estimates are as high as $600 billion.

"The costs of fixing the Year 2000 problem appear to constitute the most expensive business problem in human history," Jones said.

The Securities and Exchange Commission considers Y2K so significant that, on Aug. 5, it required public companies, investment companies and state and local governments to disclose their preparedness and contingency plans and to describe how the problem will materially affect them.

While many companies, such as AT&T, and some government agencies, including the Social Security Administration, are reportedly well along in fixing their computers, others are far behind and some are just beginning to analyze the problem. More than 80 percent of large U.S. companies are behind schedule in fixing their computer bugs, according to a survey by the Cap Gemini America consulting firm.

A tangled web

Because industries and government agencies rely on one another to do business, those that fix their own computers still could run into trouble.

For example, Social Security's computers might all be working, but Social Security checks are actually issued by the Treasury Department's Financial Management Service, which is reportedly far behind on Y2K. Manufacturers could fix their computers and still be confounded by the computer problems of their suppliers.

If power grids fail, even fixed computers won't work. Earlier this year, a Senate panel surveyed 10 of the nation's largest utilities serving 50 million people and found none had a complete contingency plan for a computer failure. One company didn't know how many lines of computer code it had. Because electricity is distributed on a grid with many suppliers, a failure at any one of them could lead to widespread disruptions.

Forecasts of trouble

Thus, the uncertainty about what will befall us 500 days from now.

Alex Patelis, an associate economist at Goldman Sachs, predicts economic disruptions comparable to the trouble caused by a major natural disaster such as a hurricane - damaging but not enough to create panic or throw the country into a recession.

Edward Yardeni, chief economist and global investment strategist at Deutsche Bank Securities, on the other hand, warns of the likelihood of major disruptions to the global economy.

Still others, fearing economic collapse and civil disorder, are laying in supplies of ammunition, water and canned goods - preparing for Y2K as they might for a nuclear attack.

Drew Cutter, who lives in a farming community in northwestern Ohio, says his neighbors and family "are kind of pooh-poohing it." But he says he's not taking any chances. He bought a pair of 55-gallon drums to store water, and he plans to buy a generator.

"You're going to have glitches that may last a day or a week," he predicts. "I don't take it, like, head-for-hills stuff. I don't get that radical, where I buy a gun, but I just want to be prepared. It's better to know now than, say, next summer."

Koskinen, who has been criticized by some for his optimism, says the biggest danger is overreaction. "If everybody decides to take their money out of the stock market, there won't be a stock market. If a few people do it, it's not a problem. When a few million people do it, it's a self-fulfilling prophecy."

"Civilization as we know it is not going to end," Bennett says. "I'm not yet

ready to build a shelter in the back yard. But it might not be a bad idea to

have a little extra food and water."