No one can accuse Genzyme Corp., the biotech giant based in Cambridge, Mass., of being a technological laggard. Its scientists work at the forefront of biological research, while computer-controlled production equipment churns out hundreds of different advanced compounds.
So imagine the reaction of Robert Cowie, Genzyme's chief information officer, when he realized in 1996 that nearly every major system at Genzyme would be unable to cope with dates in the next century--the so-called Year 2000, or Y2K, computer bug. If left unfixed, the production equipment, the research and development computers used to analyze DNA, and the computers handling order-taking and billing would all fail on or before Jan. 1, 2000. Thanks to an early start, Cowie expects to have all his systems repaired well ahead of the deadline. Other businesses won't be so fortunate, he believes. "People will have to be complete by the end of this year, or it will be too late," says Cowie. "But there are still many companies that haven't done anything yet."
If Cowie is right about the overall lack of preparedness for Year 2000--and there's growing evidence that he is--a lot of businesses are in for a nasty shock. Up to now, skeptics have been able to pooh-pooh Year 2000 as a relatively easy-to-fix bug, an example of overheated hype by consultants looking for a quick buck. But there's growing alarm in Washington and elsewhere. The Securities & Exchange Commission has recently been strongly urging companies to start reporting the effect of Year 2000 on their earnings. "Many people think it is vastly exaggerated," says Percy Barnevik, chairman of Swiss-Swedish industrial giant ABB Asea Brown Boveri Ltd., one of the world's largest companies. "But it is a really big challenge."
ARMIES OF PROGRAMMERS. Indeed, the Y2K bug is shaping up to have a profoundly negative impact on the U.S. economy--starting almost immediately. According to a new analysis prepared for BUSINESS WEEK by Standard & Poor's DRI, the growth rate in 1999 will be 0.3 percentage points lower as companies divert resources to fix the problem. Then Y2K could cut half a percentage point off growth in 2000 and early 2001. That would be the same size as the expected economic damage from the turmoil in East Asia.
All told, the Year 2000 bug could cost the U.S. about $119 billion in lost economic output between now and 2001. "The Year 2000 bug will certainly hurt the economy," says DRI chief economist David A. Wyss, "and it could be a real killer if more of the problems are not fixed."
It's not only the growth rate that will be affected. Starting in 1999, inflation will be higher than it otherwise would have been and productivity growth will be lower. The reason: Instead of creating or installing new productivity-enhancing programs, every company that uses computers is diverting money and skilled workers into patching old programs.
Think of it as a town threatened by a rising river. Every able-bodied person--no matter what their job--is put to work stacking sandbags, while economic activity in the rest of the town slows down. In the case of the Y2K bug, experienced programmers and computer-science PhDs are doing what is essentially unproductive labor.
And the amount of skilled labor needed is enormous. Finding, fixing, and testing all Y2K-affected software would require over 700,000 person-years, calculates Capers Jones, head of Software Productivity Research, a firm that tracks programmer productivity. "People are spending huge sums of money fixing problems that are 35 or 40 years old, just to stay in business," says John Bace, research director at Gartner Group Inc.
Indeed, the demand for Year 2000 fixes helps explain why management consulting and accounting firms have added a stunning 200,000 new workers over the past two years. Moreover, with programmers' wages soaring because of the overpowering need to fix the problem, Year 2000 has the potential to push up inflation.
Then, at the beginning of 2000, the economy is almost certain to be hit by some computer failures, since it's becoming clear that many companies have gotten a late start attacking their Y2K problems. The BUSINESS WEEK/DRI estimate assumes that 85% of software programs will be fixed or replaced, which may be optimistic: A December, 1997, survey by Howard Rubin, a computer-science professor at Hunter College in New York, indicated that two out of three large companies did not yet have detailed plans in place to address Year 2000. Small companies and government agencies are even further behind.
Year 2000 is going to pose a serious policy dilemma for the Federal Reserve, which has been trying to steer a tricky course. Over the next couple of years, Fed Chairman Alan Greenspan is going to be under increasing pressure to raise interest rates as labor shortages for information-technology workers send wages skyrocketing. The problem: A rate hike in 1999 would run the risk of worsening the post-2000 slowdown.
Some analysts even worry that Y2K could send the economy into a recession. Edward E. Yardeni, chief economist at investment bank Deutsche Morgan Grenfell Inc., sees a 40% chance of a sharp downturn. One way that could happen is if there's a major failure in the government's computer systems. Each week, the federal government sends out $32 billion in Social Security and payroll checks and payments for such mundane items as rent. Even a short delay could be a major shock to the economy.
Unfortunately, when the Office of Management & Budget issues the next official report on the government's Y2K readiness on Mar. 15, the news is not likely to be good. On Feb. 6, the Defense Dept.'s Inspector General issued a report saying that the military has no assurance that it is purchasing Year 2000-compliant products, "which may seriously hamper the ability of DOD to perform its administrative and warfighting mission requirements." The Internal Revenue Service is struggling to meet the deadline, as is the Health Care Financing Administration (HCFA), which handles the enormous flow of Medicare and Medicaid funds. In the worst case, the tax system and the Medicare payment system will experience severe disruptions in 2000, halting the delivery of refund and reimbursement checks. "What is going to happen is the government is going to say, `It's O.K., trust us,"' says Richard M. Kovacevich, CEO of Minneapolis-based Norwest Corp. "It could get scary."
NOT LOSING SLEEP. To be sure, there are some, especially in Silicon Valley, who aren't concerned. "Frankly, I don't lose any sleep on worrying about being Year 2000-ready," says Eric Benhamou, CEO of 3Com Corp., the large network-equipment maker. Adds Tony Hampel, group manager for Year 2000 Marketing at Sun Microsystems Inc.: "Year 2000 is an annoyance, a speed bump. We're overassessing the end-of-the-world aspect of the Year 2000 problem."
Perhaps. But other knowledgeable people are more worried about the widespread economic impact of the Y2K bug, including the Federal Reserve. "The failure to get it right will affect the integrity of the payment system, financial markets, and the performance of the domestic and the global economies," said William McDonough, president of the Federal Reserve Bank of New York, in a recent speech.
How can a single computer bug turn out to be so potentially hazardous to the economy? In theory, any single instance of the 2000 bug is relatively easy to repair. But the sheer number and complexity of computer programs that use dates turn the process of fixing the programs--also known as "remediation"--into a herculean task.
Consider a household analogy. If you have to change one lightbulb or one electric switch in your house, there's no problem. But if you have to replace the lightbulbs and electric switches in everything you own, including your refrigerator, your car dashboard, and your furnace, it becomes a very time-consuming and expensive process.
Indeed, what makes Year 2000 particularly vexing is that it affects both stand-alone computers and the embedded processors built into all sorts of modern equipment, from automated factory equipment to power plants to cars to cellular telephones. Last fall, Phillips Petroleum Co. engineers ran Year 2000 tests on an oil-and-gas production platform in the North Sea. The result: In a simulation, an essential safety system for detecting harmful gases such as hydrogen sulfide got confused and shut down. In real life, that would have rendered the platform unusable. Similar problems can occur in almost any sort of modern manufacturing that involves sensors and "smart" machinery. "There will be facilities where they go in and turn on the machines and they won't go on," says Dean Kothmann, head of the technology division at engineering firm Black & Veatch, the world's largest provider of power plants.
In particular, electric utilities are only now becoming aware that programmable controllers--which have replaced mechanical relays in virtually all electricity-generating plants and control rooms--may behave badly or even freeze up when 2000 arrives. Many utilities are just getting a handle on the problem. "It's probably six months too soon for anyone to try to guess the complete extent of the problem," says Charlie Siebenthal, manager of the Year 2000 program at the Electric Power Research Institute, the industry group that serves as an information clearinghouse. "We don't know" if electricity flow will be affected, he said.
Nuclear power plants, of course, pose an especially worrisome problem. While their basic safety systems should continue to work, other important systems could malfunction because of the 2000 bug. In one Year 2000 test, notes Jared S. Wermiel, who is leading the millennium bug effort at the Nuclear Regulatory Commission, the security computer at a nuclear power plant failed by opening vital areas that are normally locked. For that reason, the NRC is in the process of issuing a letter requesting confirmation from utilities that their plants will operate safely come Jan. 1, 2000. Given the complexity and the need to test, "it wouldn't surprise me if certain plants find that they are not Year 2000-ready and have to shut down," says Wermeil.
By contrast, the securities industry, big banks, and the Federal Reserve have been taking Year 2000 seriously for years. Chase Manhattan Corp. is spending $250 million on the problem, while Wells Fargo & Co. will ultimately deploy 400 people to fix Y2K. Later this year and next year, Wall Street firms will run an industrywide test simulating the rollover to Jan. 1, 2000.
But smaller banks are lagging way behind. According to an extensive new survey from accounting firm Grant Thornton, community banks are only spending an average of $7,000 each on the bug. "They are not taking it seriously enough," says Diane Casey, national director of Thornton's financial-services practice. Only 44% have tested their vaults and other time-sensitive security systems for Year 2000, raising the possibility that they will either be locked out or that doors and vaults will spring open on New Year's Day, 2000.
PREOCCUPIED IN ASIA. These smaller banks, which typically outsource check processing and other computer-dependent operations to outside service providers, are assuming that those companies will handle the problem. But bank-service providers themselves are faced with the tough prospect of fixing their own systems. If outside vendors don't get the problem corrected, "banks are then in a real bind," says Mark L. O'Dell, director of the bank-technology division at the Office of the Comptroller of the Currency.
Another danger to the U.S. financial system comes from abroad. European and Asian banks are far behind their U.S. counterparts. In Europe, banks are focusing on the enormous task of converting their financial programs to handle the euro, rather than fixing Y2K. "Western Europe will not achieve Year 2000 compliance for even 65% of the applications that need Year 2000 repair," says software expert Jones.
And the situation in Asia may be even worse, since banks in those countries are more preoccupied with surviving the current economic threat than worrying about Year 2000. "Y2K could trigger a whole new round of country debt negotiations," warned Philip Kozloff, a member of Citibank's Credit Policy Committee, speaking at a Securities Industry Assn. conference in January. Adds Yardeni: "If foreign banks fail to comply, that could have a serious impact on world trade."
Closer to home, state and local governments are lagging even further behind the federal government. In Massachusetts, for example, according to a report released on Feb. 3, more than 40% of state agencies had not yet even begun an effort to become Year 2000-compliant. "It is likely that certain systems will not attain Year 2000 compliance in time and that alternative processing methods will be needed," wrote State Auditor A. Joseph DeNucci, whose office prepared the report.
Massachusetts isn't alone. Only about one-third of states are in decent shape, estimates Larry Olson, chief information officer of Pennsylvania, which got an early start on fixing Year 2000 problems. "Everybody else will have significant problems." What's more, the problems reach down to local levels, including school systems, which must run payroll systems, track student records, and operate other date-sensitive systems. The fixes could be quite expensive: The Columbus (Ohio) school system, with 65,000 students, may have to pay $8 million to fix its most significant applications.
"NO SILVER-BULLET SOLUTION." Is there any chance of avoiding major economic upheaval from Y2K? Yes, but it will take a huge effort. Smaller companies and organizations still have time to fix their typically less complicated computers and software. And utilities and manufacturers with embedded chip problems can fix them--if they start right now. "It's not something they can study for a year," says Kothmann.
But a large company or organization that has not yet started dealing with Year 2000 is in big trouble. While software tools do exist for identifying trouble spots and speeding fixes, they only solve part of the problem. "There is no silver-bullet solution," says Jean-Charles Andre, director of Year 2000 Services for IBM Global Services.
In particular, no software tool can eliminate the need for testing Year 2000 fixes once they are made--and testing a complicated software program is so time-consuming that it may prove to be one of the biggest barriers to finishing on time. A large company that has not started making fixes will have great difficulty completing tests before the deadline, according to Jones. The same is true for governments: "If you're not well on your way now, you just don't have time to do everything," says Olson.
Companies and other organizations may also have trouble finding the trained workers they need to fix their Year 2000 woes. A just-released study from the Information Technology Association of America shows that the U.S. already has 350,000 job vacancies for computer scientists and programmers that are not getting filled. "The big issue for corporations is going to be around resources," says Cowie. "If you need to change your systems, and you have to get programmers in to do that, it will be very hard."
To be sure, farsighted companies have up to now solved their Y2K problems by simply installing new software and hardware. But such major productivity-enhancing projects take even longer than fixing old software. "Until now, many people have been ripping things out, installing new software," says William T. Ruckle, managing director of Ernst & Young's Year 2000 business. "It's getting too late to do that." Adds ABB's Barnevik: "Some of the biggest problems will come when people have planned to substitute new systems for old ones. If there are problems, it will be too late to rebuild the old system."
What happens next, aside from everyone working very hard to fix the problem? Some organizations that have their Year 2000 bug under control are getting ready to turn it into a competitive advantage. "Solving this problem gives us a marketing tool that we plan on using," says Pennsylvania Governor Tom Ridge.
Most managers, though, are going to have to start thinking about contingency plans and workarounds. "There's no question that things will fail," says Judith List, who runs the Year 2000 effort for Bellcore, which develops communications software. "What Bellcore recommends is that you focus on the mission-critical systems first."
Year 2000 is a unique and unprecedented economic event. Previous jolts to the economy, like the gulf war and the oil price shocks of the 1970s, were surprises. But Year 2000 is the first economic disaster to arrive on a schedule. The great danger is that the U.S. economy will not be ready to welcome the millennium.