SURVEY THE MILLENNIUM BUG
THE most predictable thing about the Year 2000 problem is the uncertainty it will generate. That uncertainty will spread as the millennium approaches, and could easily tip into panic. There is little doubt that this will have a palpable economic effect in the second half of next year. Harder to predict is the economic impact of the disruption that the bug will cause.
Most of the numbers on the prospective damage are actually estimates of what is being spent on debugging. Those are difficult enough. In some countries, tax breaks for Year 2000 investments are an invitation to push through lots of other software spending at the same time. Conversely, some companies that are doing their own repair work may find themselves understating the costs if they simply take the money from their general information-technology budgets.
Two much-quoted estimates for the global cost of software repairs, calculated in different ways, come to similar orders of magnitude. Thus Software Productivity Research comes up with $530 billion (plus just over $1 trillion to test, retest, sort out embedded systems and pay off the lawyers), whereas Gartner Group suggests a range between $300 billion and $600 billion. The eventual figure may be lower: the Federal Reserve guesses that it will cost rather more than $50 billion to sort out Americas private sector; and the United States accounts for around one-fifth of world output but one-third of its computer base. If the Feds estimate is right, the eventual global figure, at least before lawyers bills, might be $200 billion-$300 billion.
These are daunting sums (though bear in mind that total IT spending between 1997 and 2000 is likely to be around $1.6 trillion even without those extra costs). What effect will they have on the broader economy? Many companies feel that Year 2000 spending is money down the drain: the equivalent of spending to meet pollution standards, say, the main virtue of which is to buy protection against lawsuits.
In fact, spending on bug eradication is of two sorts. Where work on the problem began early enough, lasting gains have been achieved by replacing ancient software and equipment. But much work, such as the adjustments made to old software code, will merely buy protection from disruption.
For the consultants who do the eradicating, spending of either kind is equally welcome. But this is one of those instances where, because of the way countries measure GDP, a disaster appears to stimulate growth when all it really does is stimulate activity to limit the damage. When the Exxon Valdez spilt oil in Alaskas Prince William Sound, the consequent clean-up created a local boom. In the same way, spending on bug protection, although it will boost GDP, will not improve long-term economic growth. Indeed, it will slow productivity growth, because workers are spending more time on tedious tasks and tests instead of thinking up new ideas.
Investment in new software (which in most countries still perversely appears in the national accounts as a cost of production) should certainly increase productivity and make countries richer, but much of the investment triggered by the Year 2000 problem is not necessarily additional; it would probably have been made anyway and is simply being brought forward. It just looks bigger because salaries for IT staff are currently being inflated by bug-induced demand.
The overall economic effect of bug-squashing will depend on which kind of spending predominates. As the millennium approaches, the balance is likely to tilt in favour of the down-the-drain sort, probably causing a small reduction in economic growth. The Fed expects about a tenth of a percentage point to be shaved off American GDP over the two years to the turn of the millennium. HSBC, a bank, reckons British GDP will be slowed by around 0.5% this year, but not at all next year.
Returning from the party of the century, will people face a prolonged recession inflicted by the bug? Mr Yardeni at Deutsche Bank Securities thinks so. But he freely admits that his forecast of a 70% probability of a bug-induced post-millennial recession is not based on any economic model. Its a way for me to convey my concern.
Others take a more sanguine view. Americas Council of Economic Advisers, looking for a precedent, has combed through a disheartening list of strikes, meteorological catastrophes and big technical failures in the past and concluded, unexpectedly, that such events hardly ever show up in GDP, even during the quarter in which they occur. When the workers at United Parcel Services went on strike last year, most of their work simply passed to other shippers. The devastating power cut in Auckland earlier this year, which was widely expected to take 0.3% off New Zealands GDP, did not even cause a blip in the statistics.
The sheer scale of the Year 2000 problem may make it different. But how different? Economists at ING Barings tried to model the effects of post-millennial glitches on the Netherlands. They made a number of assumptions to simplify the calculations, such as that most countries with strong links with the Dutch economy will have reached a similar state of readiness (unlikely); that one-fifth of all businesses will lose two-months-worth of production, half of which will be made good later in the year; that production problems abroad will reduce Dutch exports by two percentage points; that private consumption will drop by 0.25%; and that demand for IT and similar services will enjoy a boost. They concluded that, on these assumptions, final demand in 2000 will be reduced by 1.4%, with the greatest loss suffered by exports.
That would hardly amount to a recession. But if it happened, it might come at a bad time, when global demand may already be floundering, and when the end of the millennium will rattle nerves anyway. Just to add to the general sense of doom, there may well be signs in the sky: the Leonides, the largest shower of meteors for 30 years, is due in November 1999, and may knock out a few communications satellites.
However, a panic may also create a frenetic little boom of its own. Many companies and individuals will stock up to prepare for bug-related problems. Firms worried about the reliability of key suppliers will buy in stocks of components, and bring forward projects into 1999. Businessmen nervous about flying after the turn of the year will take their trip in late 1999 instead. As for consumers, in many countries they already stock up before Christmas or the New Year. In December next year, they will fill up their freezers, cars and walletsjust in case.
That, together with all the spending on millennial celebrations, will create a frenzy of activity in the final months of 1999. The first months of the new century will be marked by a giant hangover, depicted in chart 9 on the previous page. To some extent, that too will be a matter of timing: all those stocks laid in before the event will have to be used up before companies and consumers spend any more. But two other factors will also be at work.
First, glitches will undoubtedly cause some loss of productivity. Much of it may be made up later, but not all: a newspaper unpublished, or a flight cancelled, will be lost for good. Second, confidence will be affected as the full extent of the Year 2000 problem becomes clear and media reports seize on the more sensational aspects. Lots of things are bound to go wrong, and even in the absence of any big disaster anything without an obvious cause will be blamed on the bug. That will instil caution in companies and pessimism among individuals.
So, even though the millennium bug is unlikely to cause a global recession, it will certainly aggravate a downturn. Its direct effects will be unpleasant in many ways. Around the world, it will disrupt power supplies, telephone services, transport and hospitals. It will drive some companies into bankruptcy. It will kill people, perhaps just a few, but conceivably several thousand. Coming at the end of a century in which wealth and living standards have improved so dramatically, it will be a sharp reminder of just how much those improvements depend on technology; and how easily that technology can go wrong.