Millennium bomb could trigger bank crisis

Jon Moynihan on how computer failure could create huge liquidity problems - and give gold a boost

It has been called the millennium bomb, the year 2000 problem and the 00 issue, yet it could equally be described as a many-tentacled octopus, or many-headed hydra, because its threats come in such a variety of forms.

Some threats are real; others are, or might turn out to be, imaginary. Separating real from imaginary is important because what has already been established is that not enough computer programmers exist in the world to check, in time, all the potentially compromised software. Some 00 events are already occurring, as with the 104-year-olds who are this year receiving letters summoning them to primary school.

What kind of disaster can be triggered by a computer's failure to distinguish between, say, 1900 and 2000? First, it is important to understand that it's not just lines of software that are potentially compromised.

Back in the 1970s when computer chips were created, lines of code, written in languages such as Cobol, now seen as simplistic, were hard- wired bodily into computer chip circuitry.

Because, to save significant amounts of programming, the Cobol or other programmers would in many instances have left off the first two digits of a year, that same lacuna would often have been translated straight into the chip.

Thus, not only computer code, but also some computer chips, are potentially compromised.

Many companies will be comfortable that for the large majority of their software and chips, there is no year 2000 problem. However, the degree of confidence will not always be absolute.

For example, it might take a certain amount of sang froid to board one of those flights they are currently advertising which will, they say, fly you through time zones into the millennium at midnight on 31 December 1999.

Will the plane definitely keep flying as midnight arrives? How bad can it get?

There is feverish activity behind the scenes to try to ensure we avoid any of the catastrophe scenarios that have been identified. However, new threats continue to emerge. Here's a reasonably plausible one: the great banking liquidity lock-up of 1999.

The payment systems of banks around the world utilise millions of lines of computer code. Every day, thousands of billions of dollars pass through these payment systems, from bank to bank; from bank to customer; from customer to bank; and on through to other customers.

If, on 1 January 2000, the software employed by a bank to mediate these payments fails to recognise the date, and therefore fails to make one or all payments that day, those involved in those transactions don't get the money they were expecting.

It needs only a small number of banks to have an 00 problem in their software for a significant number of payments not to be made.

In the interbank money transfer system, a failure to transmit somewhere between 5 to 10 per cent of daily payments would quite possibly be enough to lock up the entire system - if you haven't paid me, I can't pay him.

Doubtless, all banks are striving hard to have this problem solved before 31 December. But many have not yet tested the robustness of their payments software to be entirely sure.

International payments systems are so intertwined that the problem could be precipitated by banks not in the developed world - where many have not yet examined their payments system software - but by those in lesser developed countries where resources to investigate such problems are very scarce, so even less has been done.

The problem extends beyond banks, to the companies with payment systems tied to banks, which themselves are responsible for large liquidity flows into and out of the banking system.

How many of these will fail to sort out their software in time? How many need to fail before the banking system locks up and a systemic liquidity crisis is created? How prepared will businesses be, in the run-up during 1998 and 1999, to take a chance on the fact that there won't be a liquidity crisis?

The question has pertinence because any company or indeed individual who decides that there could be a problem on 1 January 2000 would be most likely to react to such a problem by creating, and hoarding, liquidity for themselves.

In other words, companies and individuals concerned about this problem will start to amass cash and other liquid instruments against the potential of a short-term or even long-term systemic liquidity crisis.

If that were so, then we are back to the problem that if only 5-10 per cent of assets are liquidated - or withdrawn from the system and placed in cash, gold or whatever - there will again be the potential for a liquidity crisis.

At that stage, central banks will be faced with the decision whether to flood the system with liquidity (thus potentially threatening a wave of inflation and economic dislocation), or to let the affected banks fail.

Even this solution may not be available to them if the liquidity crisis gathers speed very quickly. And even with such actions, commerce could be crippled for days, weeks or even months.

As with all liquidity crises, the threat could be self-fulfiling. Fears that it could happen could ensure that it does.

Anyone believing that a liquidity crisis could occur will want to pop out and buy their gold coins before everyone else does.

As the price of gold goes up - possibly to astronomic levels - and local liquidity scares occur, the lines round the block at bank opening time start swelling.

The great liquidity crisis of 1999? Impossible, you may say. However, there is nothing particularly outlandish in the scenario outlined above, which is made all the more feasible by the enormous reliance payments systems in the 1990s now place on their computer software. The scenario may, or may not, happen.

What should be done? Preventative measures could be taken.

The Hong Kong Monetary Authority has made it a requirement for all banks under its supervision to have a millennium compliance programme in place by the end of June 1997, guaranteeing that they will have identified and sorted out all problems by the time 2000 arrives. Similar requirements have yet to be seen in other banking jurisdictions.

Companies, unlike banks, do not face any statutory requirement. Certification of millennium-worthiness, similar say to ISO-9000 certification or an S&P credit rating, has not yet emerged and is an option - but one that would have to be developed soon.

Banks and big corporations themselves should pay particular attention to their payments software and hardware in the run-up to the millennium. And for those who subscribe to disaster scenarios - perhaps a small punt on the price of gold?

Jon Moynihan is group chief executive of PA Consulting Group.

PA, in association with Taskforce 2000 and the OTI, produced the report, Defusing the Millennium Bomb, which tracks the progress of UK companies to achieving Millennium compliance. Copies can be obtained from Clare Fortune on 0l7l-333-5367