Y2K – What went wrong?

This is an adaptation of a talk I gave to the New South Wales branch of the Australian Skeptics on February 5, 2000. It was a follow-up to the address given in November 1999 and was publicised before the end of 1999 using the theme – "See him grovel or see him gloat".

I'm going to cover five areas here – what went wrong, how much was spent on Y2K, how much expenditure was predicted, what the prophets are saying now, and how this all relates to skepticism. I will stick to the computer issue, because nobody except lunatics expected the religious millennialism to have any effect at all.

What went wrong?

Surprisingly, not much at all. Even someone as skeptical of the hype as I was was amazed at how little went wrong. I monitored the Australian government's alert line from midnight on the fateful day and the messages became more and more desperate as nothing happened. At one stage four countries were declared to be on amber alert

• GAMBIA 31-Dec-1999 10:59 GMT Govt Svcs
• NICARAGUA 31-Dec-1999 01:39 GMT Water
• NIGERIA 30-Dec-1999 23:54 GMT Health/ Hospitals
• SRI LANKA 31-Dec-1999 11:11 GMT Air Trans

As you can see, these are major industrialised nations where problems could initiate a domino effect across world financial markets or maybe precipitate nuclear war. Some hours later, when it was reported that only Gambia and Nicaragua were still at risk, I felt much more confident.

When other reports of problems started coming in, I became afraid again. An ATM in London rejected some credit cards. An EFTPOS machine in Coles in Melbourne did the same. A water authority somewhere issued invoices with an incorrect date for payment on them. A video store in Adelaide told customers that their videos were being returned a hundred years late (If I ran a video store, I would have done this deliberately to get myself some free publicity). Worst of all, the Western Australian birth registry people issued some incorrect birth certificates.

The first crash of a financial institution in Australia was reported in the first week of February 2000. Monthly statements from Diner's Club are usually listed in date order from oldest to most recent, but statements issued during January were only sorted by two-digit years and showed January charges before December 1999. Is it any wonder that interest rates had to rise, precipitating a stock market crash that brought Telstra shares free-falling back to what people paid for them in the first place?

If I sound cynical, I apologise. I meant to hide it better. Things could have been very bad indeed. We have already had two plane crashes this year big enough to make the evening news, but imagine how it could have been if Mobil hadn't put the Araldite in the petrol to save us from participating in a plague of plummeting Pipers. Some things have got better, though. We had lots of train derailments around Sydney in December but only one so far this year. We have also only had one IOC corruption scandal and two ticketing fiascos (but only one of them from SOCOG).

Since the start of the year, several more significant dates have passed without problems. There was the 3rd and 4th of January when people returned to work and switched their computers on, the 31st January when billing systems had to process the first end-of-month for the millennium, and, of course, 2/2/2000 which was not only the first date where the day equalled the month but the first all-even date since 8th August 888. Of course, 11 days were lost to the calendar since that date, so maybe we should adjust for that and worry about the 13th instead.

How much was spent.

I have an official Y2K participant's mug. It has a picture of Ned Kelly on it. I would like to start off by saying that a lot of people did very good work in inspecting and modifying computer systems during the time leading up to 31/12/99 and there is no doubt that the fact that I can joke about trivial problems now is in part due to this effort. A lot of money was spent on replacing equipment and software that was a few years old, but this should not be classed as Y2K expenditure – it was just the effective management and maintenance of assets that sensibly companies should always have been practising. I have no problem with money well spent. I just can't be convinced that the amounts of expenditure reported could be justified. In fact, I believe that we have just seen the largest fraud ever perpetrated on Australian taxpayers, consumers and shareholders. Someone commented to me the other day that it didn't matter how much was spent because it all added to a vibrant economy. Tell that to people who remember the houses of cards that poured money through the stock market in the early 1980s.

Estimates of expenditure in Australia range from $10 billion dollars to $15 billion dollars. The range of these estimates gives some idea of how reliable they are. The most commonly-quoted figure is $12 billion, which just happens to be 10 times the amount that Alan Bond stole, so it's not that much really, is it? Even the lowest figure suggested that no amount of damage that could happen after the start of the new year could exceed the financial damage that had already been done. The $15 billion figure came from a highly-respected business statistics and intelligence organisation and was strongly influenced by their prior estimate of remediation costs. More about them later. I suspect that nobody has any idea of how much money was really spent.

The big dangers to life, limb and society were going to come from the banks, the airlines, telecommunications and utilities. The only utilities I ever heard mentioned in NSW were electricity and water. For some reason gas was never going to be a problem in NSW although I believe Victorians thought differently. Let's look at water first. Water runs downhill, and I thought that that was unlikely to change with a change of date. The water that comes out of the taps at my place comes from a great big green tank on top of a hill in Baulkham Hills. When the level in these tanks drops, pumps are started elsewhere to fill them up again from other reservoirs, much like the way your toilet cistern refills itself when the level drops. I imagine that the tank supplying my place has been supplying and refilling more than usual during the very hot weather over the last couple of days, but that has been due to increased demand because of the hot weather, not because the dates or times are anything significant. Electricity is similarly driven by demand and again consumption over the last few days would have been a big spike because of air conditioning, pool filters and water pumping. Someone from one of the electricity companies finally snapped at a reporter's nonsensical questioning just before new year and told him that electricity would keep coming because they just didn't use transistors to switch 133,000 volts. Figures I've seen suggest that the water and electricity people spent about $300 million in testing and remediation, which is not a lot of money when you consider the sizes of the operations.

Banks are supposed to have spent about $1 billion between them. Again, this is not a lot when you consider the IT budgets of these organisations, but even then there is evidence of waste and extortion. I have a home loan and two car leases which the banks seem to have had no trouble setting up so that they ran across the end of time. In any case, the banks have all rewritten their major software applications over the last few years so no problems should have been expected. There were problems of management fear, though. I remember speaking to someone senior from ANZ who said that the minimum expenditure for the bank was going to be $110 million because they had 55 million lines of computer code and the consultants had told them that it cost $2 per line to fix things. This is a direct analogy to a plumber using the amount of pipe in your house as a basis for charging for fixing a leaking tap. If the consultants just locked themselves in a room for a year and played cards, the bank would have willingly paid an invoice of $109 million and thought they were getting a bargain.

At the airlines, Qantas spent about $200 million to find out that nothing serious was wrong. Ansett spent $45 million, which suggests that someone there was thinking.

The telecommunications industry is interesting because we only had one company in the business ten years ago. Total expenditure has been estimated at $1 billion, with Telstra accounting for $400 million of that. I don't have much problem with Telstra spending big because they had a lot of history to catch up on. I just wonder where the rest went. Why didn't anyone at Optus have the guts to say that, as every piece of equipment, every software program, and every procedure (either automated or manual) in the place was less than eight years old, why was any money being spent at all?

Looking at these big risk areas and making allowances for slack and waste, let's grant them total expenditure of $3 billion. That is still a long way from 10 or 15 billion.

What was predicted.

Imagine you see a child's toy ball on the floor and want to guess how much it weighs. You know a bit of information about it. It's obviously more dense than air so there is a low limit on the mass and you can also probably guess that it's not made of solid uranium. How accurate would you think your estimate of its mass would be given just those facts? Now I want you to think about how confident you would be of your estimate if instead of seeing the ball itself, I had just given you a black and white photograph with no background detail.

You might think that nobody would estimate anything with so little information, but that was about how much was available to people estimating what would have to be spent to find and fix all possible Y2K problems in Australia. This was a problem where nobody even knew what had to be looked for, let alone what sort of fixes would have to be applied. Nobody had even the slightest clue about how much remediation was going to cost, nor, for that matter, how much damage would be caused if nothing was done.

None of this stopped people from trying. One international business intelligence organisation estimated that preventive expenditure in Australia would be $20 billion dollars. This number was a total fabrication. It was plucked from the air and sent out to subscribers to their information service, with an invoice attached to it of course. The same company is the one who is saying that actual expenditure was $15 billion dollars and letting us know how lucky and efficient we were to get away with only that much. This figure is also ethereal, but had to be above other estimates of damage in order to prevent them from looking foolish. Being wrong by 1/3 is much better than being out by 100% when you are in the prediction business. In a delightful irony, these geniuses didn't know that Netscape Navigator has a bug in the date routines in the JavaScript programming language, so the date on their web site clicked over from 31/12/1999 to 1/1/19100. Remember that this is the foremost (and most expensive) Y2K expert outfit in the world. Laugh? I nearly wet myself.

Just in case you think I am generalising about people who dispense business research, I will tell you about the reliability of numbers from another international group. I collected and supplied Australian retail computer prices to these people for more than a year. I had a quite complex model which drew in the wholesale pricing from the largest distributors in the country and applied formulas which took into account different markups based on brand perceptions, recommended prices, etc. The figures were spot checked against retailers like Harvey Norman and also against prices available from manufacturers like Compaq, HP and IBM through their web sites and marketing information. I was told that my prices were way too high, and was given an example of a particular notebook computer which someone claimed to be selling for about $7000. I pointed out that this gave a gross margin of $117 and therefore was not a realistic price. Eventually I was told that they no longer wanted to pay me for wrong numbers when they could get accurate information for free. These people charge in the thousands of dollars per year for access to their pricing reports which are worth less than a free catalogue from Harvey Norman or Dick Smith.

What the prophets are saying now.

Most of them aren't saying anything at all. Like most of the great predictors of messianic events in the past, they have either become invisible or are claiming that their predictions were misunderstood. At ten minutes past midnight on the morning of 1/1 2000, I posted a message to the Australian Computer Society's Y2K mailing list which just said "I just turned this thing on and everything seems to be working." I was almost immediately informed that nobody had said that things were going to fail on 1/1/2000, it was all going to happen after the weekend when people went back to work. 30% of small business in Australia would fail during January, just wait until next week, and so on. Oh yeah? This was coming from people who had been claiming just days before that a large percentage of computers would not even turn on, let alone run. I later posted a joke newsflash that some of you may have seen, saying that Boris Yeltsin had resigned because his handlers were worried about his robotic components and that a plane hijacking had ended because the kidnappers did not want to be in Afghani airspace on 1/1/2000. Australia's foremost doom-and-gloomer accused me of passing on rumours when I should have been telling people how disaster was really going to happen on 31st January. It has been going on like this ever since, with new Armageddon dates coming and going with increased desperation from the prophets. As I have mentioned, we have made it through the end of January and the first day-equals-month date. We have 29th February, 3rd march, 4th April etc to go, plus 30th June, 31st December and any number of other magic dates. One pundit has said that Y2K problems may even go on for three years.

Some of the prophets have just taken the approach to history shown in the book 1984, where history is rewritten to suit the feelings of the day. These people have just reinvented themselves as Goods and Services Tax experts, despite GST having nothing to do with computers because it is an accounting problem. I have already been told that GST will cost more than twice what Y2K cost (which was immediately transformed into $24 billion dollars, which is 2 times 12). One supplier of multi-million dollar accounting software packages has already told its clients that implementing GST will take 3 to 4 months, with many consultants needed on site at all times during that period. This is despite selling the software in other countries where GST applies and having just extorted millions out of the same customers to look for Y2K problems.

Why is this relevant to skepticism?

Being a sceptic is not just never believing things without evidence or proof, it is asking that the possibility of proof exists. It's the opposite of gullibility. I resisted much of the Y2K hype for the same reason that I reject homeopathy, iridology, telekinesis, ESP and alien abductions. There is no mechanism of action. There is no way that the results claimed can happen given what we know about how the universe works. There is nothing that can stop water coming out of my taps while there is water in the reservoir and intact pipes between there and me. There is nothing in the date that can stop my phone working, but the bill might be wrong if I happened to be making a timed call at midnight.

We saw supposedly intelligent and astute business leaders conned because they would not apply critical thinking when makers of invisible garments came to offer to make them some clothes. We saw them pay protection money to magicians who promised to remove curses and bring the sun back from the eclipse. We saw them purchase talismans to ward off evil spirits. We saw them sacrifice shareholders' funds just because someone authoritative asked them to. Is it any wonder that these same people join pyramid schemes, or happily accept increased health insurance premiums to cover quackery, or consult feng shui charlatans about office design, or listen to pundits who can predict the stock market, or spend huge amounts of money on ground up weeds to treat disease?

Have we learnt anything from the Y2K experience? When I look at the Y2K scamsters repackaging their spiels by the simple process of sticking hand-lettered "GST" labels on them, I can only doubt that we have.

Notes for non-Australians:

• Ned Kelly is Australia's favourite outlaw and folk hero. He was hanged on November 11, 1880. It is a mystery why we revere someone who stole things and shot people, but at least he wasn't a sportsman. Yes, I do have a coffee mug with his picture on it, but I also have a Microsoft mug.
• Telstra is the largest telecommunications organisation in the country and used to be a monopoly. Optus was set up as the first competitor and had no customers before 1993.
• Alan Bond was an "entrepreneur" in the 1980s. At the time this talk was given he was serving a prison sentence for looting a public company. His management skills are amply demonstrated by the fact that he went broke running a brewery. In Australia!
• "GST" refers to a new Goods and Services Tax (or valued-added tax) which will be introduced into Australia on 1 July, 2000. Politicians have assured us that we will all pay less income tax, that the cost of living will fall, that companies will be more profitable, and that nobody will really notice anything different. Perhaps we should change the name of Australia to "Utopia". Airlines who confidently flew at midnight on 31/12/1999 will stay grounded during July 2000 to avoid winged pigs.
• The IOC is the International Olympic Committee and SOCOG is the Sydney Organising Committee for the Olympic Games. The corruption and incompetence in both organisations have been providing us with better entertainment than anything we might see at the real Games in September.
• Harvey Norman is a retail chain which claims to sell more computers than anyone else in the country.
• The ANZ is a very large bank which actually paid a feng shui "expert" for advice. On this advice, they renumbered the floors in their head office building and cut the corners off desks. People trust them with money.
• About 5000 piston-engined light aircraft were grounded indefinitely in January 2000 because Mobil made petrol that damaged the engines. Planes have only been using petrol for 97 years and it is easy to make a mistake with new technology.