|
25 Nov 99
Dear David:
For what it's worth, I don't see Y2K getting serious until late in January. The economy is a marvelous thing with lots of momentum, and like a bull elephant in charge mode, just shooting out its heart won't bring it to an immediate halt.
Setting aside the wild-card possibilities, I don't see any way that the oil distribution / refining system isn't going to experience substantial shortfalls. The modern lifestyle is based on cheap, regular and adequate supplies of gasoline. Given that the oil crisis of '73 only represented a 7% actual shortfall in supply, I'm expecting serious disruptions to start showing up in the later half of January.
Given our dependence on foreign oil from Venezuela and Saudi Arabia, automotive parts from Mexico and Brazil, our involvement with Japanese banks, etc., I was particularly impressed by the testimony two months ago of Jacqueline Williams-Bridgers, the U.S. State Department's Inspector General, to the U.S. Senate's special committee on Y2K.
She said, "These assessments suggest that the global community is likely to experience varying degrees of Y2K-related failures in every sector, in every region, and at every economic level. As such, the risk of disruption will likely extend to the international trade arena, where a breakdown in any part of the global supply chain would have a serious impact on the U.S. and world economies."
My question for those who feel that Y2K will only be a bump in the road would be, "Which part of every is unclear?"
Given the current level of dependency on automobiles in our culture, we can't afford more than a percent or two disruption in overall gasoline availability without triggering adverse social reactions (those who were around in '73 know what I mean). The odds that the interuption in gasoline availability will be no more than a couple of percent isn't very good.
We'll know one of the key factors quickly enough. The railroads have announced that they're shutting down for 36 hours during the rollover in order to test their systems. If the imbedded chips in the switching system make it through, then I expect that we're going to be in decent shape overall.
If, however, we effectively loose the rail system because they can't trust the switching/reporting network, then the dominos are really going to fall. The majority of the electric plants out there use coal, and they only have room to stock about a week's supply.
How people will react in late January is a real guess. If they start hurting, they're going to want to blame someone for that hurt. That's the part that makes me nervous.
It's interesting to watch the dominos starting to fall already: Hershey, Multifoods, Volkswagen. Half a billion in losses in those three companies alone in the last month is telling, but it's the nature of the losses that's most illuminating.
A couple of years back, as Information Technology managers started to look seriously at Y2K, they had two choices: fix the old or create the new. Many a company made the reasonable decision to start from scratch and recode every critical system they had. Some like Boeng spent more than a billion dollars, and it looks like they're in the clear. That isn't surprising given the nature of their business and their capacity for taking on complicated technical projects that have to be completed on deadline.
Still, they had some interesting surprises. Like the test on the 757. They rolled the clock forward, and then went looking for problems. Two hours of checking didn't turn up anything, so the plane started to taxi to the runway. Half-way there, the computer shut down. It turns out that the bug had caused the computer's cooling fan to fail to come on. After running for two hours without a fan, the computer overheated and shut down.
Lots of other companies also took the second path. While this plan has a great potential payoff, i.e. the latest software running on the latest machines, there is one really big risk. If the software isn't completed on time, a situation known to occur on occassion :-), the company can find itself caught between an old system that will no longer work after the first of the year, and a new system that won't work yet.
Even worse, in trying to make the transition, they can corrupt their data base and be left with the added task of having to do a new physical inventory.
This is the sort of thing that's starting to show up. It's not being talked about on the evening Happy News, but that's hardly surprising. For example, Hersey's Y2K program just cost them their largest sales weekend of the year. You may not have noticed, but Hersey didn't make it for Halloween this year, and probably won't make it for Christmas. Lost sales are said to be in the $100M range.
Now, it's true that western civilization won't grind to a halt because of a lack of chocolate, but it's also true that there are a lot of human faces in pain disguised by a figure like that. The loss of one end-product manufacturer involves losses for the business that supply them, as well as the businesses which depend on their employee's paychecks.
Multifood is an example of a good idea that didn't quite work. Their business consists of distributing and vending food products. Their distributing software was way old, but the vending side of their business was more up-to-date. And so they decided to expand the vending software to include their distribution operations.
In practice, they discovered a few glitches. For example, their distributing division used the term "each" to describe a case of something, while their vending side thought "each" meant one unit. The result is that some folks who ordered a case of something got one bottle, while others who wanted a bottle received a case.
That lead to a host of unhappy customers, and a lot of time lost filling out return slips and processing additional shipments, but the real problem is that this fubar has rendered their inventory records worthless. Not only do they have to correct the software problem, but now they also have to rebuild their inventory records based on physical counts.
The upshot is that their latest 10-Q Quaterly report filed with the Securities and Exchange Commission is projecting a 25% drop in profits for fiscal '99, word of which led to a one day drop of $112M in the value of their stock.
Volkswagen Europe is another example of a company that took the second path. Fortunately, they finished their new software with two months to spare. The problem is that it takes a lot more than two months to test complex software, and to save time, they passed up the step of operating the old and new systems in parallel. The upshot there is that the new software corrupted their old database, and they haven't been able to ship parts for the last two weeks.
The best description I've heard recently is the comparison to Christmas tree lights; not the old series kind where the burning out of one bulb would shut down the string, Rather, it's more like the new kind that have a by-pass circuit in the base so that if a filament fails, the base shorts out to let the power keep flowing. The problem there is that the rest of the bulbs then have to deal with the fractionally higher voltage. As more filaments fail, the rest of the filaments have to carry the load, and tend to burn out correspondingly more quickly. In time, this leads to a cascading failure of all the lights left in the string.
While none of these examples puts the world at risk, each of them does put a bunch of people through a whole lot of grief. As light after light on the productivity tree winks out, the tension level rises.
And then we get to the IRS's assurance that they have their contingency plan in place. They expect to be able to issue 10,000 refund checks by hand per day. Given that last year they issued some 80 million checks, that should only take them some 36 years working at that rate. Faster if they authorize overtime :-)
How are people going to react if they're told to not expect their tax refund in this lifetime?
Anyway, just some thoughts on a rainy morning. It's going to be a very interesting year.
Take care old friend, and do stay in touch.
Walt
|