 Print-friendly version Send this to a friend Posted 9/20/2004
Contrarian Chronicles
About Contrarian Chronicles
Learn the Contrarian Chronicles lingo
Subscribe to Market Rap on Fleckenstein Capital
Related Articles
Chip glut troubles even the Dead Fish
What Intels 5 big problems mean for tech
Intel's twin woes: excess capacity, slack demand
Contrarian Chronicles
Recent articles: The 'recovery' is living on borrowed time, 9/13/2004 'Buy and hold' can mean 'hold and lose', 8/23/2004 The great tech recovery is coming undone, 8/16/2004 More...
| | Contrarian Chronicles Under techs hood, things dont look so good
The chip industry says the soft patch is just a one-quarter event. But if 'absolutely everyone' is saying the same thing, odds are the problems are deeper than that.
By Bill Fleckenstein
Let's jump right into the connect-the-dots department for a look at my recent favorite subject: chip land and its fabled one-quarter problem.
I'll start with two questions: If nearly every company in an industry announces it has a problem, is that problem less or more likely to be solved in a single quarter? And, is it less or more likely that it's something greater than a minor transitory issue?
Now to hone in on specific names and look at how their news undercuts the "one-quarter" thesis: Beginning with last Tuesday night's data, Xilinx (XLNX, news, msgs) said that it has almost two quarters of inventory built up. Obviously, it would be nearly impossible for it to work this off in just one quarter.
The semiconductor industry often experiences periods of double and triple ordering. That's one reason why lead times explode, and it's the fear of the explosion in lead times that reinforces the double and triple ordering. This fact alone suggests that the problem will not be solved in one quarter.
Meanwhile, to further compound Xilinx's problems, Celestica (CLS, news, msgs) said that it saw cutbacks, most likely from Cisco Systems (CSCO, news, msgs). Cisco is Celestica's biggest customer and also a very large Xilinx customer. (Celestica builds products for Cisco, using parts from Xilinx and others, and said on the call that its inventories were going higher this quarter again, since the deterioration just accelerated.) So, the possibility that Xilinx will get out of two quarters' worth of inventory this quarter when its major customers are cutting back doesn't seem very likely.
In addition, Xilinx will probably have to cut its wafer starts at the foundries. (Think United Microelectronics (UMC, news, msgs) and Taiwan Semiconductor (TSM, news, msgs).) Now if this goes on for very long, the foundries will need to re-evaluate their capital-spending plans. That's how this news will ultimately affect semiconductor-capital-equipment companies.
If this were just one isolated incident, that forecast for the equipment arena might be a bit of a stretch. But besides Xilinx and Celestica, let's take a look at some other companies that have pre-announced, which also happen to count Cisco as their largest customer: Altera (ALTR, news, msgs), Cypress Semiconductor (CY, news, msgs) and Integrated Device Technology (IDTI, news, msgs). In all probability, Cisco is cutting back orders to the vendors it buys parts from in addition to its contract manufacturer. Cisco has excess inventory on its books, and it is seeing orders that are smaller than expected.
When I first got into the investment business 25 years ago, I remember some company that I cared about describing a problem as confined to one quarter. Being green, I naturally believed them. It turns out that their problem wasn't a one-quarter problem. In my entire career, I do not ever recall something that was supposed to be a one-quarter problem actually turning out to be a one-quarter problem.
Nevertheless, a one-quarter problem is the present battle cry. For the "other people's money" crowd, it's got a clear-cut urgency. As a friend recently noted: "The arrogance of the behavioral finance (i.e. price action is everything) crowd is getting thick. Do they think they can rent positions and engineer rallies to get paid every year?" Yes, it appears they do think that.
Whitewater for Big Blue? We could go through this same analysis on the IT side. Quite likely, Celestica is also referring to IBM (IBM, news, msgs). (Big Blue, I believe, is uniquely positioned to begin hitting on no cylinders, but that's a story for another day.) While IBM tends to make more of its own product than do some of the other companies, the point survives that this is more than a one-quarter problem. Note that Celestica said some of its largest customers, implying that it was more than just one each.
This sort of digging can only lead to conjecture, not certainty. But just from what happened Tuesday night, one can start to see that this problem, in all likelihood, will be far more severe than the whistling-past-the-graveyard crowd had wanted to believe. There are too many problems in too many places for this to be a short-term problem.
Intels heartburn Intel (INTC, news, msgs), where the problems are even more severe, has an estimate of approximately the same. Normally, the third quarter is a pretty good period for Intel on a seasonal basis, but I think we're liable to see earnings in the mid-20 cent region. (Editors note: Zacks says analysts are expecting about 27 cents per share.)
But we can't forget that in Intel's case, they have an absolutely huge amount of inventory on the balance sheet. When they start selling that inventory, it will put tremendous pressure on Intels average selling prices. That will therefore hurt their revenues, therefore hurt their margins and therefore hurt their earnings.
Down the road, there's always the possibility that part or all of the inventory will have to be written off. That will not be an "excludable" one-time charge. It will flow through the income statement.
The yoke of AMD and overcapacity Intel also has serious competition from Advanced Micro Devices (AMD, news, msgs), something I have been discussing regularly for a year now. While many folks have not been paying attention, AMD has made serious progress, in terms of taking market share away from Intel, largely because it has built a better mousetrap. Intel has been trying to milk the same, tired X86 structure for well over 15 years now, and it has been swooped by AMD. (Check BusinessWeeks article on AMD: Suddenly -- It's AMD Inside.) The story chronicles the knockout punch that lies ahead for the onetime champ.
Intel has the problem selling into an end market that's totally saturated. It also has that inventory problem. That suggests it's suffering from too much capacity. (Running those factories at lower volume will really hurt margins.) Meantime, it has a competitor with a better product. Not exactly something you would want to buy at 20 times an estimate that you can't hit.
I think in Intel's case, 80 cents for next year will certainly turn out to be a better bet than current estimates of around $1.20. But it's liable to be much lower. In another quarter or so, Id guess we'll be able to make some better estimates.
For now, people have rather gleefully made the bet that, because bad news on the earnings front is so prevalent, this must mean that it's the bottom again. I continue to feel this is a dangerous conclusion. Yet, every day that the market shrugs off bad news, the bad-news bulls only get emboldened. I think they have started to play this gambit a little too soon, as they're liable to get a continuation of unfriendly news clear through earnings season. Therefore, they have gotten themselves all lathered up at a very poor moment in time.
Bill Fleckenstein is president of Fleckenstein Capital, which manages a hedge fund based in Seattle. He also writes a daily Market Rap column on his Fleckensteincapital.com site. His investment positions can change at any time. At the time of publication, Bill Fleckenstein was short Altera and long Altera puts; long Cisco puts; short IBM and long IBM puts; short Intel and long Intel puts; and short Xilinx and long Xilinx puts. Under no circumstances does the information in this column represent a recommendation to buy, sell or hold any security. The views and opinions expressed in Bill Fleckenstein's columns are his own and not necessarily those of MSN Money.
|