T.K's Dec.10 report:
10 December 1998
Thomas P. Kurlak, CFA
First Vice President Semiconductor Industry
Irrational Exuberance
Reason for Report: Update
Investment Highlights:
· Despite popular opinion, we see no evidence of a broad-based
industry recovery.
Fundamental Highlights:
· PC seasonal pick up driven by retail sub $1000 models.
· Corporate PC demand not materially affected by Y2K.
· PCs not enough to drive overall industry recovery – only affects
36% of market.
Semiconductor Industry – 10 December 1998
What Has Changed – Irrational
Exuberance
Yesterday's New York Times market page stated that
semiconductor stocks rose amid optimism that the “recent
boom in chip sales” will continue. However, there is no
evidence of a broad-based chip industry recovery, let alone
a boom. As we recently reported, the industry's own
October trade data shows this.
Actual sales fell 11.4% in October from September. The
three-month rolling average increased 6% because of
dropping off the vacation month of July and adding
October. The October three month average is always up
from September for this reason. The actual always falls
because September is a five-week month. Year over year,
sales fell 2%. Compared to average monthly sales for all
of 1998, October was up a little vs. down a little last year,
but hardly a boom. And, that modest increase is due
entirely to higher microprocessor sales for the seasonal PC
pick up.
What has whipped the overall semiconductor group into
such a sharp price recovery in such a short time? With
industry sales still very weak and SIA data yet to reveal
any growth over 1997, what is the group's appeal to
investors now?
So far, there is no question that PC sales are up from the
first half, which is producing a recovery for
microprocessors. But what is behind this PC pick up? Is it
sustainable? Is another inventory correction coming? And
what about corporate buying and Y2K? Answers to these
questions help with analyzing PC driven Intel (INTC, B-3-
3-7, $119) and Advance Micro Devices (AMD, C-2-2-9,
$31), but what about the other 80% of the chip industry?
First, on PC demand, Merrill Lynch believes 1999 can be a
reasonably healthy year in units (up 15%) but average
prices will continue to decline as the low end mix rises. So
downward price pressure on PC chips should continue. In
Q4 of this year it appears that sub $1000 PCs will
comprise 41% of total PCs or 13 million (8 million
Celeron, 5 million AMD K-6) out of 31 million. This is
way above the 1998 average of 25% for sub $1000 PCs
implying that a major part of the Q4 PC sales pick up is
seasonal retail buying. This can go away in Q1.
Also, corporations are spending their PC budgets before
they are cut in 1999 due to lower profits and industry
consolidation. Profits are down or looking weaker in 1999
in finance, banking, consumer products, drugs,
manufacturing, oil, chemicals and many other large PC
markets.
Will the Y2K issue increase PC buying dramatically to
offset down corporate profits? It is not clear, but we see
companies such as Merrill Lynch solving Y2K without
much change in PC buying by simply testing existing
equipment, which overwhelmingly passes. Practically all
leading Pentium, Pentium II and AMD K-6 PCs are Y2K
compliant. If more horsepower is not needed then why
replace a perfectly good PC in 1999? It appears that
software fixes in mainframes to deal with Y2K are taking
up a lot of the IT budget. We would not be surprised if
corporate PC buys are flat or fall off in 1999.
What about the other 80% of the semiconductor industry
not directly tied to PCs where no broad recovery is
apparent? About 20% of that 80% is helped indirectly by
the seasonal PC recovery in Q4. Therefore, some fraction
of most major chip houses sees some Q4 pick up. But not
enough to make an attractive earnings recovery out of.
Indeed, many companies still see enough weakness in the
rest of their business to make a forecast of general
recovery hard to support.
We believe earnings prospects for 1999 still look
lackluster for most semiconductor companies. We also
believe the stock prices of most of these companies are
inflated by expectations of a new up cycle that does not
appear to be developing. Yes, pricing has stabilized due to
cut backs in production and capital spending but this has
no impact on demand. And while production is down,
capacity is not. Unused capacity is sufficient to allow for
an estimated 45% increase in semiconductor production.
Indeed, Intel is cutting its spending sharply now due to
excess capacity and smaller die sizes and this is being
repeated around the world. Ample chip capacity exists.
Spot chip shortages are not amounting to anything because
factories can just raise output. This has already become
evident in DRAMs with the Korean production increase.
|
