GS: Global Themes: Revisiting 'Tech's split
personality'
Several important tech developments have occurred around the globe over the past week.
It is important to consider everything from US CIO spending plans to Asian technology
margins. There are investment implications for semi equipment, semis (DRAMs in
particular), and LCDs. It is also timely to revisit the 4Q tech trade idea following the 11%
rise in the GSTI since September 30th.
Consider the following points:
**Semi equipment orders are being pushed out by semi customers and panel makers
alike. With utilization at Asian foundries declining (TSMC now expects 4Q utilization to
be only 85%) and multiple LCD makers pushing out next generation plans in response to
plummeting prices, new equipment purchases have now turned and 4Q04 and 2005 order
guidance is being reduced across most equipment suppliers. Most notably, Applied
Material, the industry bellwether, finally broke rank and capitulated on guidance,
pointing to a 35% QoQ drop for 4Q04.
**LCDs prices have collapsed much more rapidly than expected. For example, 17'' panel
prices have fallen from $294 in spring 2004 to only $165 now. Similar declines have
occurred at 15'' and 30'' sizes as well. Critically, this is causing cash losses for some 2nd
tier Taiwanese LCD makers. See Exhibit 1 and 2.
**DRAMs have become very profitable, in contrast, for at least two reasons. First, as
other PC component prices have fallen (LCDs in particular), there has been more room
for memory content in "the box." Second, notebook (NB) demand has been quite robust,
tracking at an above normal +22% QoQ increase for 4Q. With this backdrop in mind,
some Taiwanese pure-play DRAM makers are now earning well over 40% margins as
seasonal demand reaches a peak. See Exhibit 2.
**Our tech team's newest CIO Survey still points to rather anemic corporate IT capital
spending growth for 2005. Though results of the survey did move up slightly from the
August reading, the forecast is still only for 2.5% growth. This remains well below
pre-bubble trend and, more importantly, below current bottom-up consensus revenue
estimates for 2005, which are currently modeling over 9% growth.
Investment implications: With these points in mind and based on our analysts' work
around the global sector, we see 5 investment conclusions.
1. CONSIDER SELECTIVE PARING OF THE TECH TRADE. As we pointed out in
our "Technology's split personality" report in early October, the typical 4Q tech rally has
been concentrated between early October and late November, with December actually
losing ground (see Exhibit 3). We're thankful for the 11.4+% gain in the GSTI over the
past 7 weeks, and would think about carving positions post Thanksgiving. This +5.8%
relative gain (the S&P 500 is up 5.6%) is almost exactly 'normal'; the 1994-2003 average
Oct-Nov outperformance was 6.2%. There are certainly tech ideas we like, as mentioned
below; but equally, there are segments of tech we would avoid, especially given the lackluster
outlook for next year.
2. ONE AREA OF CONCERN IS DRAMs. Based on our global semi analysts' (Andrew Root,
Helen Huang, and Matt Gehl) work, DRAM fortunes appear ready to reverse. At 9% of the cost of
a PC, memory/box is full; seasonal demand is set to slow; NB upgrades have moved down the
corporate priority list according to the GS survey; and capacity will dramatically expand in the
1Q/2Q of 2005. DRAM and NAND are increasingly intertwined, and the team is modeling
oversupply for both. NAND prices have already started to fall, and DRAM should follow shortly;
a 30% decline is expected for next year. It is hard to imagine the DRAM margin in exhibit 2
staying at cyclical peak levels. Stocks such as Micron, Infineon, and Powerchip are all at risk in
this scenario.
3. SEMI EQUIPMENT GROUP ALSO DESERVES CAUTION. Jim Covello expects quarters of
negative growth as current excess capacity is worked off in the semiconductor industry. Moreover,
current DRAM spending plans look suspicious at 5-10% growth. Based on our semi teams' view on
memory, if pricing rolls over, spending plans will likely decline as well. This may catalyze the next
leg down for semi equipment orders. AMAT, ASML, and Tokyo Electron are all susceptible.
4. ARE YOU A CONTRARIAN? IT'S TIME TO REVISIT A PACKAGE OF LCD STOCKS. As
mentioned above, panel prices have collapsed, and elasticity of both demand and supply are
kicking in. Matt Fassler's time-lapse survey found a 3-4% reduction in just the past 3 weeks at Best
Buy and Circuit City, LCDs are becoming more competitive vs. plasma TVs, and anecdotal
evidence suggests consumers are buying TVs. On the opposite end of the supply chain, Asian
manufacturers are responding to lower prices by cutting 2005 expansion plans, especially on for
6G and 7G fabs. As such, supply and demand should be better balanced; Helen Huang and Yuji
Fujimori's newly adjusted model forecasts 10% oversupply in 2005, down from a 21% excess
previously. An even better balance may result with potentially greater supply restraint. As seen in
exhibit 2 LCD margins are nearing trough levels that were reached in late 2002. A similar trough is
very plausible now. With this in mind, Chi Mei Optoelectronics and LG Philips LCD have been
upgraded to Outperform and Nitto Denko has been added to our Asian Current Investment List
(CIL). AU Optronics, Sharp, Nippon Electric Glass, Corning, and Merck KGaA are all also part of
the theme. Given the market caps and volatility involved with many of these stocks, this theme is
conducive to a basket purchase.
5. LASTLY, STOCK SPECIFICS ARE KEY FOR 2005. Given the challenging outlook expected
for 2005, we remain focused on a core group of company-specific ideas. Low cost position, market
share gains, spending priorities, non-Tech drivers, dominant positions, etc... are characteristics
we're looking for in order to transcend the hurdle of a feeble tech environment. In addition to the
LCD ideas above, Dell, Symantec, Affiliated Computer Services, Paychex, eBay, Ericsson, and
Sony are the stocks mentioned in our Oct 4th 2004 report.
I, Tim Craighead, hereby certify.. |
