GS US SEMI EQUIP WEEKLY: USE RULE OF
THREE TO AVOID HEADFAKE RALLIES
Summary: (1) The Rule of Three framework is likely to help investors avoid headfake
rallies during the downturn, (2) Short interest ratios indicate that the Street isn't as bearish
as many would like to believe, (3) Looking forward, next catalysts are Intel mid-quarter
update on Thursday night and the start of negative preannouncements, which we expect to
begin shortly after Labor Day. We expect Intel to maintain their previous capex range of
$3.6 billion to $4.0 billion on the call despite negative Street chatter around their capex,
and (4) News, Events and Price Performance.
THE RULE OF THREE FRAMEWORK IS LIKELY TO HELP INVESTORS AVOID
HEADFAKE RALLIES DURING THE DOWNTURN. With several companies
acknowledging last week that the semi equipment cycle has weakened, we received an
increased number of questions from clients wondering whether it was time to begin to
"dip their toes in the water" on the stocks now that some of the bad news is out of the
way. While getting the first signs of management capitulation (at least from the back-end)
is one important criteria for defining the bottom for the stocks, we believe there is still a
lot more that has to happen for the stocks to find a true bottom. Most certainly, there will
be headfake rallies along the way to the bottom driven by seasonality, swings in investor
sentiment, and other short-term trading drivers. However, the semi equipment stocks
aren't likely to find a true bottom until the Rule of Three framework is satisfied. We first
laid out the framework back in March of 2003 in an effort to convince investors that it
was time to buy the semi equipment stocks as the cycle was nearing a bottom and
business was poised to improve. Ironically, we highlight the same framework today in an
effort to convey our belief that it is too early for investors to begin buying the stocks in
anticipation of the next cyclical upturn.
Below we provide a review of the Rule of Three framework; this framework has been
used repeatedly over the last 15 years by Goldman analysts covering other industries to
accurately predict cyclical bottoms. Given that long-term semi capex growth rates aren't
even in-line with nominal GDP growth (the CAGR of semi capex from the peak of the
1996 cycle to the peak of this cycle is 1%, assuming 2004 is the peak), we believe that
the Street will continue to increasingly treat the semi equipment stocks like any other
slow/no growth cyclical business. We believe these rules create a framework for buying
stocks in a cyclical business. The three rules are simple:
1. Are industry cash flow margins bottoming-out? 2. Have management teams
capitulated? 3. Is there a glimmer of hope for an upturn?
Rules One and Two define whether an industry has reached a bottom. Rule Three helps
to answer the question of whether an industry is starting to improve off of the bottom.
Rule 1: Are industry cash flow (i.e. EBITDA) margins bottoming-out? Rule 1 asks
whether or not the semi equipment companies have reached the pain threshold to stop
chasing unprofitable business for the purpose of gaining market share. DOWNTURNS
DON'T END UNTIL SUPPLY/DEMAND DYNAMICS GET BACK INTO
BALANCE. As long as semi equipment companies are generating significantly positive cash flow
margins during the early stages of a downturn, managements have incentive to try and chase
unprofitable business (i.e. win business by cutting price or giving away free service) and in the
process of doing so, prevent the industry from achieving that supply/demand balance and, in turn,
prolong the downturn. Cutting prices or giving away "free service" along with tool orders at the
end of a quarter, in order to hit order guidance provided to Wall Street, only serves to prolong a
downturn by giving customers more capacity than they need or capacity they don't really need at all
(but will take if they can get it very inexpensively). The last downturn didn't stop getting worse
until Applied Materials negatively preannounced a 30% order shortfall in January of 2003. Since,
Applied had negative operating margins (i.e. near break-even cash flow) in that quarter,
management appropriately refused to offer any more deals at the end of the quarter (given that to
do so would have required significant cash burn). We do not believe it was a coincidence that
Applied's order shortfall marked the bottom of the cycle (i.e. customers were no longer getting
more capacity than they needed so the supply/demand dynamic began to improve) and that 6
months later, the industry was in a tighter supply/demand situation and the next cyclical upturn
began.
For the reasons cited above, the single data point we will be watching most closely over the coming
quarters is whether or not management teams are offering deals to the chipmakers at the end of the
quarters. The more frequently we see that practice occurring over the next several quarters, the
longer the downturn is going to last. However, if semi equipment managements are very
disciplined and refuse to offer incentives to make orders at the end of the quarter, then this could
indeed be the abbreviated downturn for which many are hoping. In addition to the anecdotal
channel checks that we will do at the end of the quarters to gauge how disciplined the semi
equipment companies are, we will also watch the reported cash flow or EBITDA margins to gauge
how aggressively managements may or may not be in closing business at the end of the quarters.
Rule 2: Have management teams capitulated? This rule is more straightforward than Rule 1. The
idea here is that semi equipment companies have to get rid of "upturn" cost structures in order to
give investors the operating leverage they deserve during the beginning of the next cyclical upturn.
Only after industry managements capitulate and acknowledge that visibility for an upturn is
non-existent are they likely to appropriately cut operating expenses.
While we generally believe that semi equipment companies haven't increased their cost structures
as significantly as in previous cycles, there will indeed be costs that need to come out of the
models. For example, we need to look no further than the significant reverse operating leverage
that LTXX and CMOS showed last week when they lowered out-quarter revenue forecasts to
understand that cuts need to be made. In fact, LTXX already announced that they will be cutting
operating expenses in order to get their cost structure in-line with their updated short-term market
forecast.
Rule 3: Is there a glimmer of hope for a cyclical upturn? Rules One and Two tell us whether or not
an industry has likely reached a cyclical bottom. That's all well and good, but an industry can stay
at a cyclical bottom for some time. Rule Three helps us identify whether business upstream in the
chain has improved such that a cyclical upturn is possible in the coming quarters. We would look
to datapoints such as unsustainably low chip inventory levels, capacity utilization at the foundries
increasing or commodity chip pricing (i.e. DRAM or flash) improving in what is normally a
seasonally weaker period for fundamentals, as a sign that business could improve in the coming
quarters for the semi equipment companies.
SHORT INTEREST RATIOS INDICATE THAT THE STREET ISN'T AS BEARISH AS MANY
WOULD LIKE TO BELIEVE. Short interest data for all of the stocks in our universe (released last
week) contest the notion that everyone on the Street is negative on the semi equipment stocks. One
of the reasons we hear from investors who are looking to take overweight positions on the stocks
is that everyone else on the Street is negative, so the contrarian view (which often is the correct
view on these stocks) is to be overweight. While the entire Street is now aware of the softening
cyclical datapoints, it does not seem as if many short-term investors are positioned negatively on
the group. The short-interest ratio on most of the stocks in our universe actually declined
meaningfully for the period ended Aug. 15th. Short interest for KLAC (-13% m-o-m), NVLS
(-11% m-o-m), LRCX (-10% m-o-m), and TER (-8% m-o-m) all declined meaningfully
month-over-month. In addition to the m-o-m declines, we would also highlight that short-interest is
significantly lower than at most times over the last several years. For example, NVLS short interest
has only been lower than this month's 1 other month in the last 4 years. In addition, KLAC
short-interest is 37% lower than i
t was in Aug. of 2003, despite the fact that that period marked the beginning of a cyclical upturn
and we believe we are now at the end of the next cyclical upturn. AMAT's short interest, while up
7% m-o-m, is also down 10% y-o-y, which doesn't make sense to us given the negative cyclical
datapoints that are so prominent today. Given that stocks have held-up relatively well in the last
several weeks since the data was taken, we believe that the short interest may even be lower today.
We have three reasons why investors aren't more negative: 1) Investors are afraid to short stocks
that have declined so much YTD (see performance chart below), 2) Investors are afraid to short
stocks ahead of seasonal strength upstream in the supply chain, 3) All investors seem to believe
that everyone else is negative, so they don't want to be negative and get caught in a short squeeze,
and 4) There are still a number of investors who don't believe we are heading into a cyclical
downturn. We believe that true investor capitulation in this space is still to come and until it does,
we continue to believe that investors should avoid overweight positions in the stocks.
LOOKING FORWARD, NEXT CATALYSTS ARE INTEL MID-QUARTER UPDATE ON
THURSDAY NIGHT AND THE START OF NEGATIVE PREANNOUNCEMENTS WHICH
WE EXPECT TO BEGIN SHORTLY AFTER LABOR DAY. In the absence of any other
significant newsflow this week, we expect that the Intel mid-quarter update will generate even
more buzz than usual leading up to Thursday night. While there has been significant Street chatter
around Intel's capex budget, we don't believe that Intel will change their capex range of $3.6 billion
to $4.0 billion. With few exceptions, Intel's recent history has been to make any changes to their
capex budget on its year-end call. Our checks with the equipment suppliers indicate that Intel is
making orders for its 65nm rollout and we don't believe that Intel is far enough off of that plan to
make any changes to their budget at this time.
We expect that the datapoints in the semi equipment space will become louder post Labor Day,
which is when we would expect to see negative preannouncements from a range of companies.
There are a number of competitor conferences coming up after Labor Day and we believe that
business is tracking far enough below plan that companies won't be able to communicate to Wall
Street without updating their guidance. We cut numbers on all of the companies in our universe
that won't benefit from the buffer of SAB 101 deferred revenue recognition and believe all of the
companies on which we cut numbers are preannouncement candidates this quarter. The Street
generally expects a round of bad news during preannouncement season and that expectation is
leading some to argue that stocks might rally through the bad news (i.e. the bad news is priced in),
which looks like a reasonable argument given that NVLS was up 2% on Friday after lowering
order, shipment, gross margin, and EPS guidance. That said, we would point out that CMOS and
LTXX both traded down significantly post their reports (which were widely expected to be bad),
leading us to conclude that the magnitude of the misses and earnings reductions post those misses
will determine whether stocks shrug off the bad news from a trading perspective or if there will be
incremental trading weakness.
News, Events and Price Performance
Last week
Monday 23 August (1) Accent Optical Technologies announced that it will cooperate with the
Industrial Technology Research Institute of Taiwan to develop overlay metrology technology for
next generation tools. The new technology is anticipated to achieve overlay measurement precision
to the 65nm node and manufacturing process control to 2.0nm. (2) Mattson Technology filed a
Form S-3 shelf registration statement with the SEC, covering future sales of up to 8,861,144 shares
of already outstanding Mattson common stock held by STEAG Electronic Systems AG.
Tuesday 24 August (1) Tokyo Electron Limited has agreed to make an equity investment in
Molecular Imprints, as a contributor to MII's Series B funding round. (2) Brooks Automation
announced its real-time enterprise software initiative that will extend the functionality of its suite of
real- time manufacturing applications, enabling enterprises to create an end-to- end real-time
environment that reaches from the factory floor to enterprise- level business systems. (3) Credence
Systems announced that MediaTek has prompted several major subcontractors to purchase multiple
Sapphire NP test systems to meet its capacity demand.
Wednesday 25 August (1) Credence Systems (CMOS-$6.93; U/N) reported $0.20; GS $0.20,
Street $0.19. Please see our 8/26 note for additional details.
Thursday 26 August (1) Albany NanoTech of the University at Albany announced that its College
for Nanoscale Science and Engineering has installed and begun qualifying for 300mm wafers using
a 193nm preproduction immersion lithography system. The 300mm wafer scanner-track platform
consists of two components: the ASML TWINSCAN AT:1150i scanner, and a Tokyo Electron
CLEAN TRACK LITHIUS coater/developer system. (2) Amkor has completed its acquisition of
Unitive. According to the agreement, Amkor also obtained a 60 percent interest in Unitive
Semiconductor Taiwan Corporation, a joint venture between Unitive and various Taiwanese
investors. (3) LTX Corporation (LTXX-$5.48; NC) reported $0.14; Street $0.13. (4) Novellus
Systems (NVLS-$25.18; IL/N) hosted a previously scheduled mid-quarter update call during which
it revised its organic order growth guidance to flat, from previous guidance of flat to +5%.
Management stated that it expects Q3 revenues to be $412 million, in line with previous guidance
of $408 million to $418 million, but gross margins to decline by 1%, as a result of lower shipments
($390 million vs. previous guidance of $408 million to $418 million). EPS guidance of $0.37
remained within a previously guided range of $0.37 to $0.39. Please see our 8/27 note for
additional details.
Friday 27 August (1) Applied Materials announced the installation of the company's 1000th
chipmaking system in Southeast Asia.
This week's calendar:
Thursday 2 September; (1) Intel mid-quarter update.
GS Universe Price Performance 8/27/04 Price performance
Ticker Company Name Rtg Close Week MTD QTD YTD Y-Y
Semiconductor Capital Equipment
AEIS Advanced Energy Industries IL/N 10 4% 4% -35% -61% -56%
AMAT Applied Materials IL/N 16 2% -4% -17% -27% -24%
ATMI ATMI Inc. IL/N 19 -2% -6% -30% -18% -30%
ACLS Axcelis Technologies IL/N 8 -5% -12% -34% -20% -7%
BRKS Brooks Automation IL/N 13 -4% -10% -36% -45% -47%
CMOS Credence Systems U/N 7 -8% -23% -50% -47% -35%
ENTG Entegris IL/N 8 -2% -8% -29% -36% -43%
FORM FormFactor OP/N 17 1% -13% -22% -12% -11%
KLAC KLA-Tencor OP/N 38 2% -7% -23% -35% -35%
LRCX Lam Research IL/N 22 1% -6% -17% -31% -13%
MKSI MKS Instruments IL/N 14 -2% -7% -40% -53% -48%
NVLS Novellus Systems IL/N 25 -1% -7% -20% -40% -38%
TER Teradyne Inc. U/N 14 -2% -19% -39% -46% -22%
Mean -- -- -1% -9% -30% -36% -31%
Median -- -- -2% -7% -30% -36% -35%
Source: Factset.
I, Jim Covello, hereby certify that all of the views expressed in this re |
