Do You Trade The Semis? Read This.
When taking positions in semiconductor and semi cap
equipment names, I tend to take positions that are not longer than
six months (although not without exception). I’d rarely want to
take a one-to-two year position in almost any name
within the semiconductor stocks (or the semi cap names, for that matter) – long
or short. Product cycles, market share, and design win momentum fluctuate too
much over a six-month time frame that what you thought would occur often changes
quickly. The key is where Street expectations and market sentiment are at a
point in time, and then determining if an edge is present to exploit.Â
For
semiconductor stocks, I mostly take positions based on product cycles,
end-market exposure, and DRAM prices. Micron is a name that presents, from time
to time, short-term trading opportunities related to patterns in DRAM pricing,
which are, in my opinion, the result of inventory dumping by Micron and other
major memory makers. I can write in more detail about this some other time. Â
For semi
cap equipment stocks (such as large cap names AMAT, KLAC, NVLS, LRCX) I tend to
look at the directional bias of utilization rates, DRAM pricing, and the overall
profitability of semiconductor companies.
In
order to gauge the health of both sectors, I look at the SOX and the SMH. The
SOX is an Index of 18 semiconductor and equipment names, and is more heavily
weighted towards the semi cap equipment names than the SMH. The component stocks
are included on a market cap weighted basis, and the weights are periodically
adjusted. The way the SOX trades can be a little whacked at times because of the
weights. But I think watching the SOX’s intraday chart is a great indicator for
the overall for the direction of the technology sector. The SOX generally leads
the tech sector both ways — up and down. This makes intuitive sense, since an
uptick in chip demand then flows into original equipment manufacturers (OEMs)
such as Dell, Hewlett-Packard, the newly merged IBM-Lenovo, etc.
The SMH
includes 20 semiconductor capital equipment and semiconductor names. The
component stocks are included on a market cap weighted basis, and the weights
are adjusted from time-to-time. I typically use the SMH as an ETF to get
exposure to a group of large cap semi names. I don’t notice any particular
characteristics when trading the SMH – it’s really just a trading vehicle for
quick large cap semiconductor exposure.
Psychology
has been, and continues to be negative in the semiconductor stocks largely due
to fears of a 2005 slowdown and growing inventory levels. On balance, the sell
side is baking in minimal growth for 2005 in the low- single digit area,
although I think it’s possible that Street numbers for 2005 will be lowered
again.Â
I think
there’s more downside in the semiconductor sector than many others who are
taking the view that the worst of the news out. My main concerns for the
semiconductor sector are high inventory levels and margin contraction over the
next two to three quarters. Margins are likely to tick downward as inventories
are high at the semiconductor companies. However, it’s possible that we hear
chip sales are up, if manufacturers choose to move inventory at a discount
rather than write it off completely, and which would supposedly excuse margin
declines. Also, I think it will take longer to work through the inventory build
than people expect largely due to the timing of recent inventory accumulation,
since demand falls off in the first and second calendar quarters of 2005. That
said, we could get a great buying opportunity in a stock like INTC and some
other semiconductor names somewhere around late first quarter to early second
quarter once Street numbers are revised downward another 20 percent or so and if
inventory levels have been worked down.
Growing
inventory at the semiconductor companies has a ripple effect throughout the
technology value chain, and can present a variety of trading opportunities.Â
- For
Intel, which seems to be the company with the highest level of inventory
build, it’s bad, because to drive revenue growth the company will need to cut
prices and in so doing will forego margins – unless demand picks up
substantially. - For
AMD, a competitor to Intel, although there has been some inventory build it
hasn’t been to the extent of Intel’s inventory build. (For SanDisk it’s not an
issue because pricing has been so aggressive and end demand has been a lot
stronger for handsets than it has for PCs.) - For
OEM’s this is fairly positive because they will likely keep part of the chip
price reductions, and pass on the rest to the customer. - For
consumers, this is good because OEMs will cut pricing and in theory the
average Joe could see lower PC prices. - For IT
distributors (such as Ingram Micro, Tech Data etc.) the impact is tough to
determine. It’s a difficult area to get insight via broadline distribution
because it’s not a main item, although chips of some type are in virtually
every product sold. There are a lot of cross currents in PC land: if there’s a
fall off in demand, it might be offset by a battle between HP / IBM / Dell /
Toshiba / Sony, etc. that could keep unit sales moving. Pricing and margins
then become issues, but with this many crosscurrents, the impact on
distributors is too tough to call right now.
Over the
next few weeks, the news flow in the semiconductor sector could become positive
as Street sentiment is still awful about the sector and the fourth quarter is
seasonally the strongest time for demand. So any positive data points on the
margin will likely be viewed positively and many will “hope” the cycle just took
a quick breather. I don’t happen to think this is the case right now, and would
be concerned as we head into the seasonally soft first and second quarters. So
while there may be a quick trading rally, I think it’s unlikely to be
sustained. I think the issues above will work their way into consensus numbers
over the next two to three months, and at some point in second quarter of 2005
there may be a great chance to go long semiconductors again. But, in my opinion
for the time being there is better risk / reward being long elsewhere in
technology and outside of tech.Â
Going
forward, in the semiconductor capital equipment sector it’s all about orders.
Recently, we have been getting first indications of cancellations / push outs.
The next step will likely be capes revisions by major independent design
manufacturers (IDMs). If orders can pull back 20 percent to 25 percent, then
either the December quarter or the March quarter could provide a solid base from
which there could be a buying opportunity in these stocks. I think a shift has
really occurred though and as good as 300mm was, it also hurts the equipment
companies because of the additional capacity it has brought on. I think it’s too
early to own the semi cap stocks, but we are potentially getting closer. We need
order cancellations / push outs to accelerate over next two to three months,
which would be accompanied by a reduction in Street in estimates, to present a
potential buying opportunity early next year.