Why You Should Buy When the Chips Are Down

Semiconductor stocks are in full retreat. The Philadelphia Semiconductor Sector Index has fallen more than 20% in the past three months. Competitive pressure and  Microsoft's (MSFT) delay in launching its Vista operating system have pushed chip industry titan  Intel (INTC) down about 30% from its 2006 high. Despite its dominance in chips for wireless handsets,  Texas Instruments (TXN) also has seen its stock trading down with the sector. We believe that in some cases the market has thrown out the baby with the bath water and, for investors with a long-term view, this creates buying opportunities.

Our Approach
Let us first step back and talk about the challenges in picking chip stocks and how our approach can help long-term investors avoid some common pitfalls.

Investors are often intimidated by the daunting task of figuring out the semiconductor cycle. Why is it so hard? For starters, it's difficult to accurately forecast end-market demand. On top of that, the time it takes for chipmakers to change their capacity in response means that they have to get it right years in advance. Take the example of chipmaker  Marvell (MRVL). A small portion of Marvell's revenue is derived from its chips used in  Seagate (STX) hard disk drives, a small portion of which are in turn used in  Apple's (AAPL) iPod. Even the industry's best analysts have had trouble forecasting iPod sales correctly, so just imagine the fun analysts would have in projecting sales of Seagate and Marvell. We haven't even mentioned pricing, inventory, or capacity along the entire food chain. The iPod is but one end market for Marvell. Marvell is but one chipmaker. You get the picture.

Even if the rare talent might exist who could consistently predict the turning points in the semiconductor cycle, differing investor reactions to each cycle make it difficult to profit from that information, to say the least. For example, chip stocks have sometimes rebounded before the industry's revenues have hit bottom, and sometimes after.

We don't claim to be smarter than the research shops and industry gurus who specialize in predicting the semiconductor cycle. Rather than focusing on the cycle, however, our full-cycle valuation approach focuses on long-term value creation. As indicated by our extensive valuation work using discounted cash-flow analysis, the long-term trend in free cash flow growth is the key driver for value creation from one cycle to the next. This is where Morningstar's moat and risk analysis can give long-term investors an edge over short-term speculators. A related caveat: If you are out to make a quick killing in the market or to beat a benchmark every quarter, this is not for you. To profit from our approach, you really need a long investment horizon. You need to be able to hang in there when times are tough or take profit when others are rushing into the market.

So here's how we do it. Like other analysts, we evaluate near-term growth drivers to model the companies' financials in the near term. Unlike many other analysts, we spend most of our time understanding long-term trends in a chipmaker's end markets as well as long-term evolutions of the chipmakers' and their customers' competitive positions. For example, we believe consumer electronics is becoming the next frontier for many chipmakers. Low unit cost and large capacity will continue to drive adoption of small form-factor hard disk drives in devices such as the iPod. Even though the storage market is commodified, the storage chip market is not. Marvell's chips offer better performance at a lower cost than its rivals thanks to its cutting-edge technology. We'd then base our assumptions on our analysis. This is the process we follow in evaluating a semiconductor stock's long-term growth potential. It underlies our confidence in going against the trend when necessary and finding value in unpopular places. It will not help you pick the top or bottom (which we would advise against to begin with), but it can help long-term investors profit from cyclical chip stocks without having to worry about the cycle. For recent examples, check out the history on  Intersil  or  Silicon Laboratories (SLAB).

Our Picks
We like the long-term prospects of high-performance analog (HPA) chipmakers and programmable logic device (PLD) makers. Our favorite names in these industries are  Linear Technology ,  Maxim Integrated Products , and  Xilinx .

Both industries have high barriers to entry. In the case of HPA, the demanding nature of analog chip design coupled with a dearth of analog engineers give Linear and Maxim a real advantage. In the case of PLDs, switching costs in the form of high upfront development cost for manufacturers and training cost for design engineers to become efficient on a new platform effectively keep new entrants out. The long life cycle in the communications and industrial end markets make their product prowess last even longer. In addition, Linear's focus on profitability, Maxim's expertise in integrated products, and Xilinx's dominance in high-end PLDs further distinguish them from the crowd.

Another stock to consider is our favorite name in the communications chip space:  Broadcom . Broadcom's team of excellent engineers has arguably developed the most enviable technology portfolio in the broadband communications space, further complimented by its acquisitions. In the HPA space, we also like  Analog Devices (ADI). The company is simply not getting the respect it deserves from the market, as a result of unfavorable end market changes in 2005. We believe Analog Devices has put its problems behind it and should move back to our fair value estimate in time. In the PLD space, we'd also be buyers of  Altera  at the right price. Other than favorable industry dynamics, we believe the company's initiatives in new products will pay off in coming years. In addition, we like  ARM Holdings . Proliferation of portable electronic devices such as wireless handsets, PDAs, and MP3 players should drive license revenue growth for its popular ARM processor.

Stocks to Avoid
Investors have to be careful, though, when looking for value in the discount bin. We remain concerned about  SigmaTel's  future, as its new MP3 chip has not yet gained traction in a competitive market. We are still not enthusiastic about companies in depressed end-markets such as  Agere  and  Applied Micro Circuits . Neither are we excited about large-cap integrated companies that have failed to create shareholder value, such as  Infineon (IFX), Philips Semiconductor (a division of  Philips Electronics (PHG)), and  STMicroelectronics (STM). And a word of caution in general about buying technology stocks: many are currently under investigation for their options granting practice, which poses additional risk. We'll be watching closely as this plays out, but as of now we haven't seen the need to change any of the fair value estimates for the companies I've mentioned.