Semiconductors: Glimmers Amid the Gloom

Chipmakers are in the midst of a nasty cyclical downturn. But the signs point to a brighter future

By Thomas W. Smith

Sometimes it's downright hard not to paraphrase Dickens. So here goes: These are the worst of times -- and the best of times -- for the semiconductor industry. Thought at present chipmakers are mired in a cyclical downturn, the future holds a great deal of promise.


  Let's get to the bad part first. The semiconductor industry is presently diving to a cycle bottom. Worldwide sales of semiconductors were $16.87 billion in Jan. 2001 (latest available data), indicating a rise of 13.7% on a year-to-year basis, but down 5.7% sequentially from the December 2000 level. (These figures are from the Semiconductor Industry Association, or SIA.)

Given the many earnings warnings and descriptions of rapid deterioration of orders in February and March by chipmakers and their customers, such as Cisco Systems and Nortel, the downward sequential trend for chip sales is likely to continue into the spring and perhaps beyond. Listening to the companies' conference calls, one hears refrains such as "more sudden than most downturns," and "very little visibility into the June quarter, let alone the whole year." Or: "Inventory overhang will take longer to clear now that demand is weakening."

Trends are also downward for the semiconductor equipment makers. According to preliminary figures from the trade association Semiconductor Equipment and Materials International (SEMI), North American-based manufacturers of chip equipment received orders of $1.80 billion and made shipments totaling $2.35 billion in February 2001. That makes for an ugly book-to-bill ratio of 0.77. (A book-to-bill ratio below one, or below parity, indicates a contracting industry, whereas a ratio above one indicates expansion.) The book-to-bill peaked in March 2000 at a very robust level of 1.46. By October 2000, the ratio had dwindled to a still strong 1.16.

Then things began to fall apart: November's ratio was 1.12, December's 0.99, January's 0.80, and February came in at 0.77. Back in the cyclical trough in the summer of 1998, the book-to-bill ratio dipped below 0.60, so, based on that precedent, we might have a few more months to go before the ratio gets to cycle bottom levels. And several months after the bottom in the book-to-bill, shipments should start to climb slowly.

The suffering for the investor in the semiconductor sector can be summed up by a quick look at the Philadelphia Semiconductor Index. The index is down from its 52-week high of 1320.69, to a measly 579.01 at the close March 28, 2001. That is not far off the 52-week low of 515.31, which implies that it would not take much more bad news to tip the index to new 52-week lows.

So the industry and the stocks are way down from boom-time levels, with some likelihood of falling a little more. Seems like the worst of all possible worlds -- low and headed lower.


  However, with some historical perspective, it is easy to view 2001 as a very bright time for the semiconductor industry and for investors. First of all, look at the proliferation of semiconductors in everyday life. Personal computers are commonplace in the U.S., and penetration still has a lot further to go in other countries. Wireless phones and their PDA cousins are becoming common, with adoption faster in parts of Europe and Asia than in the U.S. Smaller, better, cheaper chips are doing many new tasks in environments such as automobiles (e.g. mapping systems, entertainment systems, automatic tire pressure readouts) and buildings (security systems, digital audio/video systems, broadband communications). The cheap video watch is almost here. In the sense that society is eagerly using chips in many new ways, this is a golden era for semiconductors. The end demand is there to support a giant industry.

A key fact of life to bear in mind is that the semiconductor industry has grown at 17% compound annual growth rate (CAGR) for the past 40 years. That is a relatively high rate of growth for a major industry.

Another key fact is that the industry is characterized by approximately four-year cycles. These cycles are usually marked by a rush to build factory capacity that eventually creates an over-capacity situation, and ultimately, an industry slowdown. However, non-industry factors such as general recessions sometimes affect the cycle. This is the case in the present slowdown: demand for chips slowed with unexpected swiftness.

The cycles are deep and reliable, in the sense that they have occurred for decades and are apt to recur for another decade, given that technological advances seem likely to continue at the rapid pace of Moore's Law for the next decade. Note well, a reliable cycle can be a valuable thing for the type of aggressive investor interested in doing the homework to buy amid a gloomy downturn and sell amid a rosy upturn.

Alas, the precise timing of industry cycle turning points is hard to predict. Furthermore, guessing when the crowd of investors will move into the stocks adds a significant complication. Nevertheless, for investors willing to sit with quality companies through what could turn out to be a protracted downturn, both chips and chip equipment are attractive industries. Valuations are generally low now, which might make this the best of times for investors seeking to rotate into a premier cyclical industry.

Along these lines, I recently upgraded my recommendation on Applied Materials (AMAT ) to 5 STARS (buy). The company is the largest supplier of wafer fabrication equipment and enjoys many competitive advantages of scale. It has good products for the coming generation of 300mm wafer fabrication plants (fabs) and sales for these advanced technologies are holding up fairly well, even as sales for older 200mm wafer fabs are down. Although near-term earnings visibility is very dim, and the downturn could turn out to last another year, we think the company is well positioned to gain market share in the next upturn, whenever it may come. In my prime scenario, orders start to look better for Applied Materials by year-end 2001, and investors may anticipate that by six months.

Smith offers additional insights on the semiconductor industry in a special BusinessWeek 50 edition of Video Views.

Smith is the semiconductor industry analyst for Standard & Poor's

Before it's here, it's on the Bloomberg Terminal. LEARN MORE