Outsourcing is a Risky Business

Electronic News, August 27, 2001 by Jeff Chappell

Analysts say the trend helps fuel industry cycles

Outsourcing has been touted as the strategy to alleviate or help manage cyclical downturns, but it may instead be a contributing factor to the cyclical nature of the semiconductor industry.

Analysts last week reached this conclusion during a luncheon panel discussion hosted by Semiconductor Equipment and Materials International (SEMI). Participants were Dan Hutcheson, president of market research and analysis firm VLSI Research Inc., San Jose and Bill McClean, president of IC Insights Inc., Scottsdale, Ariz.

Carl Johnson, president and cofounder of Infrastructure, a Dallas-based market research firm, moderated the panel.

Hutcheson suggested that while outsourcing may help chipmakers react quickly to changing market conditions and help them avoid capital-intensive risks, it moves those risks around to others in the supply chain instead of removing them. When this is combined with macroeconomic factors, it encourages cyclicality, he said.

He cited federal interest rates as an example. The federal government may reduce interest rates several times over the course of a few months to try and stimulate recovery, but it takes six months to a year for changes in interest rates to have an effect on the economy. Thus, all those interest rate drops will have a compound effect in the near future, helping to fuel an economic boom. "I think we're going to see a hard upturn," Hutcheson said.

McClean agreed that this is a possibility. IC Insights predicts that the semiconductor industry won't recover until 2002, but that it will grow 14 percent that year. "I wouldn't be surprised to see it higher," he said.

Both McClean and Hutcheson noted that while there is little evidence of an immediate recovery in chip demand, inventories have dropped considerably throughout 2001. Since the beginning of the year, the industry has used up $10 billion in semiconductor inventories, Hutcheson said. He added that chipmakers were quick to curb capacity expansion this year.

MeClean noted Cisco Systems Inc., while not placing any new orders for chips for its networking equipment this year, has continued to ship $4 billion in equipment per quarter-burning through much of its inventory. Furthermore, much of the current chip inventory in general is nearly a year old, making those products veritable semiconductor Methuselahs. Many semiconductor manufacturers will destroy existing inventories in the near future in favor of chips with more current, marketable technology, he said.

All this will likely make for a steeper upturn than the industry may expect, particularly if equipment suppliers have to put customers on allocation during a scramble for additional capacity, McClean and Hutcheson concluded. This could lead to panic buying such as the industry saw in 2000. "It will definitely happen. You will see panic buying late next year," Hutcheson said.

Reading Silicon Tea Leaves

So what are chipmakers and their suppliers supposed to look for in gauging when the next upturn will be here? McClean suggested that IC average selling prices (ASPs) are a good leading indicator of supply and demand within the semiconductor industry. And right now, he said, "ASPs are still dropping like a rock."

Hutcheson suggested looking to the portents of downturns, the Asian subcontract test and assembly houses, for leading indicators. He also noted that spot market prices of chip stocks can serve as a leading indicator. Last year spot market prices began to drop even before reports of order push-outs and cancellations in the test and assembly houses last July. As for Hutcheson's own prognostications, VLSI Research is predicting negative 29.2 percent growth in capital equipment this year and $41.1 billion in total sales, and 9.4 percent growth and $45.3 billion in sales in 2002. Both he and McClean suggested a chip recovery will likely begin to appear within the next few quarters, but it won't be until the second half of 2002 that it trickles down to equipment suppliers.

But McClean warned that there may be more going on in the industry than just the historical ups and downs. There is evidence that the compound annual growth rate (CAGR) of the semiconductor industry may be falling from its historical 16 or 17 percent to 8 or 9 percent, and it may continue at that lower rate for the foreseeable future, he said. In tracking the amount of semiconductors in electronic systems throughout the course of the industry, 20 percent seems to be a saturation point; when the industry hits this point, which it did in 2000, CAGR drops off. "We've seen it already. It's down to 9 percent," he added.

This theory is also fueled by the fact that IC ASPs haven't risen since the early 19902s, McClean said. The last upturn depended on unit volumes, not prices, and was thus extremely sensitive to economic conditions. "It's very simple. It's supply and demand," he said.