A glimmer of hope for semiconductors

By Daryl Delano, Economist

In the middle of this year the semiconductor market was in a state of near chaos, the result of a confluence of factors. Is there no light at the end of the tunnel? Finally and fortunately, the answer is a qualified yes. Bolstered by positive financial news from Santa Clara-based bellwether Intel Corp., it's now possible to discern some hopeful signs for the market, some glimmer of the early morning light.

The dollar value of total worldwide semiconductor sales increased 22.3% between May and June of this year, following a 5.3% decline the month before. June is normally a very strong month for the industry, so be careful in interpreting this strong number; between May and June of 1997 semiconductor sales rose 20.0%. Still, the ability of the industry to achieve normal seasonal improvement this June gives reason to believe that the industry is at worst bouncing along the bottom of this severe down cycle.

Unit sales volume rose 8.1% between May and June. This June's unit sales, however, were 2.0% lower than the sales level recorded in June 1997--the first time in almost two years that unit sales in a particular month have fallen below their level of a year earlier. There were a total of 6.5% more semiconductor devices sold during the first six months of 1998 than during January-June 1997. It's been clear for some time now that 1998 will be a very slow year for unit growth in semiconductor sales, but it is doubtful that it will slip into the negative territory breached during 1992 and 1996.

The dollar value of dynamic random access memory (DRAM) sales this June was 49.3% behind its level last year. Through the first half of this year, DRAM sales values were down 34.7% from the January-June 1997 total. Microprocessor sales were 10.5% below the June 1997 level; through the first six months of 1998, the dollar value of microprocessor sales has faded by 11.2%.

No corner of the globe has escaped the problems of the semiconductor industry. Through the first six months of 1998, the dollar value of semiconductors sold to Europe almost equaled sales for the same period in 1997. Through June, year-to-date sales to this region were off a scant 0.3%. During the same period, sales to the Asia/Pacific region (excluding Japan) had fallen 4.2% below the total for the first six months of 1997. However, much more worrisome double-digit declines were recorded in sales to the other two regions of the world: North America (-11.5%) and Japan (-19.8%).

With all of this unambiguously bad news, is it really reasonable to believe that this recent small glimmer of hope will lead to market recovery in the foreseeable future? While it's by no means a sure thing, the semiconductor market has a long history of boom and bust, and its fortunes can turn on a dime. Total dollar sales grew by an average annual rate of more than 34% for the three-year period of 1993-1995, before plunging by almost 9% during 1996.

A question of when

But it's never really been a question of if the semiconductor market is coming back. We're not talking buggy whips and bowler hats here, after all. The question is always when not if the market bounces back and how steep the growth curve will be during the recovery. Nobody has been able to answer these questions of timing and magnitude of recovery (or downturn) with any precision in the past, so there's no reason to think that they're going to get it right now. However, this shouldn't stop forecasters from trying to discern those forces that drive this brutally cyclical pattern.

The people closest to the industry--analysts at the Semiconductor Industry Association (SIA)--are currently forecasting average annual growth of 18% in semiconductor sales for 1999-2001. This forecast, released in June, is still considered a reasonable "best guess" of what lies ahead for the industry, although the SIA's projection of just a 1.8% decline in 1998 dollar sales probably understates the ultimate depth of the losses this year.

Microprocessor sales in dollar terms are expected to be at best flat this year following growth of more than 26% during 1997. But SIA projects sales will grow by about 25% next year and by an average of 19% in 2000-2001. And even DRAM sales values are expected to increase solidly during 1999-2001, by an average of 32.4% annually, following a decline that will approach 30% this year.

Wishful thinking? Perhaps, but although one can quibble with the precise timing of the recovery in semiconductor markets, there is no question that the market fundamentals are solidly in place for a quick recovery when it does occur.

The inexorable forces of ever-expanding end-market demand and constantly evolving technological innovation make the semiconductor market recovery inevitable. Demand has slowed a bit in recent months, but output gains as measured by the government in their industrial production indexes remain extraordinarily strong for both the computer and communications equipment industries. There's no reason to believe that these trends will do more than moderate marginally as the U.S. and global economies slow, but continue to expand. And numerous studies document the fact that semiconductor content in the products of these end markets--as well as in many other less obvious, less traditional end markets--continues to grow.

Now implicit in all of these forecasts is the assumption that worldwide economic conditions will be better in 1999 than this year. Given the depths of Asia's current problems that would seem to be a reasonable, if still debatable, assumption. That's not to say that we've absolutely reached the bottom, particularly in the case of Japan, but the consensus view of economists remains that, taken as a whole, worldwide GDP and business investment growth will improve at least a bit next year. It would take something on the order of a true global economic catastrophe to derail the strongest sectors of the economy, most electronics/ high tech in nature. This is not to pretend that economic conditions don't matter for the semiconductor industry, but other factors--supply, demand, technological change--matter much more in the long-term scheme of things. And with worldwide GDP growth forecast to rise from 1.9% this year to 2.3% in 1999, global economic meltdown is seen as only a remote possibility.

So the demand side remains strong, by any standard. What about the supply side of the equation for the semiconductor industry? Here the waters are murkier. Yes, there are serious over-capacity and over-supply challenges facing the industry at the present time--the free market gives its unequivocal evidence of this in depressed product prices. But these problems are self-correcting; new capacity is coming on board at a dramatically reduced rate compared to even a year ago. Supply shortages are a ways off, but given the continued strong gains in end-market demand they're inevitable, just as in all past cycles. In semiconductor markets, as in commercial construction markets, there's no "soft landing" or "smooth takeoff." Turbulence is a given.

And a final factor--not specifically demand- or supply side grounded--that will bring the semiconductor market back from the depths of despair is the little matter of technological innovation. End markets will continue to demand smaller chips, with lower power consumption, for new and improved product applications. Semiconductor content will continue to grow in existing applications, and increasing numbers of low-tech products will incorporate semiconductors into their workings. At the same time, an increasingly competitive global marketplace will force manufacturers to make the productivity-enhancing, efficiency-improving investments promised by the migration towards larger wafer design.

So the long-term future is undeniably bright. That does not much ease the immediate pain and trauma that semiconductor manufacturers experienced, but it's the perspective needed for long-term success. And one final, more positive, perspective for the electronics industry as a whole: sometimes one industry's misfortune is another's good fortune. In this case, of course, any industry that uses semiconductors as input to its final products has benefited in the short term from the price declines that have accompanied chaos in this supplier industry. In the longer run, however, the health of the electronics sector depends upon a vital, efficient and profitable semiconductor industry.*