High Tech…the Boom/Bust Cycle of a Generation?

May 27, 1998

Executive Summary

The Nasdaq sells at 76 times earnings. Excluding Intel, Microsoft and Cisco, the rest of the Nasdaq (at $1.7 trillion) sells at almost 100 times earnings. If one excludes the profitable financial institution stocks on the Nasdaq, the rest of high tech on the Nasdaq may sell ---who knows---at 120 to 130 times earnings.

Across-the- board severe deterioration in the PC based high tech industry is in full blossom. PC unit growth worldwide has fallen in half. The average selling price for PCs has fallen sharply. Global semiconductor sales revenues have been negative over the past two years. Semiconductor capital equipment spending cutbacks are everywhere.

Every innovation eventually ceases to become an innovation; in doing so it follows a path of market penetration and eventual saturation. Over the last half decade, the PC based high tech industry has rapidly undergone such a process of market penetration. PCs proliferate in all businesses. Surveys now indicate that almost 50% of all US households own PC's. Market saturation, if it is not already here, cannot be far away.

In the initial stages of market penetration, relative price declines of the new product greatly expand the number of potential users. Overall market revenues rapidly.

As the market for the new product approaches saturation there is little room for further market penetration. Relative price declines have less and less of a positive impact on unit sales. The industry ceases to be a growth industry. In some cases, technological and manufacturing progress may lead to price declines that exceed consequent unit sales increases. Then, overall revenue growth may become negative.

During such periods of market saturation competitive pressures usually lead to severe deadlines in profit margins.

There are signs that something like this is happening in the computer industry. Growth in orders for high tech goods has fallen sharply between 1994 and 1997, despite some firming in the overall economic growth rate. This reflects the rise in the market penetration ratio of PCs and related products to levels from which further rapid market penetration in percentage terms is no longer possible. The demand for bits of memory has increased rapidly, but not as rapidly as prices have declined. Global revenues for semiconductor memory were $40 billion in 1995. They were only $18 billion in 1997.

With an almost doubling in microprocessor production in the pipeline it is probable that we will see declines in microprocessor prices that are so great that they will more than offset increases in unit sales Prices of PC's may then fall so rapidly that the total revenues of the PC-based hardware industry may go significantly negative.

Introduction

The high tech hardware industry experiences ongoing disappointments. Yet, so entrenched is the myth of a new high tech-led-productivity miracle among investors that high tech stocks keep rising. If earnings at the hardware companies disappoint, money moves into Internet stocks that have no near term hopes of earnings where share prices rise sevenfold in weeks on daily volumes that average six times their market float.

The Nasdaq sells at 76 times earnings. Excluding Intel, Microsoft and Cisco, the rest of the Nasdaq (at $1.5 trillion) sells at almost 100 times earnings. If one excludes the profitable financial institution stocks on the Nasdaq, the rest of high tech on the Nasdaq may sell ---who knows---at 120 to 130 times earnings. And, for all we know, these earnings, may be accounting fictions. If these companies expensed the costs of employee compensation by way of options (or did not deduct these costs on their tax books), their earnings would be much lower. Many of these companies are no doubt manufacturing earnings as AOL did years ago when it capitalized its marketing costs which exceeded its cumulative reported earnings. Who knows how insanely priced high tech stocks are relative to their real underlying earnings.

Unfortunately, the potential for disaster does not lie only in insane valuations. Equally disturbing is the prospect of a boom/bust cycle of extraordinary proportions in the underlying real high tech business of these companies. Wall Street always expects failing high tech to recover. The following analysis suggests that the current slowdown may end with the worst imaginable dynamics of market saturation.

The Downturn and the Cycle

The across-the- board severe deterioration in the PC based high tech industry that Fred Hickey and I expected is in full blossom.

PC unit growth worldwide, which peaked in 1995, has fallen in half. IDC predicts unit growth of 13% in 1998, despite sharply lower selling prices.

After having remained stable at roughly $2000 a unit for over a decade, the average selling price for PCs has fallen sharply, probably to an average of $1500 for both business and consumer applications combined. New large scale price cuts are announced everyday.

After having held steady for a while in 1995, semiconductor price declines have been huge. Global semiconductor sales revenues have been marginally negative over the past two years. 1998 could make it three years in a row. Total semiconductor memory sales revenue world wide has fallen by more than 50% in two years. Something similar may now occur in microprocessors.

In prior booms (in a quite regular four-year cycle between 1970 and 1990) semiconductor capital equipment sales peaked briefly (for a year or two) at a 40% annual rate. This time this growth rate persisted for four years. Now large capital spending cutbacks are everywhere and are drastic in their magnitude.

What is happening? According to Fred Hickey a confluence of an economic expansion and several hardware and software innovations resulted in a longer period of rapid high tech expenditures and firmer pricing in this cycle than in past cycles. In the past cycles, the entire high tech boom/bust cycle lasted four years. Hardware revenue growth peaked at a 40% annual rate. At the trough, it went flat if the down cycle occurred in an economic expansion. It went deeply negative if the down cycle coincided with a recession. In this cycle peak rates of growth were sustained for at least a year or two longer than in past cycles. Does a higher and more sustained boom presage a deeper and longer downturn?

Clearly, the downturn has begun in this cycle. The issue now is how deep and how far does it run? In the past we have focused on the fact that the extreme nature of the boom would spawn extremes in the downturn. Of greatest significance was the outlook for excessive new supply. This high tech boom has coincided with the greatest stock market bubble in US history. High tech along with finance has been its leadership sector. This has allowed for unprecedented equity issuance, which has resulted in unprecedented capital availability for investment and unprecedented ease of entry for new competitors. This capital availability has been augmented by a period in the mid 1990's of unusually firm pricing and profitability in the industry and by a commitment by governments (in the Far East) to support at virtually all costs the establishment of beachheads in these critical industries of the future. It is this supply response above all that has created the severe price degradation in the semiconductor industry.

New Innovations: From Market Penetration to Market Saturation

Of greater interest, however, may be the demand dynamics in this industry going forward. Every new innovation experiences recurrent booms, subsequent oversupply, and consequent busts. However, every innovation eventually ceases to become an innovation; in doing so it follows a path of market penetration and eventual saturation. Initially, every new innovation (canals, railroads, automobiles, dynamos, radios, televisions, etc.) begins with a very high price product which only a small market segment can afford. Progress in technology and manufacturing progressively reduce the relative price of the new product. This leads to "market penetration" whereby more and more potential users can acquire the new product. At some point market saturation is reached: all firms or households who will buy or use the new product now have access to it. The market becomes largely a replacement market with overall unit growth bounded by the slow growth of the economy overall.

It is very clear that, over the last half decade, the PC based high tech industry has rapidly undergone such a process of market penetration. A decade and a half ago few households and businesses used PC's . Today PCs proliferate in all businesses. Surveys now indicate that almost 50% of all U.S. households own PC's. The PC is a sophisticated product. Educational levels, even literacy, are inadequate for a significant percentage of the U.S. population. It is quite amazing that so high a percentage of all households own PCs. Clearly, market saturation, if it is not already here, cannot be far away.

Fred Hickey makes the following pertinent comment:

"The PC industry's problems are many and permanent. At 90 million units per year, it is impossible to keep growing at 25%. Computer Intelligence estimates 45% of U. S. households had a PC at the end of 1997. Eighty percent of homes with incomes over $100,000 have a PC, 65% with incomes in the $50K-$60K range. According to Forrester Research, 36% of consumers who do not own a PC said they have no interest in buying one, no matter how low the price. The problem is complexity. Todd Thibodeaux, VP and senior economist of the Consumer Electronics Manufacturers Association told Barron's recently that he expects it will take eight to 10 more years to get the next 20% of American households to buy."

According to international studies, consumer market penetration of PCs is much lower in Europe and Japan then it is in the US, leaving room for further rapid growth in consumer demand. However, for all high tech goods (computers, robotics, etc.) market penetration ratios in the business sector are fairly comparable overall across the G-7. Of course, penetration ratios are much, much lower in emerging economies. However, in these economies the potential market is bounded by their much lower gross national incomes, despite their multitudes of consumers and firms.

The progress of a new innovation from market entry through market penetration to market saturation generates a consistent pattern of change in the elasticity of demand of the product with respect to relative price. In the initial stages of market penetration, relative price declines of the new product greatly expand the number of potential users. With further declines in relative price rapid market penetration results in very large sales increases. Therefore, overall market revenues expand, and rapidly. In the jargon we say the elasticity of demand of the new product greatly exceeds unity.

This all changes as the market for the new product approaches saturation. Then, because there is little room for further market penetration, relative price declines have less and less of a positive impact on unit sales. Price declines soon offset unit sales increases. The elasticity of demand approaches unity. The industry ceases to be a growth industry. In some cases, technological and manufacturing progress may lead to price declines that exceed consequent unit sales increases. Then, overall revenue growth may become negative.

During such periods of market saturation competitive pressures usually lead to severe deadlines in profit margins. Not only does the transitory "Schumpeterian" return to innovation dissipate; a former growth industry geared to rapid sales growth enters a period of "consolidation" in which competition becomes unusually intense. Such periods of consolidation are usually characterized by profit rates well below the prior-industry average and bankruptcies at the competitive fringe.

Given the huge advance in market penetration by PC's in the 1990's, it is likely that the PC industry is now in the throes of market saturation dynamics . This is ironic, since most market participants believe that the 1990's have ushered in a new era of unprecedented high tech based growth. They ignore the cycles of innovation and maturity of all prior new products and the obvious fact that any industry with 50% of the consumer market can no longer grow 25% per annum in a 3% per annum growth economy.

Productive capacity growth in the industry follows with a lag the prior growth trend and the availability of capital, which in this cycle has been extraordinary. Investment in personal computers and their accessories and components have been the leading edge of recent reported accelerating growth in the manufacturing capital stock. Even if one ignores the "over investment" in these high tech industries in the Far East, supply in the high tech industry is expanding at a rapid rate.

Yet, demand has slowed sharply. If we look at nominal order growth for this same industry we see a virtual collapse in orders between 1994 and 1997 despite some firming in the overall economic growth rate.

Throughout history the ruling class has always sought to own gold and silver because they represent purity and longevity.