You are here

Dr. Rodgers' Speech to the American Electronics Association: The American Semiconductor Industry: Winners or Whiners?, May 21, 1990 | Cypress Semiconductor

Dr. Rodgers' Speech to the American Electronics Association: The American Semiconductor Industry: Winners or Whiners?, May 21, 1990

Last Updated: 
May 24, 2012

Dr. Rodgers' Speech to the American Electronics Association: The American Semiconductor Industry: Winners or Whiners?, May 21, 1990

May 21, 1990

T.J. Rodgers

Combined and Edited Remarks made to the American Electronics Association Seattle,
Washington, May 21, 1990 and to the Institute of Electrical Engineers VLSI Multilevel
Conference, Santa Clara, California, June 12, 1990
T.J. Rodgers
President & CEO
Cypress Semiconductor Corporation


Tonight I am going to give a unique speech. What makes it unique is that I run a Silicon Valley semiconductor company-and I am not going to whine once about the Japanese in 30 minutes.

I want to reaffirm right now that the United States is the best country in which to start a semiconductor company, that U.S. semiconductor companies should be growing and profitable, and that if they are not, the fault lies with their management. In the future, if you hear that Cypress is not doing well, you will not hear me complaining about the Japanese or the lack of government support. You will hear me talking about what I did wrong. By the end of this talk, I hope that you will be as worried as I am that the government could poison our semiconductor industry with well-meant, but ill-conceived schemes and subsidies that would numb our competitive passion.

First, I will synopsize the standard sky-is-falling argument, the purported collapse of our $50-billion semiconductor industry that supports our $750-billion computer industry. Next, I will recount some commonly proposed but ineffective solutions to these problems-government subsidies, government projects, and government sanctions against the Japanese-ineffective remedies proposed by troubled large semiconductor companies. Then, I have a six-question test for you on some basic facts about our competitiveness relative to the Japanese. Having given that reality test, I will return to argue that the so-called competitiveness remedies are not really remedies, but burdens that will further aggravate our problems. Finally, I will discuss what we ought to do to ensure a competitive American high-tech industry.


The sky-is-falling story. You have all heard it. According to Dataquest, in 1982, America shipped 51% of all semiconductors worldwide and the Japanese shipped 35%. By 1989, we had slipped to 35% market share, conceding 51% to the Japanese. In just seven years, we traded places, from a 16-point advantage to a 16-point disadvantage, or a loss of 32 points of relative market share. Our government response prompted by that crisis is the National Advisory Committee on Semiconductor or NACS, a group formed to study the problem and report to the President's science advisor. The NACS report, entitled "An Industry at Risk," warns "timely action is urgently needed to arrest the deteriorating global position of the United States semiconductor industry."

Industry veterans have joined the chorus of apologists. Jerry Sanders is the president of the 21-year-old Silicon Valley semiconductor company, Advanced Micro Devices (AMD). Cumulatively, in the last five years, AMD has not made a profit, and it has barely grown. Jerry Sanders has said the U.S. semiconductor industry has "been severely wounded by unfair and predatory foreign competition." Charlie Sporck is the president of National Semiconductor, another big company in Silicon Valley with a record similar to that of AMD. He has said the U.S. semiconductor industry should "form a series of consortia, such as Sematech, to win back market share." In perhaps the most irksome statement that I have heard on the subject, Dick Elkus, the chairman of a Silicon Valley equipment company, likened the United States high-tech industry to a bloated rhinoceros, lying dead on the ground, being picked apart by vultures. His audience of naïve Stanford students laughed. They sat there and laughed at the U.S. being likened to a dead, bloated rhino being eaten by vultures.


The "bloated rhino" group has offered solutions. The first, the Japanese-U.S. trade agreement, is a reprisal for the Japanese attack on the dynamic random access memory (DRAM) market. DRAMs, as you know, are the highest volume chips manufactured and compose the main memory in most computers. The Japanese were successful in their attack because they beat us in DRAM manufacturing. American semiconductor companies voluntarily left the DRAM market because it was not as profitable as their other businesses. Gordon Moore, the chairman of Intel, admitted at a Congressional hearing that Intel had walked away from the DRAM market-shelving a working megabit DRAM-in order to invest in the microprocessor market where profits were better. Most of the U.S. semiconductor companies followed suit. So far, so good, but then, the semiconductor industry elders who abandoned the DRAM market came to Congress and argued, "My God, we don't have any DRAMs. DRAMs are critical. America needs DRAMs to be militarily secure, let alone competitive. We need government to help fight the Japanese."

So, the government negotiated the Japanese-U.S. semiconductor trade agreement. How did it work? Even though the American companies had voluntarily walked away from the DRAM market, it somehow followed that we had to penalize the Japanese. (Now, I should say here that I believe the Japanese did trade unfairly in the DRAM market and they did sell DRAMS below their cost; i.e., dumping, which is against the law.) In essence, the semiconductor trade agreement penalized the Japanese by forcing them to raise their prices. What a great penalty! First you give them the market, and then when they own it all, you say, "Okay, as a penalty for breaking the law, you are ordered to raise your prices." American computer companies paid billions of dollars in surcharges directly to the lawbreakers. Only the government could "help" our technology industry in such a creative way.

Another competitiveness "solution" frequently resorted to by the semiconductor establishment is useless litigation-typically with the rallying cry of defending intellectual property and always with a hidden agenda. It is ironically unfortunate that Texas Instruments just succeeded in getting hundreds of millions of dollars from the Japanese in a legal action enforcing its DRAM patents. The unfortunate aspect of TI's windfall is that most large semiconductor companies have now followed suit by turning their legal departments into profit and loss centers. I am not exaggerating. The level of useless litigation in the semiconductor industry is at an all-time high. As an example, Advanced Micro Devices, in the last two years, has sued Cypress, Gazelle, Intel, Samsung, Lattice, Atmel and Brooktree. How can you run a company right and make money when management is distracted to that extent? Apparently, these litigious corporations espouse the theory that corporate America will become competitive again by suing for its profits. This contagion of litigation, which promises more 80386-like monopolies in the semiconductor industry, obviously will not make America more competitive.

Sematech, the government-subsidized research consortium in Austin, Texas, is another wrong answer. Although billed as a government-subsidized institution to help the semiconductor industry at large, Sematech is really a nicely crafted subsidy won by a group of only fourteen large semiconductor companies that I call the Sematech Club. The Sematech subsidy was won with gun-to-the-head lobbying tactics: "We need money. If we don't get it, the semiconductor industry will fall, and if semiconductors fall, computers will follow. And if computers fall, the Japanese will have won the war. We need $100 million a year." How could any Senator or Congressman, with that un-nerving argument presented to him by prominent Silicon Valley figures, do anything but capitulate? So far, Sematech has cost a quarter-billion dollars and produced no tangible results.

U.S. Memories, the DRAM production consortium proposed by the Sematech Club, was a "moon shot" DRAM re-entry program with a billion-dollar price tag. Fortunately, U.S. Memories died from the disinterest caused by its unrealistic business plan and preposterous price tag.

The National Advisory Committee on Semiconductors (NACS) is another bad idea. This government-sponsored body produces bad ideas equal in scale to U.S. Memories, for instance, doubling the Sematech dole to $200 million a year. On an even more colossal scope, NACS has proposed the formation of the Consumer Electronics Capital Corporation, a gigantic "multi-hundred-million dollar," government-run venture capital firm. This is the last thing we need. Silicon Valley already has too much venture money chasing too few deals. The problem with venture funding is that venture capitalists cannot afford to invest because they cannot make a reasonable return on investment in today's stock market-another government-induced problem I will discuss later.


Let me ask you six questions to test your knowledge of Japanese-American competitiveness. Some of these facts were presented by Dr. Thomas Moore of the Hoover Institute at Stanford during the "bloated rhino" panel discussion.

Question number 1:            Which foreign country owns the most American assets?

Question number 2:            Which country is the most productive in GNP per capita?

Question number 3:            From what country do we import the most?

Question number 4:            Which country has the largest economy in the world?

Question number 5:            Which country has the most favorable balance of trade? (I'll give you a hint: it's not the U.S.)

Question number 6:            Which country has the most engineers?

If your answer to each of these six questions was "Japan," you were wrong six times. What country owns the most U.S. assets? Britain; Japan is a distant second. Moreover, the grand total percentage of U.S. assets owned by the Japanese and mourned so loudly by the whiners is only four-tenths of one percent, mostly trophy property such as skyscrapers in three cities and Oahu beach-front property and golf-courses-not quite the serous economic threat it's made out to be. As a matter of fact, if we are so worried about U.S. assets, why aren't we in an uproar over the loss of our Mobil gas stations to British Petroleum? Why aren't we Brit-bashing? And finally, tell me the truth: if you had a choice between two identical hotels in New York City, one owned by Akio Morita, the chairman of Sony, and the other by Leona Helmsley, where would you choose to stay?

Which country is the most productive? The U.S. is, by far. Germany follows us and, depending on which study you read, Britain or Japan is next. Various studies put American productivity 30% to 100% ahead of that of Japan.

From what country do we import the most? Canada, not Japan. Logs, not Toyotas.

Who has the largest economy? The American economy is 2.5 times bigger than the Japanese economy which also lags that of the Soviet Union. We have, by far, the largest economy in the world. The Japanese economy is growing only slightly faster than ours. Dr. Moore of the Hoover Institute calculates that at the current growth rate differential, the Japanese economy will match ours in 189 years. We have a lot of time to fix our problems properly; we should not allow ourselves to be stampeded into stopgap solutions.

Which country has the best balance of trade? Germany. With the exception of one or two companies like Siemens, Germany is a country of small companies like the U.S., not a country of massive, vertically integrated companies like Japan. Thus, our economic model works in Germany to produce a better balance of trade surplus than Japan's.

And which country has more engineers? The National Academy of Sciences estimates that the United States has twice as many engineers and scientists as the Japanese. Unfortunately, 40 percent of our engineers work on defense-related systems, while the Japanese put most of their engineers into commercial activities leading to exports. In my opinion, the real "peace dividend" will come not from a Department of Defense budget reduction, but from putting some of our most talented engineers to work on salable products rather than on things we bury in the ground and hope we never have to use.

The American economy is very large, extremely competitive, and growing. Although some industries have lost significant market share, the American economy is growing only slightly less fast than the Japanese economy. We have the time to make a measured response to our problems, rather than succumbing to ill-conceived government intervention. By stampeding prematurely in the wrong direction, the government will do more harm than good.


We should stop complaining that the Japanese are the root cause of all of our problems. I do believe that the Japanese market is less open to us than our market is to them. We need to address that problem and we need fair trade, but those issues are secondary and not the root cause of our problems. Also, our pre-occupation with Japan prevents us from getting on with solving our problems at home.

I believe that approximately two-thirds of our problems are our own fault. And when I say "our," I really mean the people in this room-the management of companies in the United States. If you are losing, you are not losing because our economic system is poor or because you are not getting government subsidies or because the Japanese are treating you unfairly. You are losing because you are being out-managed.

Consider the example of quality control. The basic Japanese books on quality control declare unanimously that quality in any corporation starts at the top. If the chairman and president do not personally promote quality with no compromise, the principles of quality, however well taught, will not propagate throughout the company-and the workers will not be fooled for a minute. Workers in every country want quality; just ask the Honda workers in Marysville, Ohio, who export more cars from the United States than any other American car maker. Yet, how many times have you heard American managers blame poor quality on their own workers!

What do you think would happen if we had a surprise quality quiz for the top management in our semiconductor industry? Something simple like, "Please create an Ishikawa cause-and-effect diagram and a Pareto analysis chart that define and prioritize the quality problems in your own administration department." Suppose we took ten Japanese and ten Silicon Valley top semiconductor executives at random and gave them that test on quality control. Who do you think would get the better grade?

That is why we should stop focusing on the Japanese problem and start focusing on improving our ability to manage manufacturing competitively.


While American companies are suing each other, our Japanese counterparts are busy shipping us exports. Let me give you two real examples from Cypress's legal history to illustrate just how absurd some modern legal skirmishes have become.

Cypress's strategy to stay lean while growing is to fund internal start-up companies. At a billion dollars in sales, we want to be a confederacy of ten vital $100-million companies, not one aging giant. When we entered the reduced instruction set computer (RISC) microprocessor business, we needed a technology manager. I heard from a venture capitalist that a well-known computer guru, Roger Ross, was seeking funding to start his own RISC company. I called Roger and said, "Why start your own company? Cypress will fund you. Our sales force will sell your product which can be manufactured in our foundries. You can still run your own independent company. I will serve on your board of directors and I will show you how to build a successful company because I just did it at Cypress." Roger bought it. He and the four other founders of Ross Technology who left Motorola were immediately sued-both as a corporation and individually. What do you say to your wife when she calls you at work and says that she has just been served papers by a $5-billion company? All of Ross's employees were directly threatened with not being able to work for another company on microprocessors-the foundation of their careers. In one meeting, the only female member of the Ross team was told by a Motorola lawyer that she "ought to start a bakery."

Cypress has been sued or threatened with legal action over 20 times in its seven-year history. The Motorola suit was a typical warning shot fired across our bow. We settled the Motorola suit in the usual way, with no money changing hands and an agreement not to disclose the terms of the settlement (so that we could both claim victory publicly-a truly classic piece of useless litigation). By the way, we have settled all but two of our legal actions, and in all of these encounters, we have never once been to a trial, nor have we ever paid a dime to anybody. Those statistics alone should give you a feeling for the merit of this type of litigation.

However, Motorola's garden-variety harassment lawsuit pales in comparison to the level of legal incompetence demonstrated by the AMD lawsuit that followed. Ross hired one engineer from AMD-just one engineer-and they sued us and him, personally. Of course, the suit alleged the usual litany of nefarious activities, starting with theft of trade secrets. The big companies were trying to sue us out of Austin; no matter who we hired, we were accused of hiring them with the intent of taking trade secrets. Fortunately, it turned out that AMD's lawyers made the embarrassing mistake which I labelled "Xerox litigation." Get this: the AMD lawyers took the Motorola lawsuit, which named five defendants, removed their names and inserted the name of the designer we hired from AMD, and then re-filed the lawsuit. They literally copied Motorola's legal document. But their grammar showed that they were not even skillful plagiarizers-they forgot to change plural to singular. The individual designer, for example, was charged withcontinuing "to violate their contractual and fiduciary obligations to plaintiff." AMD must have figured, "What the heck, Motorola is taking a whack at Cypress, let's follow suit; it only costs $65 to file the legal documents. We'll take whatever settlement Motorola negotiates from Cypress."

AMD's opportunistic suit caused me to draw the line. We arrived at their preliminary injunction hearing with good lawyers, ready to fight. We won and AMD indicated that they were ready to settle. At that point, I had to decide whether or not to settle, or to litigate and win a public contest. I decided to offer a face-saving settlement. So, I invented a new legal structure, the "selective no-hiring agreement," a creative way out, which would allow AMD to say that they won a no-hiring agreement, but which would have no impact on Cypress. Both parties agreed to my proposal that AMD would name seven employees of their 18,000 that would be off limits to Cypress hiring for six months. To avoid AMD's having to give me a list of seven key employees (not a smart thing to do), they agreed to create a list of seven names and have an AMD officer hold it. If during the next six months Cypress were going to hire anybody from AMD, I would contact John East, the designated AMD vice president, to inform him that I was about to hire "Ed Jones." If "Ed" were on the list, Cypress would wait for the duration of the six-month period before hiring. This face-saving concession ended the frivolous "Xerox litigation" suit. The conclusion to this episode is appropriately ludicrous. We were about to hire somebody from AMD, so I called up John East's office and said, "We are going to hire somebody from AMD; I want to see if he is on the list." The response was that the AMD list-keeper, John East, had resigned (he is now running a start-up company) and that the list was lost. "Ed Jones" is now a Cypress employee.

That story is good for a laugh, but think about it in the context of an intractably litigious company lobbying in Washington for hundreds of millions of dollars in subsidies for Sematech-and simultaneously blaming the Japanese for their woes. It is our problem, not a Japanese problem. If you analyze the opportunity costs of the three-year legal war between AMD and Intel over rights to the 80386, the probable $1 million per month those companies have been spending on legal fees would have been enough to fund the four Cypress subsidiary companies which will account for half of our growth this year.


Once the "Save America" Sematech pitch succeeded in generating a $100-million-per-year subsidy, Washington went on with its business. The operation of Sematech transferred to its board of directors-you guessed it-the same group that did the lobbying. Apparently, one of their first orders of business was to determine how to restrict membership by preventing the participation of the majority of the American semiconductor companies-companies for which the Sematech subsidy was intended.

The mechanism to block Sematech entry was a clever dues structure which was ostensibly open to all, but really closed to most. Sematech dues are set at one percent of sales or $1 million per year, whichever is greater. As an example, that dues structure would have forced a $20-million semiconductor company to pay dues equaling five percent of its sales-five times the dues for a $100-million or larger member of the Club. Sematech's original, impractical business plan also drove away many potential Sematech members. The plan was to advance semiconductor manufacturing technology without doing any manufacturing-an oxymoron in our business where "learning curves" dictate that manufacturing prowess can be gained only by actually manufacturing an ever increasing cumulative unit volume.

The tactics of Sematech's directors were successful in channeling the $100-million yearly subsidy to a select club of only fourteen semiconductor companies-twelve of them having over $1 billion in annual sales. That's the reason for the apparent short-sightedness of the other 279 of America's 293 semiconductor companies who chose not to sign up for "free" government money. By the way, Sematech dues top out at $15 million per year, so companies over $1.5 billion in sales get a dues discount. For example, a Sematech member with $3 billion in sales would pay one-half percent of sales for dues-one tenth the rate of the $20-million company discussed above.


If the government-oriented answers are wrong, what should we do? The right answers are free markets, entrepreneurs and tough competition. I think our free enterprise system has worked spectacularly well for over two hundred years. It has built the largest, most productive and most prosperous economy that ever existed-all without burdening its people with the kind of personal constraints that other countries impose in the name of national competitiveness. If our government would just leave the economy alone-stop some of its counter-productive practices-we would begin to take back some of the ground we have lost to the Japanese in semiconductors.

A recent article in the San Jose Mercury News tried to make the point that smaller entrepreneurial semiconductor companies cannot take up the slack caused by poorly performing big semiconductor companies. The article compared the combined revenues of a group of Silicon Valley start-ups, including Cypress, Altera, Xilinx, Weitek, and others, and showed that this group of companies was slightly smaller in revenue than National Semiconductor's $1.6 billion yearly figure. To me, the article proved the very point it was trying to refute. Although the entrepreneurial bunch of companies is not now quite as big as National, they are growing a lot faster, and they are a lot more profitable. In fact, based on the current growth rate of the entrepreneurial group, AMD and National, in four years, the entrepreneurial group will be bigger than AMD and National combined. That momentum will make a significant contribution to our national economic picture in the very near future. Remember, Intel was "only" a fledgling $137 million start-up company in 1975.

The second misconception relating to entrepreneurial companies is that the U.S. needs huge companies like the Japanese monoliths-the only type of company capable of making the huge investments required to produce DRAMs, synchrotron rings, microprocessors and the like. That "critical mass" theory also falls apart quickly with a little analysis.

Take DRAMs. We have already talked about DRAM-inventor Intel bailing out for economic reasons, along with AMD and National Semiconductor-all three $1 billion-plus companies when they quit. Which lone American company kept manufacturing DRAMs in America despite the withering Japanese attack? Micron Technology, a Boise, Idaho, start-up which is ten times smaller than Intel.

Look at SRAMs. The Japanese are attacking that market right now. Who is left in the SRAM market? The largest SRAM company in the U.S. is our arch-rival, Integrated Device Technology, which ships about $115 million in SRAMs per year. We are second with about $100 million in annual SRAM revenue. Thus, the top two SRAM companies in the U.S. are not big companies, but venture-funded start-ups. The entrepreneurs are the ones holding the fort in the areas where the Japanese have attacked, not Intel who invented and brought to market the SRAM, not National who bailed out, and not AMD who announced recently for the third time that they were quitting the static RAM business because they could not make any money.

Entrepreneurs are the right answer, and yet we are often impeded or even attacked by our own government. Government funding of organizations like Sematech in essence picks fourteen "winners" and 279 "losers." We should have no "industrial policy," if that term means meddling-the picking of winners and losers. Even Clyde Prestowitz, the author of the semiconductor trade agreement that I described earlier, has said in public, "We all know that if the government gets in the business of picking winners and losers, it will only pick losers." We seem to be in a headlong rush to manage our economy in an attempt to emulate a Japanese model that Tom Peters has pointed out does not exist: that government-formed consortia chartered to compete in a given field are responsible for Japan's success. Peters points out that it was the cut-throat internal competition among fourteen Japanese car companies that damaged our three slow-moving car companies, not any form of government guidance. (The Japanese Ministry of Industry and Trade even told Soichiro Honda not to start a car company!)


We should not relax the antitrust laws. Big U.S. semiconductor companies continue to lobby to reduce the penalties for antitrust convictions from triple damages to single damages. Reflect on that premise: America will be more competitive if companies convicted of antitrust law violations pay smaller fines. The very executives who orate vehemently for strong fair trade laws-the laws which provide what they love to label a "level playing field" for us and the Japanese-are the same ones who are trying to weaken the antitrust laws-the laws which have been so successful in maintaining that same level field for large and small companies within the United States. I feel both sets of "fairness laws" are important and neither should be gutted. Frankly, as the president of a $200-million entrepreneurial semiconductor company, I fear the internal antitrust threat much more than the external fair-trade threat.


Our government is good at throwing money at projects, but not at managing them. Given that assertion, education is a great place to throw money. Just think, Washington could put a check in the mail with a letter saying, "Dear Donald Kennedy: We like Stanford (my alma mater). We want Stanford to offer more integrated circuits courses and to produce more graduate students in this field. Here is our first of ten yearly checks for $20 million. Please do your best to make the integrated circuit industry in the United States more competitive. Use your own judgement on how to spend the money. Please send us a quarterly report, not to exceed five pages, and tell us how you spent the money."

I guarantee that the $20 million would be put to good use; there would be new graduate students, new professorships, new equipment and new laboratories, all created in a thoughtful and economical way. $20 million to Stanford, $20 million to Berkeley, $20 million to Cal Tech, $20 million to MIT, and $20 million somewhere else; $100 million total. That is what we are not doing, because the $100 million is getting buried at Sematech every year. Wouldn't a $20 million yearly grant to each of the top five electronics universities in the United States help America more than Sematech's fourteen-company subsidy?


The most damaging act our government wreaks on our high-tech industry is deficit spending. The balance of trade deficit gets all the press-we like Japan bashing-but our own budget deficit really does most of the damage. A recent Business Week lead editorial reminded us that the United States carried a balance of trade deficit from the Revolutionary War through World War II. We seem to have done well despite that problem. So why are we so worried about it now? It is a symptom, and, unfortunately, we are treating that symptom, not the problem. The problem, and the root cause of the trade deficit, is the budget deficit. The budget deficit creates a vicious cycle that is difficult to break. Here is the short version. Budget deficit borrowing drives up interest rates. A high cost of capital, especially in the capital-intensive semiconductor industry, causes companies to become uncompetitive. Consumers then buy from competitive companies, many of which are Japanese, driving up the balance of trade deficit.

Let me explain the vicious cycle in more detail. Deficit spending drives up the cost companies have to pay for capital in two ways. When the government overspends, it borrows huge sums of money. To attract all that money, it drives interest rates very high, well above those in Japan. Cypress and other companies have to pay even higher interest rates yet to attract money away from the government. Thus, government has forced American companies into a situation which is nothing more than a milder version of the one faced by companies who have issued high-interest "junk bonds" and are staggering under that debt load. The net result: during our period of decline, American companies were paying 9% interest while the Japanese companies were paying 3%. The average cost of the silicon, plastic, copper, gold and other raw materials in a Cypress chip is only about 28 cents. Believe it or not, the capital costs-the money we, in effect, borrow to build our factories-average twice that amount or about 56 cents per unit. In Japan, that same capital would cost about 19 cents per unit. Because of the high capital cost of chip making equipment, our manufacturing costs are much more heavily influenced by the cost of money than by the cost of silicon. Doubling the cost of the silicon wafers we use would have less impact on our bottom line than raising the interest rate we pay by only two percentage points.

Government overspending, and borrowing to pay for it, has a double impact on capital-intensive, high-tech companies. As the government borrows, interest rates go up. Investors then take their money out of the stock market and put it into high-yielding, safe, government-secured debt. As investors sell off their high-tech stocks, stock prices go down, thus impairing the ability of high-tech companies to raise needed capital by selling their stock. So, excess government borrowing impairs our ability to raise money by either of the common means: by borrowing or by selling stock.

Because so much money has been lured from the stock markets by government borrowing-selling T-bills-prices in our market for high-tech stocks are about four times lower than comparable stock prices on the Tokyo Stock Exchange. Our stock market price-to-earnings ratio is 15; theirs is 60. In apples-to-apples terms, American high-tech companies have to sell four times more stock than comparable Japanese companies to raise the same amount of money. As the cost of money goes up, American semiconductor companies get less and less competitive relative to the Japanese. Worse yet, some American companies may save money by putting off critical new manufacturing equipment purchases and become inherently, not just financially, non-competitive. Breaking the U.S. industry out of that vicious deficit spending cycle is the single most important step government can take in improving our industrial competitiveness. Instead, the government continues to overspend and tries to patch up the problem with Sematechs!

Here is a striking example concerning the cost of capital. NMB is a third-tier Japanese semiconductor company, a very good manufacturing company that is the subsidiary of a Japanese ball-bearing company. Compared to Cypress, although they manufacture well, they have only a few products, no marketing, no process development, no product development and almost no sales capability-they buy their products and technologies and use sales agents. If Cypress were competing with NMB in an effort to raise money, I could go to Wall Street and win. Yet, NMB just raised $472 million in their initial public offering. Although Cypress is a highly regarded company on Wall Street, we needed six offerings to raise just $150 million. Why must a highly-regarded American company work so hard to get capital, when a third-tier Japanese semiconductor company raises $472 million right out of the starting gate? Worse than that, what would happen if Cypress and NMB both discovered at the same time " the answer," the way to win, the way to become the dominant semiconductor company in the world? If "the answer" cost $472 million, NMB would become the world's best semiconductor company and Cypress would shrink into obscurity. Why? NMB would simply go to the Tokyo Stock Exchange and sell some stock to raise $472 million. We would face the ultimate paradox: Cypress as a company is worth about $472 million, so we would have to sell the whole company to raise the money to become the world's best semiconductor company. The fact that our P.E. ratio is fifteen and theirs is 60 becomes more than a boring financial statistic when viewed in those terms. If the economic war boils down to just a war of capital formation (which it certainly does not), the system is rigged in Japan's favor. We need to solve our capital cost problems, but make sure they are not used as the excuse for poor performance: Intel has over $1 billion in the bank and we have about $100 million left from the $150 million we raised.


How do you solve the cost of capital problem? After first eliminating the budget deficit, lower the capital gains tax. With a lower capital gains tax, investments in high-tech companies offering a capital gain will be more attractive than putting money into T-bills. Why don't we have a low capital gains tax? Simple, politicians. The public lets them get away with standing up and saying, "I can't support a tax break for the rich." As long as we either really believe that line or we just allow our politicians to get away with that sort of pseudo-populist rhetoric, the problem will continue.

Note well: the maximum capital gains rate in Japan is one percent of the assets sold. Perhaps our unemployed semiconductor workers will take solace in the fact that we did not give "tax breaks to the rich" to finance the building of semiconductor plants where they might have made a living. Instead, we have cleverly made sure that the idle money of the rich was put to good use-loaned to the government-not used to build factories. Voters should make "tax break for the rich" rhetoric as instantly fatal to political careers as the public disclosure of substance abuse.

There is another tax area in which crowd-pleasing rhetoric is contrary to good industrial policy. I refer to those "fat corporations that don't pay any taxes." If Congress does not give our industry R&D tax credits and investment tax credits-incentives for us to invest-then we continue to under-invest and continue to slip in our ability to compete.


There are subtle ways that government action has inhibited our ability to raise capital. The deliberate devaluation of the dollar, as pointed out by George Gilder in a recent Forbes editorial, also limits companies' access to capital. I have discussed the vicious cycle in which the budget deficit comes around to aggravate the balance of trade. When you attack the balance of trade symptom, rather than its root cause, sometimes you actually end up making matters worse. In an effort to reverse the balance of trade deficit,  we have deliberately manipulated the yen-dollar ratio to make dollars and, therefore, our exports cheap. But, after five years of effort, why hasn't the problem vanished? The answer is that our solution to the problem has itself aggravated the balance of trade problem.

To illustrate the problem, consider the semiconductor world as being represented by two companies-Nippon Electronic Corporation representing Japan and Intel representing U.S.-each having a billion dollars in sales and, therefore, 50% market share. Now imagine devaluing the dollar by a factor of two, as we did during the 1982 to 1989 period. After the devaluation, Intel would still be a billion-dollar company because a dollar is a dollar. But measured in dollar terms, Nippon Electronic would become a $2 billion company as a result of devaluation. So, Nippon Electronic would double its sales overnight by virtue of a forced change in the currency exchange rate and become twice the size of Intel in dollar terms. I can already see the headlines: "Japan captures 67% semiconductor share; U.S. slips to 33%. Industry leaders demand government action."

If you remember the statistics I gave you earlier, in 1982 the U.S. had a 51% to 35% lead in world semiconductor market share, but by 1989, our lead had vanished and we were behind 35% to 51%. In seven years, we lost 32 points of relative market share. What the people who use these statistics never tell you is that the yen-dollar exchange rate used in the calculations also changed from 248 in 1982 to 139 in 1989. If you take the exchange rate out of it, if you use the low 1989 exchange ratio for all calculations, 27 of the 32 points of market share erosion are due to variation in the exchange rate. Only five points of the market share loss are the result of non-monetary causes.

You could argue that the exchange rate had to be changed because of a weak dollar, resulting from excessive imports. That may be true to some extent, but to a first approximation, most of the apparent erosion of our semiconductor industry has come from the exchange rate: first, Japan's driving an unrealistic 248 yen-dollar exchange rate in 1982 to make its exports cheap, and then, the U.S. central bank's effort to drive an equally unrealistic 139 yen-dollar exchange rate in 1989 to try to achieve the opposite effect. The unwanted side-effect of the cheap-dollar policy is that when our government cuts the value of the dollar in half, it literally cuts Intel's sales in half; it cuts Cypress's sales in half; it cuts Intel's profit in half; it cuts Cypress's profit in half-and it also cuts our respective stock values in half. As we discussed earlier, low stock prices-be they caused by excessive government borrowing, or by high capital gains taxes, or by government manipulation of currency exchange rates-limit semiconductor companies' access to the money they need to be competitive. Thus, another vicious cycle prevents cheap dollars from solving the balance of trade problem: devaluing the dollar to fix the balance of trade devalues the stock of American high-tech companies, which in turn restricts access to capital for those companies, making them less competitive, and stimulating imports that in turn damage the balance of trade. It's 1984 doublespeak: fixing the balance of trade hurts the balance of trade-if you do it with currency value manipulation.


First of all, as corporate shareholder-voters, you can insist on having leaders-not apologizers-leaders who realize that the U.S. is the best environment for companies and who offer no excuses for problems other than their own performance. You can demand leaders who use engineers and not lawyers to make profits. You can elect leaders who recognize that two-thirds of the American competitiveness problem is self-inflicted, not something to be blamed on Japanese trade practices or government policies. As shareholders, you can rise up against corporate leaders who milk their companies or run them ineptly.

You get what you are willing to endure, Mr. Shareholder. You have a crappy corporate president because you allow him to perform poorly and you don't do anything about it. You allow him to pay himself $1 million a year for poor performance. That's why your company doesn't make money, Mr. Shareholder, and you ought to do something about it.


I think it was very healthy for our industry that we all had a chance to watch Mr. Tellup of Lockheed look unemployment straight in the eyes when 40 percent of his shareholders told him to improve his performance or depart. A recent letter to the editor of the San Jose Mercury News expressed shareholder intolerance very well.


"The semiconductor industry has once again begun to whine because President Bush has taken some steps that may reduce the amount of taxpayer subsidies that

the industry has been enjoying. I'm having a very difficult time with this. As a stockholder of a major semiconductor manufacturer based in Santa Clara County, I recently received a proxy statement that indicated that the chairman of this firm had received compensation of nearly $1 million in 1989. The officer received this for managing a company whose stock value has declined by 80 percent over the past five years and that has shown a net loss over the same five years.

Before tax dollars are spent on this industry, largess as I have just described ought to be restricted. And perhaps more competent management should be brought in to run such a vital industry."

Wouldn't it be great if in a congressional whining session, some congressman leaned into the microphone and said, well if this industry is so damned important, how come we haven't got somebody better than you running it?


I think the future looks good for the American semiconductor industry. If you re-cast the Dataquest semiconductor market share figures with any constant yen-dollar ratio, we have not lost ground to the Japanese in semiconductors since 1985. Our period of precipitous decline was from 1975 to 1985-a period when our manufacturing was so bad, even our entrepreneurial edge could not compensate for it. We bottomed out in 1985 when our manufacturing disadvantage became small enough to be offset by our entrepreneurial advantages. We have been at a stalemate with the Japanese-"trench warfare" according to Intel CEO Andy Grove-for the last five years. In the 1990's, I predict a reversal, as American manufacturing continues to become more competitive with Japanese manufacturing. Japanese quality control and manufacturing management techniques are well documented and readily learned. There is no Japanese societal magic; their techniques can be replicated here in the United States. Hewlett-Packard and Motorola are examples of U.S. companies with excellent quality control, companies which are competitive with the Japanese right now.

Harvard professor Michael Porter seems to be talking about semiconductors in his new book, Why Nations Triumph: "In many ways, what the U.S. needs most is a philosophical shift. Defensiveness and a loss of confidence have crept into American industry and government. A mind-set has developed that U.S. industry is helped by devaluation, feeble antitrust enforcement, lax regulations, cooperation among leading competitors, policies that create a monopoly in particular technologies, and "temporary" protection. As appealing as these policies may seem in the short run, they will only make further loss of competitive advantage more likely."

American semiconductor companies should start spending more time on winning and less on whining.