You are here

When Accountants Attack Profits: The GAAP Accounting Exodus | Cypress Semiconductor

When Accountants Attack Profits: The GAAP Accounting Exodus

Last Updated: 
Sep 03, 2009

The Enron debacle has cast a shadow on the credibility of corporate America. Companies are turning over every stone to make sure that financial reports are both accurate and presented in a way to convey credibility. One would think that in this environment, my company, Cypress Semiconductor, would be reporting exclusively with the “seal of approval” of the Generally Accepted Accounting Principles (GAAP). But we do not. We publish two sets of financial numbers: one in the GAAP standard, and the other in the “pro forma” standard, which we have turned to as the basis of our communication to our investors over the last few years.

Why would we issue two sets of financial numbers, when a simple, single report would have served better in the current environment? Simply stated, the GAAP accounting rules, as mandated by the Financial Accounting Standards Board (FASB), no longer conform to reality. Our GAAP-based report misrepresents the finances of our company to our investors, necessitating the second, pro forma report.

The root cause of the problem is that the Financial Accounting Standards Board, the organization that sets accounting standards, has prioritized some dubious accounting theories over clarity of reporting, accuracy of reporting and even common sense, as I will demonstrate. One problem relates to accounting for acquisitions. Another potentially more important problem (at least for Silicon Valley) will be created if FASB demands the expensing of stock options. After detailing several problems with GAAP reporting accuracy, I will describe an alternative: a simple and accurate system for asset reporting proposed by Walter Schuetze, a former chief accountant of the Securities and Exchange Commission (SEC).

What company would risk deviating from GAAP accounting standards in the post-Enron era? The answer: Intel, Advanced Micro Devices, Conexant, Fairchild, LSI Logic, Motorola, STMicroelectronics, Texas Instruments, Cypress and many others. Indeed, a PriceWaterhouse- Coopers survey found that 74% of semiconductor companies issue pro forma earnings statements in addition to legally mandated GAAP statements. I will focus today only on the semiconductor industry, but the GAAP exodus is happening in other high-technology industries, as well.

The event that caused most semiconductor companies to break ranks with GAAP-only accounting was the gradual elimination of the pooling accounting method for mergers and acquisitions starting in 1998. A final ban on pooling accounting came in 2001. When FASB forced companies to begin accounting for acquisitions using the alternative method of “purchase accounting,” the earnings of the acquiring companies were suddenly decimated—with charges that did not reflect the reality of their acquisitions. After that change, investors, and the analysts that interpret results for them, had to subtract out the phantom purchase-accounting losses, known as “goodwill charges,” before comparing current earnings to prior results. This correction of the GAAP financial statements became a basis for the exodus from GAAP to pro forma accounting.

Figure 1 (below) shows why the elimination of pooling accounting for acquisitions distorts profit and loss (P&L) statements, necessitating the alternative pro forma reporting method.

Consider two hypothetical companies, ACO and BCO, that each share 50% of a U.S. market, with ACO dominating in the West and BCO dominating in the East. Suppose further that each company has identical finances, and that the president of BCO is retiring, causing BCO to accept an acquisition offer from ACO, a one-for-one stock swap, to form ABCO.

Under pooling accounting rules, the revenue and profit of ACO and BCO are added to create ABCO, a company twice as big as ACO or BCO. To pay for the acquisition, ACO issues 100 million new shares to the shareholders of BCO, making the combined share count of ABCO 200 million shares. Since both the earnings and the share count double, the earnings per share remains constant. And, assuming that the shareholders value ABCO with the same price-to-earnings (P-E) ratio as they did both ACO and BCO, ABCO’s share price remains constant at $40. With twice as much revenue and profit, and twice as many shares outstanding, the market capitalization value of ABCO is simply the sum of the market capitalization of the two companies. This accounting treatment is simple, transparent, straightforward—and forbidden.

In contrast, as shown in Figure 2 (below), the purchase accounting treatment of the same transaction creates a complicated and misleading result— due to the FASB mandate to create a fictitious “goodwill” balance sheet entry. BCO’s “goodwill” is defined to be the difference between the market capitalization of BCO and the value of BCO’s assets; that is, goodwill is the market capitalization of BCO that cannot be accounted for by physical assets. The purchase accounting treatment of the ABCO merger is given below.

In this case, the $2 billion goodwill value of BCO, a quantity which jumps up and down with share price and is nowhere reported in the GAAP financial statements of BCO, is forced onto the balance sheet of ABCO as an asset, despite the fact that the asset is a fiction and has no cash backing whatsoever. Furthermore, that asset is required to be amortized over a period of years, as if it were a piece of equipment. Assuming that the $2 billion goodwill value of BCO is amortized over a typical period of five years, ABCO’s earnings will be decimated with a charge of $400 million per year over that period.

Thus, purchase accounting requires that companies place a fictional asset on their books. The amortization of that asset over an arbitrarily determined period, or the revaluation of that asset, produces a fictional loss on ABCO’s profit and loss statement, which, in this example, completely wipes out ABCO’s very-real profits.

Under purchase accounting, ABCO is a $4 billion company which has no earnings. Under pooling accounting, ABCO is a $4 billion company that earns $400 million a year pretax—the combined amount that ACO and BCO made before the acquisition. Of course, pooling accounting reflects reality, since neither of the companies became less profitable due to paper shuffling. FASB accounting theorists provide convoluted explanations of why a company making $400 million per year of profit doesn’t really make $400 million a year of profit. When common sense fails, the FASB high priests always fall back on the unassailable religion of “accounting principles.” (Unassailable principles, I might add, that were the opposite—and unassailably so—just years before.)

This departure from reporting reality, as exemplified by the ABCO example, is why both companies and investors have moved on to pro forma accounting. If you look at the Cypress earnings estimates posted on the First Call™ website, you will find that the analysts—each from a prestigious Wall Street firm— project our earnings with the phrase “excluding goodwill” in the report title. The analysts are not interested in GAAP balance-sheet manipulations; they want to know our real earnings. Similarly, the investing public rewards or penalizes us not on a GAAP basis, but for making or not making our pro forma earnings estimates.

In prior years, M&A transactions were less prevalent and confined primarily to maturing industries undergoing consolidation.

Today, M&A is a way of life. Cisco Systems, the Silicon Valley data communications powerhouse, set records for growth by using acquisition as a tool. Cisco has acquired dozens of companies, primarily to get access to talent and technology. Cisco’s strategy allows them to purchase known-good R&D en masse, rather than building R&D one hire at a time—and then enduring the risk that their R&D projects may be unsuccessful. In this new environment, technology companies must either make strategic acquisitions, or fall behind their competitors. For example, Cypress made only one acquisition during its first decade of operation, but has made 13 acquisitions in the last three years.

The fuel driving our surge of acquisitions is the downturn of 2001, during which many good companies had trouble getting a second or third round of venture financing. For Cypress, these companies represent an incredible opportunity to assimilate talent and technology. Unfortunately, young companies typically have almost no hard assets. Under purchase accounting rules, we are therefore forced to write off almost the entire acquisition price as an expense. It is quite possible that an acquiring company taking advantage of a few acquisition opportunities in today’s market could not report a GAAP profit over the next five years, even if it were really making an actual 20% cash profit on a pro forma basis.

I am under no illusion concerning my job function: It is to raise the share price for our shareholders, primarily through the mechanism of raising our earnings per share, which, in turn, comes primarily from revenue growth. Of course, my job is to be done with a long-term perspective, which I’ve held for the 20 years since co-founding Cypress in 1982. I believe in making profit the right way— by hiring good people, by creating innovative technologies and products, and by manufacturing them with excellent quality and costs low enough to make a good profit. Short-term, profit-enhancing maneuvers, such as cutting R&D expenses to tidy up the P&L statement, are always destructive to the long-term goals of raising earnings per share and share price, and should always be avoided, regardless of the short-term consequences.

When Cypress was founded, Jerry Sanders, then president of Advanced Micro Devices, the company that I left to form Cypress, declared that there was no room for new startups and that the semiconductor industry was already consolidating toward a “Big Three,” like the car industry. Cypress had to fight through that real barrier to entry: the attack of established companies publicly discrediting us. Our first big customer, Digital Equipment Corporation, would not even talk to us during our first year, until AMD itself shut down one of their computer assembly lines by losing the manufacturing recipe for a product we also made, a Static Random Access Memory, or SRAM. Cypress did not lose its SRAM recipe and we recorded our first significant revenue in 1984.

Just as our business took off, Japan began battering the American semiconductor industry. Semiconductor industry leaders whined, claiming unfair play, but the reality, as I testified five times before congressional and senate committees, was that the Japanese were simply better than we were at that time. Typical Japanese product quality in 1982 was approximately 50 defective parts per million shipped, while American quality was in the range of 1,000 or more defective parts per million. Just as American consumers kicked out General Motors in the 1970s to embrace Toyota because of quality, so did American engineers kick out American semiconductor companies in the 1980s to embrace Nippon Electric Corporation (NEC), Toshiba and Hitachi. The president of NEC also said in an interview that he did not think a new company like Cypress Semiconductor could make it.

We fought our way through the Japanese quality wars. Despite four college degrees, I had never been taught the concepts and mathematics of quality control. My tenure at two American semiconductor companies did not add much quality-control education either. When the Japanese quality attack threatened our company, we hired experts, put together crash quality-education courses for the executive team, and, when we had mastered the subject, passed on that knowledge to the rest of the corporation. Cypress’s quality level is now 25 parts per million. And last year, Cypress surpassed NEC in revenues from our primary memory business, SRAMs. Indeed, we passed by all Japanese semiconductor companies in SRAMs, to become No. 2 in the world behind Korea’s Samsung.

Today we are facing our latest challenge: The communications revolution. The semiconductor industry is moving abruptly away from its decades-long focus on the computer industry toward the datacom industry and its arcane technologies. As in the case of the quality revolution, we are again facing a knowledge gap. In the last year, Cypress executive staff has attended many lengthy seminars on how data is “framed” on the synchronous optical network (SONET); how data moves from a fiber on the SONET network through a photodiode, transimpedance amplifier, limiting amplifier, serializer-deserializer, and demultiplexer; and how your digitized and compressed voice is encoded in codedivision multiple access (CDMA) cellular phone systems. It’s been a long, tough, challenging—and exciting—year. One very challenging characteristic of the semiconductor industry is that we make the bricks and mortar of every electronic system. So, when an electronics era changes, we either fail or move on to become experts in the new field. I learn as much every year now as I did when I was in graduate school.

To summarize, my job is to raise shareholder value by increasing our profits through revenue growth. So far, Cypress has delivered those profits, despite the best efforts of established companies to thwart the Cypress startup, despite the withering attack of the Japanese semiconductor industry in the 1980s, and despite having to get the equivalent of a new Ph.D. every four or five years just to be able to understand the language of our customers.

I’ll be damned if I’ll let accountants take Cypress’s real and hard-earned profits away from our financial report card with phony losses.

I remember a meeting in 1998 in which we were discussing whether or not to acquire a small startup company with good people and great technology. At the time, the SEC had begun to restrict the use of pooling accounting. We could not find a way to structure this acquisition without triggering a punitive purchase accounting treatment. Get this: We nearly opted not to make an acquisition that was highly favorable to our shareholders because of punitive accounting, all in the name of “more transparent” financial reporting.

Our investment bankers said that we could break through the dilemma by using “synthetic pooling.” I’d never heard that term before. To use synthetic pooling meant to make the acquisition using purchase accounting, but with a new strategy. The strategy of synthetic pooling was not to optimize the purchase accounting treatment—that is, to minimize the impact of goodwill losses on the P&L statement—but to use purchase accounting at its punitive best, to deliberately create an ugly, large, goodwill hole in reported profits. The purpose was to highlight, not to hide, the inaccuracies of purchase accounting. The investment bankers claimed that institutional investors, the pros who hold 65% of our shares, would ignore the phony goodwill losses and give us credit for our real earnings and for making the right move for our shareholders. We did a case study of several companies that had employed synthetic pooling, which greatly reduced, or even eliminated, their profits, at least on paper. Their investors ignored the goodwill charges. The companies maintained normal P-E ratios based on their pro forma profits, which eliminated goodwill losses. The analysts who followed those companies also converted to pro forma earnings estimates to provide accurate analysis.

Given our modest size and stature at the time, we were not willing to blaze the “synthetic pooling” trail with the pioneers. We acquired the company and have suffered a goodwill charge to GAAP earnings ever since, despite the fact that it is now a profitable part of Cypress. A few years later, we noted that many semiconductor companies—and the analysts following them—were eliminating goodwill charges from their reports and earnings estimates. We finally joined the crowd when over half of the semiconductor industry had begun reporting pro forma earnings.


Business is good in Silicon Valley, even during hard times. I think we have set some high standards for creating value and treating our employees well. At Cypress—and we are just a typical Silicon Valley example—our average San Jose employee, including our manufacturing workers, earns $107,000 a year in cash and benefits. Capital gains on stock options add to that figure. Silicon Valley workers do not find it necessary to join unions, because they are treated well and they are all shareholders. Contrary to the current political rhetoric on “fat cat” stock options, approximately 82% of the options granted by Cypress every year are given to rank and file employees. The other 18% go to our board of directors, our executive staff and me. Three of the 10 most affluent towns in the Unites States are Silicon Valley bedroom communities. San Jose’s per capita income is double that of the United States, at large. We invent the future and spread the created wealth throughout the organization. It works well. Why would anyone want to change it?

Again we return to accounting principles and the belief that they should somehow cause us to change the way we run our businesses toward perfection, as viewed from the double-entry ledger. In 1994, FASB tried to force the expensing of stock options; that is, forcing companies to report a fictional loss when they grant options to their employees. FASB came to Silicon Valley for a so-called hearing on the topic, but they made it very clear their minds were already made up prior to the meeting. That was a big mistake. San Jose, a place that cares a lot more about electrons than elections, got mad and held some rallies. The politicians shut FASB down quickly.

Unfortunately, the accountants are supported by a few politicians in a current attack on stock options. Sen. Carl Levin (D., Mich.) has openly attacked companies and stock options throughout his career. Perhaps his animosity toward the auto industry has something to do with his views. But why should Silicon Valley suffer collateral damage in Levin’s battle of the big auto union versus the big auto company? Levin’s coauthor, Sen. John McCain (R., Ariz.) is simpler to figure out: He has an unbridled ambition to become president of the United States. McCain scored in popularity polls during the last election for his position on campaignreform. He is looking for another populist campaign issue. I met with McCain in his office, in an effort to convince him that corporate welfare should be eliminated. He dismissed my presentation (calling billions in corporate welfare “peanuts”) and then poked his index finger toward my chest, and said, “But we are going to do one thing you really won’t like; we are going to take away your big, fat stock options.” McCain believes attacking the “fat cat’s” paychecks will help get him elected. So, for reasons that have nothing to do with the reality of running a business right, for the sake of some accounting theories which don’t make sense, and for the sake of politicians who have unrelated axes to grind, stock options are again under attack.

Using the Enron episode as an excuse, the Senators scrambled to relaunch a word-for-word copy of Levin’s failed 1997 anti-stock-option legislation. It’s virtually impossible to have a rational, reality-based discussion with these politicians on why leaving the success of Silicon Valley alone is a good idea, because what they’re trying to “fix” is unrelated to improving our businesses. When a company raises money, it faces a choice between using debt or equity methods. If it chooses the debt method, it borrows money, and its earnings per share go down because it must pay interest on that debt. If it chooses the equity method, it sells stock, and its earnings per share go down because of the greater number of outstanding shares; that is, because of dilution. As you might imagine, in a free market, the coupling between debt and equity financing is very tight. Often, the decision on which funding method to use boils down to a fraction of a percentage point in the cost of money. But, here is the main point: Companies either pay for raising money in the form of losses incurred by paying interest on a debt or they pay for raising money by the dilution incurred from issuing shares; they never pay both ways. The Levin- McCain proposal (which is a very close cousin of the 1994 FASB proposal) would force us to treat stock options as both an expense and a dilution. In other words, the use of employee stock options would not only trigger a loss on the P&L statement, but it would further reduce earnings per share due to increased share count. Accounting theorists babble for hours on why this “double taxation” of stock options makes sense, but it simply does not.

Even in a technical accounting sense, the Levin-McCain proposal is inconsistent with current accounting methods for related financial instruments. For example, our company raises money— about $1 billion over the last 10 years—by selling a widely used security called a convertible debenture. A convertible debenture offering is a hybrid between a stock or equity offering and a debt offering. The value proposition behind a convertible debenture offering is that companies sell stock options to shareholders at prices above the current market price. Until the company actually delivers the higher share price, it is obliged to treat the offering as a debt, paying interest to the convertible debenture holders. Once the offering or “strike price” is reached, the company then pays off the debt permanently by issuing shares, as it would in an ordinary stock offering. The SEC rules for the accounting of convertible debentures require two calculations. In one calculation, the convertible debenture is not counted as debt (that is, the interest payments are ignored), but the shares underlying the convertible debenture are counted in the shares outstanding, diluting earnings per share. In the second calculation, the shares underlying the convertible debenture are ignored (that is, there is no dilution), but the interest payments on the debt are taken as a loss. The SEC rule is that a company with an outstanding convertible debenture must then report the lower of the two results. So, in a directly comparable case, the SEC itself requires that the stock options inherent in a convertible debenture are treated either as a charge to earnings or as dilution, but not both. McCain, Levin and FASB seem to be going out of their way to create a new punitive rule for stock options. Is their purpose to make Silicon Valley more like Detroit?

Consequently, if accountants require us to expense stock options it will represent another unrealistic charge against earnings that I predict both companies and analysts will ignore. If successful, FASB and McCain-Levin will have driven GAAP another step away from being a useful accounting method that investors can use to measure companies accurately.

The word “theology” relates to a set of beliefs that are supported by faith and not required to be supported by reality.

I have used the word in analogy to describe beliefs unsupported by reality in accounting and politics. In many professions, certainly accounting, law and even engineering, the selfconsistent logic of the profession sometimes causes its practitioners to float in a theoretical space above reality. But business has a way of forcing people to deal with the reality of issues. Reality is where I live. If we do not produce what our customers want, we will be ignored and our competitors will prevail. Our employees expect their checks every other Thursday. Analysts and investors expect us to achieve our earnings estimates and that our share price will appreciate. Everybody expects that we will be ethical and obey the laws—just ask the ex-CEO of Enron to explain the penalties for violating that rule.

Many of us regard the founding of Hewlett-Packard as the genesis of Silicon Valley. It is instructive to actually read the postulates of the HP Way. The first cited objective is to make a profit. Yes, in the HP Way, which is generally considered to be one of the most humanistic forms of management, the first order of business is to make a profit. Of course, making a profit the HP Way involves innovation and treating people right, the hallmarks for which HP is more well known.

Contrary to a popular saying, I submit to you that profit— money—is the root of all good. It is profit that enables Silicon Valley companies to pay their employees well— and that enables those employees to support their families and communities well. It is that environment which provides Silicon Valley workers with exciting jobs —not just paychecks—and that inspires them to engage in further invention. That invention, in turn, creates more wealth—wealth that is mostly reinvested to stimulate further invention. What I have just described is one cycle of the economic engine that drives Silicon Valley. Yet, we are asked to tinker with that incredible engine of prosperity, not because of any real issue, but because of the theologies of politicians and accountants.

There are many examples of how theology creates systems inferior to those that are reality-based. An obvious example: One group of theologians may declare it is “right” for women to be veiled, uneducated and not allowed to appear on the street alone. The realists would say that any society that denigrated and ignored 50% of its brain power was destined to be—and stay—poor. That’s an easy example.

The Japanese economy is a more subtle example of economic reality versus economic theology. Many reasonably consider Japan to be a high-technology, capitalist economy. But why has their economy been in a decade-long recession, even when times were good here and elsewhere? The simple answer is that the Japanese government does not truly believe in free market capitalism, which they tout when they’re selling cars or TVs to us but ignore at home.

The economic theology of Japan is that a ruling group of men—the right men, from the right schools, working for the right companies, which are members of the right company groups, or keiretsus—have the unquestioned right to control the assets of the country.

Who controls the assets of America? You do. And I might add that you are rather brutal about it. You save money and invest it in the stock market, either directly, or through your 401K plan, or through the bank where you save it. When, for example, your 401K underperforms, you check off a box on your computer form to fire your 401K account manager with no warning and no mercy. When the investment firm you fired loses your funds, they immediately have to dump some of the stocks that they hold. That’s Cypress stock, sometimes. I’ve been fired like that. My net worth was cut in half in 30 days in 2001. I accept your sometimes brutal decisiveness. Why should the college accounts of your children suffer to make me wealthier?

You invest in the stock market to get a better return than the 4% interest you can earn at a bank. Do you know where typical Japanese citizens invest their money and what interest rate they get? They invest in the Japanese Post Office at 0.25% interest. Unlike you, they are not allowed the free-market response of firing an underperforming 401K manager. Their government “protects” them from the complicated money market that you deal with—and they pay dearly for it.

Who gets the nearly free money in Japan? The right men from the right companies enjoy financing at interest rates 10 times lower than ours. And the right men can underperform all they want, because no one is able to check that box on the computer form to fire them.

Unfortunately for the American semiconductor industry, the anti-capitalist flaws in the Japanese economic system took more than a decade to smack the Japanese semiconductor industry into the stone wall of economic reality. At first, that nearly free money made Japanese companies very competitive. In our industry, 70% of the cost of a silicon chip is due to depreciation, the cost of the money we have invested in the factory that built the wafers. With the Japanese cost of money 10 times lower than ours, their chips automatically cost less than half of ours. That financial fact—combined with our industry’s quality problems—took its toll on the American chip industry in the late 1980s, when Japan surpassed the U.S. to dominate the world semiconductor market. But the American pay-as-you-go system has prevailed in the long term.

The Japanese economy fell ill as Japan’s economic belief system came into conflict with economic reality. Here’s a vignette of how it happened. The keiretsu bank lent the hypothetical Japanese Semiconductor Company, JSEM, $1 billion to build its first wafer fabrication plant, which I will call Fab 1. That plant was optimized to give up capital productivity in return for better quality and higher human productivity. Why not? The capital was cheap, quality was an effective competitive weapon, and people were expensive. The enhanced financial competitiveness of Fab 1 was not used to make extra profit, but to lower prices to drive competitors out of the market. And it worked; Japan took over the No. 1 market share position from the U.S. in 1986.

When it came time for JSEM to build Fab 2 two years later, because of its low-price strategy, JSEM had no retained earnings from Fab 1 to fund Fab 2, so it borrowed $1.5 billion more from the keiretsu bank to build a second capital-inefficient plant. JSEM then owed its keiretsu bank $2.5 billion, but continued to seek market share rather than profits.

JSEM’s Fab 3 was planned to be built two years after Fab 2 at a projected cost of $2 billion. But JSEM’s meager profits would not support the interest on $4.5 billion of debt. Furthermore, since JSEM built its original Fab 1, the American semiconductor industry had caught up in quality. The American semiconductor plants—built with much less capital because of the high price of money in the American markets—had also become more costcompetitive than JSEM’s plants. There were even bigger problems: JSEM’s keiretsu bank had lent too much money to too many companies that didn’t pay it back. The keiretsu bank ran out of money and even the national bank of Japan had a liquidity crisis, which endures today. There were no more “free money” loans to be had. And, since 1993, the American semiconductor industry is back solidly in first place, with 51% market share in 2001, nearly double Japan’s 28%.

In a capitalist free market, investors coldheartedly select only the best investments. CEOs must be productive with your capital or be fired with a check mark on your 401K computer card. Japan is often described as a capitalist country, but it is certainly not. Under Japan’s economic theocracy ordinary citizens have little control over their money—the ruling class makes the “right” decisions about how money is spent. Do you see the politburo with capitalist camouflage here? The only difference is that the ruling class in Japan is selected on a merit basis. Unfortunately, when America engages in “government-industry partnerships,” a.k.a., subsidies, corporate welfare, tariffs, and the like, we engage in the same destructive practices that devastated Japan’s economy and put the Soviet Union out of business.

The tactical nature of the reality-theology conflict for us in Silicon Valley is that we broadly ignore the theologians in Washington for most of our careers. What we care about is uninfluenced by the latest, rehashed conservativeliberal debate, or by the blundering statements or sexual peccadilloes of the latest president. We are 100% consumed by increasing our knowledge of the technologies whose frontiers we are pushing back. A conflict with Washington ignites only when our companies are forced either to go along with the latest, dumb—and destructive—idea from Washington (or Sacramento or Norwalk, Connecticut, FASB’s hometown), or to fight it. That fight is why I don’t have much time to fight ordinary pork barrel politics anymore. Senator Snoot can build a parking garage named after himself in the name of homeland security at my expense—as long as his other laws don’t threaten my Silicon Valley homeland.

Japanese companies generally choose not to fight but to go along with bad economic theologies—to their detriment. That’s one reason I keep a copy of Soichiro Honda’s epitaph and reread it every few years. He was an engineer with drive and integrity who created a great company by listening to his customers and then by pulling off the tough engineering required to meet their needs. He worked outside the theology of the Japanese system and was shunned for it. He succeeded even when the Japanese Ministry of International Trade and Industry (MITI) told him not to go into the automobile business. And, while General Motors was firing engineers and hiring lawyers to prove that the clean air standards could not ever be met (my god, we’d use up all the platinum in the world on those catalytic converters), Honda introduced its CVCC engine in 1974 on a Civic model that met the tougher California air quality standards—without even needing a catalytic converter. I owned that car; it was clean, got 35 miles per gallon—and could lay rubber. I’ve owned a Honda ever since then and have two of them, now.

When attacked by Washington, American companies split into two camps, the realists, exemplified by many Silicon Valley companies, and the politically correct, who subscribe to the theology of the day (the government should subsidize our R&D expenditures; our board of directors must be as race- and gender-diverse as our hometown in order to run the company properly, the government should put a tariff on our foreign competitors to keep us healthy, our CEO needs a $30 million jet to be efficient, we can’t get top management talent without $20 million severance agreements, etc.) Do you see some of the Fortune 100 here?

Consider Bethlehem Steel, the company that made the steel for the Golden Gate Bridge and that just helped convince President Bush to put a punitive 30% tariff on imported steel. The sad story of Bethlehem Steel is outlined by Professor Jim Collins (an author of Built to Last), in his new book, Good to Great. While Bethlehem Steel was building a new, luxury office tower in an “X” shape to maximize the number of corner offices, and while Bethlehem executives golfed on the executive course and showered according to their pecking order in the hierarchy, Bethlehem’s successful competitors— including Texas’ Nucor Steel—went about building the new, efficient steel mills that were demanded by the marketplace. Of course, Bethlehem’s newest set of tariff training wheels will continue to prevent the company from learning how to pedal unassisted. The company—weakened by yet another round of steel subsidies—will make it through this economic trough and perish in the next. Good riddance. In the reality-theology debates, the sad fact for me is that despite their pro-capitalist rhetoric, many of my fellow CEOs are not free market capitalists. When they review a new, proposed legal attack on the free market, they don’t see it as an attack on capitalism to be fought with vigor, they see it as a potential opportunity to gain some legislated advantage—and then put their lobbyists to work, goldmining. The semiconductor industry mostly lines up on the free market side. Even the Semiconductor Industry Association, our trade organization, has a policy not to solicit corporate welfare. That policy was initiated in 1996, when the partially government-funded semiconductor research consortium, Sematech, returned to Washington an unused $200-million portion a government grant given out as a reaction to the Japanese attack on the semiconductor industry.


By 1997, it was clear that the pooling accounting edict was not working, as judged by the GAAP exodus. Senator Phil Graham, chairman of the Senate Banking Committee, held a hearing to address the problem. Both Ed Jenkins (the new, moderate and more reasonable head of FASB), and I testified at that hearing. The hearing was classic politics: How to declare that FASB and the SEC had been right about the accounting change all along—but, at the same time, to change the rules again to reverse the GAAP exodus.

Under the new proposal, goodwill from an acquisition would still be put on the balance sheet, but the requirement to depreciate it—the rule that caused the fictitious quarterly losses—would be eliminated. At first, the idea seemed clever and workable, but it has not brought our industry back into the GAAP camp.

The Senate hearing was a menagerie that showed just how Washington works. All sorts of government-campfollower “entrepreneurs” testified. Each had an angle on how to make money from the new accounting scheme. An investment banker testified on how his industry would be needed for “impairment testing,” a mandated process in which each of the goodwill entries on a company’s balance sheet would have to be re-evaluated every year. (I started to get visions of $1 million invoices for these new outside “services.”) A commercial banker testified that goodwill could be an asset to secure a loan. (He was either irrational or saw an opportunity to get 15% interest on junk collateral.) There was also the testimony of accountants and securities lawyers who saw an opportunity to profit from the new rule.

I used the forum to present a bigger and more powerful investorrights concept: Forcing goodwill to be put on the balance sheet is not just an accounting distortion; it is fundamentally wrong, because it allows the government to take control of private property. Think about what the abstract accounting term “goodwill” really represents. It is the difference between the market capitalization value of a company and the value of that company’s hard assets. Equivalently, it is the value the company’s shareholders give to the company for being more than just bricks and mortar. My company has market capitalization of $2.5 billion and assets of $1 billion. Our shareholders have therefore awarded us—for our company’s knowledge, human talent, technologies, etc.—a premium of $1.5 billion over the share value we can account for with cash and possessions. That’s a private, free market transaction between our 77,033 shareholders and us. It’s the bonus they’ve chosen to give us for our performance. No government agency should try—on principle—to distort that free-market transaction. We’re proud of the goodwill value we’ve earned from our shareholders. On the other end of the spectrum are companies whose market capitalization is less than their asset value. That’s when corporate raiders swoop down and chop up companies to sell off their assets.

One extremely important aspect of free markets is that people must have the absolute right to make bad decisions and lose money as a result. Without that freedom, people won’t have the incentive to correct counterproductive behavior. Think about the boom. Absurdly, it spawned three companies that sold pet food over the Internet. They had almost no assets, but their market capitalization, and hence goodwill value, was sky high—right up until they discovered that pet food was heavy, and therefore expensive to ship, and that it was simpler to buy Alpo at the supermarket. The demise of the companies forcefully reminded investors that their ownership and control of goodwill also carried with it the responsibility to dispense goodwill carefully. They did not. They paid for it. And, hopefully, they learned from the experience. I am 53 years old, and have worked all but five years of my career to build Cypress Semiconductor. Our shareholders have awarded us an extra $1.5 billion over our asset value—60% of our share price—for doing a good job. Suppose General Electric acquired Cypress to enter the semiconductor industry and asked me to run Cypress as a wholly owned GE subsidiary. GE shareholders should value the hypothetical Cypress acquisition for themselves, by setting GE’s share price and hence goodwill value after the acquisition. Should they set GE’s post-acquisition share price to reflect only Cypress’s acquired assets? Or should they reflect the whole Cypress acquisition price in GE’s share price? Or should they give GE an even higher share price to reflect an important acquisition that fills in a strategic gap in GE’s technology portfolio? These decisions rightfully belong to GE shareholders.

In other words, GE shareholders should “vote” in the free market to add or subtract goodwill value from GE’s share price, based on their evaluation of whether or not GE made a smart move in acquiring Cypress and whether or not GE overpaid. But free market principles notwithstanding, under either the old or patched FASB rule, the goodwill value of GE’s hypothetical Cypress acquisition would be confiscated from GE shareholders. Under the result, as the hypothetical accounting high priest might put it to GE’s shareholders, “No, you don’t get to value your acquisition’s goodwill. We are going to seize the goodwill value created by Cypress, put it on GE’s balance sheet, and depreciate its value to zero over five years. Of course, we’re doing this so you can understand GE’s finances better.” Under the new, book-but-don’t-depreciate rule, the accounting high priest would say, “No, you don’t get to value Cypress’s goodwill, even though you paid to acquire it. We will force GE to hire an investment banker to value Cypress’s goodwill for you on a yearly basis.” Do you see a state-controlled economy here?

My Senate testimony on investor rights was ignored and we now have a new cottage industry “helping” us manage our goodwill “assets.” I even tried to get the SEC to listen to my goodwill-belongs-to-the-shareholders pitch. Arthur Levitt, then head of the SEC, agreed to hear me. Unfortunately the man, touted as the champion of the investor, fell asleep during my passionate pitch on investor rights (which is hard to do, since I practiced my 15-minute pitch and delivered it with energy, four feet from the chairman). When he woke up, Levitt gave me the bureaucrat’s dismissal by telling me that his hands were tied, that the SEC just followed the rules set by FASB. (Sure, Arthur.) When the new, book-but-don’t-depreciate rules for goodwill took effect, I hoped that my company could return to GAAP-only accounting. I would like nothing better than to report—accurately —our quarterly results in one, standard format. But, like most top-down edicts, the new rule has not fixed the problem of distorted earnings reports. The problem is that there are alternative methods besides using goodwill to account for the value of a company beyond its tangible assets. Other accounting fictions such as “in process technology,” “existing technology,” “license value,” and “non-compete contracts” can be put on the books as phony assets, instead of goodwill. Each theoretical “asset” has its own different, convoluted accounting rules. Acquiring companies tend to divide up their goodwill problem into multiple categories to optimize their specific finances. The new FASB rule only fixed the goodwill part of the problem. The rest of the phony losses, resulting from the other junk on the balance sheet, still decimate otherwise healthy P&L statements. Are you starting to wonder how real your GAAP reports are?

Consequently, the semiconductor industry and the analysts that follow it did not accept the book-but-don’tdepreciate olive branch extended by FASB, because it did not fix other P&L distortion problems. Furthermore, the GAAP vs. pro forma accounting issue cannot be put to rest until we have resolved whether or not GAAP will require the expensing of stock options. If the “double taxation” of stock options happens, many hightechnology companies will never make a profit—at least according to GAAP.

Recently, I wrote an editorial for The Wall Street Journal attacking the proposed McCain-Levin stock-option legislation. I received an email from Mr. Walter Schuetze, a former chief accountant of the SEC. He supported my position, but not as a fervent defender of stock options. He is really a fervent defender of the integrity of the balance sheet. The financial reporting problems I have discussed all involve distorting the balance sheet with assets that do not have any hard backing. What do you do with “goodwill” or “non-compete contracts” if you are awarded such assets in a bankruptcy? Mr. Schuetze, a logical and plain-spoken Texan, says clearly what ought to be reported as “assets” on a balance sheet:

“I think that we should define assets by reference to real things, not abstractions. I think that we should define assets as follows: Cash, claims to cash, for example, accounts and notes receivable, and things that can be sold for cash, for example, a truck.”

Do you think a balance sheet meeting his criterion would be difficult for shareholders to understand? FASB does. Have you checked recently to see how many of the socalled assets on the balance sheets of your portfolio companies are accounting fictions, rather than cash or other saleable assets? Can you even infer that information from the financial reports you get? Mr. Schuetze also said to me, “I think FASB and the SEC have done a terrible job on reporting. The financial reports today are almost unreadable.” So there’s more than just one engineer/CEO with that opinion.


First, FASB and the SEC should sign up to the Hippocratic oath “to do no harm”—that is, they should leave stock options alone. The current system works exceptionally well; it built Silicon Valley. Some greedy CEOs in some politically correct companies will overpay themselves, sometimes breathtakingly so. That’s not an “abuse” to be “corrected” by a new law that will hurt the majority of well run companies in Silicon Valley. That situation is a mistake made by a corporation in the free market. The mistake should be corrected by the free market—and it will be corrected. That correction may take time, just as in the case of the Japanese economy or the bubble, but economic reality will eventually penalize corporations