To Kill A Giant:

Wintel, the Oligopoly

Ted Lewis

Binary Critic

I’d Rather Be Rich Than Famous

Apple has been defeated, Novell has nothing left but the inertia of its (huge) installed base, Sun Microsystems and other workstation vendors are living on borrowed time, and IBM, HP, Unisys, and DEC never really mattered. The AIM (Apple/IBM/Motorola) challenge to Intel’s domination of the microprocessor instruction set has fizzled and retreated to the relatively backwoods of embedded systems. AMD cannot make enough K6 chips to be a threat, and everyone else is looking for a niche. Meanwhile, Wintel gets stronger every day. Is competition dead in the computer industry?

The computer industry looks like it is shaping up to repeat the oil industry story, circa 1900. Remember when Standard Oil owned "everything?" President Roosevelt capitulated to the muckrakers and broke up Standard to save the American way of competition. Today, the Wintel "monopoly" reminds most of us of Standard Oil, which has prompted some to complain like muckraker Ida Tarbell. This new activism has prompted Wintel, the Oligopoly to ask, "ask not how we can make more money, rather ask how we can keep the department of justice off our backs?" Microsoft and Intel would rather be rich than famous, er infamous, in the eyes of the DOJ.

This brings us to the big question: should the DOJ trust-bust Microsoft and Intel?

The Wintel Oligopoly - Is Enough, Enough?

Many of Wintel’s detractors look forward to the day when DOJ puts a stop to further advances of the Wintel oligopoly. But so far, DOJ has mostly kept its hands off. For example, in its first major encounter with DOJ, the dominant PC software company got the DOJ to back off in exchange for an end to price-fixing (Operating Systems license fees). The 1995 consent decree allowed Microsoft to monopolize the desktop operating systems space but prevented it from strong-arming vendors into buying other Microsoft "applications" in order to keep their Windows license. Note that Microsoft got what it wanted in 1995 - free reign over the operating systems market in exchange for a slap on the hand. This decree will be an important piece of paper in the legal browser wars that are about to take place in 1998.

This is nothing new. DOJ is constantly investigating Microsoft. In a lower profile investigation DOJ looked into Microsoft’s efforts to "buy control of the Internet" after the Redmond red meat-eaters bought, invested in, or partnered with over 20 Internet-related companies. Microsoft spent $2 billion in partnerships and acquisitions over the past 3 years - a drop in its $9 billion bucket of cash. For comparisons, see Table I.

One can argue the existence of an Internet domination pattern: Microsoft bought Vermeer ( FrontPage); Eshop to get a storefront server for e-commerce; Interse for its web traffic analyzer; WebTV (control of standards), VXTreme (video streaming), and entered into joint ventures with MSFDC (First Data Corp. handles electronic payment), and MSNBC (web-linked TV News venture with GE/NBC). It has stakes in VDOnet (video streaming), Progressive Networks (RealAudio), UUNet (Internet service backbone), and ComCast (Cable TV). Does an Internet emphasis show in these deals?

The giant software company hasn’t stopped its buying spree, and revenues keep adding to its cash reserves. Figure 1 shows IBM in the lead as the richest software company with 1996 revenues of over $13 billion. IBM’s income was greater than Microsoft and Oracle combined! But Figure 1 also illustrates what scares people. By the end of the decade, Microsoft will become the gorilla of software, and sprint ahead of the pack as we go beyond 2000. In short, the American public is worried that Microsoft is too rich and powerful to be a fair competitor.

What about Intel? It has been investigated twice in the past decade - once in 1991 - 1993 , and again in 1997. The current investigation centers around the claim that "Intel monopolizes or attempts to monopolize or otherwise restrict price or non-price competition in the development or sale of microprocessor or other computer components or related intellectual property," according to the recent subpoena. Intel revenues ($20 Billion) are far less than IBM, GM, or GE, but it has an 85% share of the global microprocessor market, and like any good competitor, is trying to vertically integrate the entire PC industry. Just as Ford Motor Company wants to make the whole car, rather than just the engines, Intel wants to make the whole computer. It’s new Pentium II slot (versus the older socket) has literally driven competitors off the road. It designs and sells glue chips surrounding the microprocessor, motherboards, disk drive circuit boards, modems, and networking cards. Recently, it raised eyebrows by announcing its intention to buy Chips & Technologies (graphics chips). Partnerships with networking companies and expansion into the lithography machinery and software segments suggests that the dominant chip maker will expand its influence over the entire industry - the DOJ willing.

While Microsoft is rich, Intel’s market share gives it a near-monopoly of the hardware business. Together, Microsoft and Intel form a cartel that some say is unbreakable. Shouldn’t this be stopped?

The Ida Tarbells of Software

Netscape and Sun Microsystems have positioned themselves as the Ida Tarbells of software. You remember Ida - the woman who brought down Standard Oil around the turn of the century. Her series of articles (later turned into a book) about how Standard Oil’s monopolistic price-fixing crushed its competitors eventually led to the break-up of Standard. Earlier, in 1887, the Interstate Commerce Commission had been formed to protect farmers against greedy railroad companies. The ICC set shipping rates and micro-managed the shipping business down to the smallest detail. Isn’t this kind of regulation necessary, again?

Many in the computer industry see an analog: Wintel is like the railroads in the 1880s because it controls the transportation system of desktop computing. Wintel is so big that it can control prices and the distribution of products to consumers, claim its detractors. Some even claim that Wintel crushes innovation by either stifling it with heavy marketing campaigns, or buying up the small guys before they have a chance to flourish. Said behavior has prompted Netscape to take legal action against Microsoft in what has become the "browser war" and Sun to launch its own crusade against Microsoft in what has become the "Java standardization war." AMD is fighting its own "slot war" against Intel, in a sideshow that is perhaps already lost as Intel cuts prices and enters the "under $550 PC" market segment. Business is war.

Netscape filed papers with the U. S. Department of Justice to halt Microsoft's alleged predatory pricing and bundling practices. Gary Reback, Netscape's counsel, charged that Microsoft offered clandestine payments, discounts, and other financial favors to companies who agreed to stop selling Navigator and started selling Explorer. He said, "Microsoft tactics include manipulating the disclosure of APIs, and the bundling of products such as FrontPage, Internet Explorer, and Microsoft's Internet Server with Microsoft's monopoly operating systems." Even Ralph Nader’s Consumer Project on Technology has entered the fray. In October 1997, it circulated e-mail to the DOJ asking DOJ to take action against Microsoft, and later held a conference that consisted mostly of complaining sessions. Nader’s raiders claim Microsoft should not be allowed to bundle Internet Explorer with its operating system in ways that are unavailable to other firms. Specifically, IE should not "own a place of prominence" on the screen to the detriment of other products.

I predict that DOJ will force Microsoft to back away from its bundling of IE with Windows 95, but this will have no effect on the industry, because Windows 98 will absorb IE. And because the 1995 decree permits Microsoft to add any capability it wants to its operating system, by the end of 1998 the industry will be right back where it is now - under the domination of the Wintel oligopoly. Bill Gate’s shrewd plan it to test the DOJ in 1997, and hammer them in 1998. If DOJ forces Microsoft to unbundle IE in 1997, Gates still has time to modify Windows 98 before it floods the market in the fall of 1998. If DOJ turns the other way, Gates will have cleared the way for an integrated Windows and IE in Windows 98. Either way, Microsoft is on top of the situation.

Fair Isn’t Free

Lets get real: the American free enterprise system is not really free. Rather it tries to emulate a fair enterprise system where the government mediates disputes among capitalists, environmentalists, socialists, and college professors like me. The difference between free and fair is subtle. In a fair enterprise system, government regulation is used to level the playing field, e.g. equalize advantages of size, market share, abundant resources, and sheer wit. Government’s role is to balance things. The question is, does the Wintel oligopoly need to be balanced? I argue that in the case of Wintel, the Ida Tarbells may have reason to fear and loathe Microsoft and Intel, but they don’t have a case for litigation. Wintel may be one tough competitor, but in general it is not playing on a tilted playing field. Remember, the DOJ is interested in protecting consumers, not competitors. As long as consumers have cost-effective alternatives to Wintel, DOJ won’t do much. Before you send hate e-mail, let me explain.

The Break-up of Microsoft

To argue in favor of a breakup, DOJ must prove that competition is dead or dying because its rules give an unfair advantage to Wintel. With a few exceptions - which have mostly been addressed already - the rules have been fair to the ABM (Anything But Microsoft) forces. Let me take you through a few of the arguments.

ABMs complain that Microsoft stifles competition because it is in too many segments of the industry. Specifically, it has an unfair advantage in the market because Microsoft makes both operating system and its applications. It should, therefore, be split up into at least 2 companies - an operating systems company and an applications development company. The key word here is "unfair," and the argument eventually leads back to Microsoft’s large percentage market share in operating systems and applications. But the argument is flawed because Microsoft’s dominance has not short-changed the consumer. Rather, it has been achieved because consumers refuse to buy non-Microsoft products in large quantities - even when they have cheaper alternatives. It is the customer who is being unfair, if anyone.

Consumers have had a choice among operating systems since the beginning of the PC era. In 1981, the original IBM PC shipped with three operating system options: MS-DOS, UCSD p-System, or CP/M-86 from Digital Research. Even today, consumers can opt for IBM’s OS/2, SCO’s UNIX, or Sun’s Solaris on Intel. And, if consumers do not want to be locked into an Intel platform, Apple’s Macintosh, SGI’s MIPS, or AMD K6 machines offer a choice. [Contrary to rumor, the Macintosh has been shown to provide a better solution with a total cost of ownership that is much lower than the Wintel solution. So, please do not complain that this alternative is more costly - it isn’t].

On the applications side, nearly every application that Microsoft sells has a competitor. Corel, Lotus, and Novell provide alternatives to MS Office. Of course there is Netscape for browsers and servers. Lotus probably has the best groupware - Lotus Notes/Domino Server versus Microsoft Exchange. In the tools arena, there is Borland, Symantec, and IBM as alternatives. [Lets be frank - perhaps the best development environments in several important categories do not belong to Microsoft, but instead belong to IBM (VisualAge for C++), Borland (Delphi for client/server), Symantec (Visual Cafe for Java), and Metrowerks (Cross platform development)]. It is difficult to claim that Microsoft has no competition with so many alternatives. If you think Microsoft is too big, stop buying its products!

A similar case against dominant toy maker Toys R Us erupted in 1997 when the Federal Trade Commission busted Toys R Us for restricting sale of Mr. Potato Head and Barbie Doll to the giant warehouse stores. By withholding best-selling merchandise from club retailers such as Sam’s and Price-Costco, Toys R Us was able to charge 10 - 20% more for its products. Like Microsoft, Toys R Us used its large market share to bully retailers. But, unlike Microsoft, Toys R Us got into trouble because its actions resulted in higher prices and fewer consumer options. This is the subtle difference: Microsoft’s actions result in lower prices - not higher. Microsoft does not prevent its competitors from selling their wares - it only makes it more difficult. In short, Microsoft makes competitors compete on price, quality, and volume - not availability. And as long as competitors exist, consumers have a choice. In this case, choice is what makes the market fair. Many of these choices come from companies that are as big and powerful as Microsoft, e.g. IBM. Consumers can opt for a non-Microsoft operating system, non-Intel hardware, and non-Microsoft applications. So, why don’t they?

After the Break-up

Suppose the DOJ did break Microsoft into smaller companies. According to economists Heilbroner and Thurow, an industry with one or two giant firms and a tail of smaller firms does not operate very differently - if at all - than does an industry with five or six leading companies. There are many examples, but the break-up of Ma Bell is the most dramatic illustration that breaking up monopolies does little to benefit consumers. Local telephone calls are more costly today than before the break-up, and innovation in the telephone business certainly has not accelerated.

There is one more irritation: unsavory practices - like threats - that limit innovation in the industry. A final argument in favor of breaking up Wintel claims Wintel kills innovation even before it can get started. Like Stac Electronics (file compression) and WebTV(Internet TV), many start-ups either get smothered by Microsoft or bought out in their early stages. This is one area where the rules may need to be changed. If innovative companies can prove Microsoft prevented them from bringing their innovative technology to market, litigation may be warranted. But, they would still have to convince DOJ that their products would ultimately benefit consumers - not merely themselves.

Several Microsoft’s are no better (for competition) than one Microsoft. In fact, Microsoft may need its size in order to survive the Next Big Thing in computing. Back in 1887 when the ICC was created to regulate the railroads, nobody could choose a car, plane, or truck over a box-car and train. Today, technology changes so rapidly that breaking up Microsoft might simply wipe out the software industry, altogether! According to Heilbroner and Thurow, events tend to change things faster than the regulations governing them. Internet Time is simply too quick for most governments.

A careful examination of Microsoft’s revenues reveals its vulnerability - it is a 2-product company. Without its upgrade business from MS Office and Windows 95 products, Microsoft would be a mere shadow of itself. According to venture capitalist Roger McNamee, Microsoft cannot sustain a $6-10 billion enterprise on revenues derived from a saturated operating systems and office suite market. "The ability of Microsoft and Intel to dominate the agenda of our industry may be peaking," says McNamee, who is a partner in Interval Capital Partners. In order to grow, Microsoft must enter some other market. The only ones big enough to generate $6- 10 billion are Internets and Intranets. Therefore, Microsoft must compete on the Internet and Intranet battlefield. Hence its buying spree of companies deep in Internet country.

 

The End of Intel

The currently on-going investigation of Intel by the FTC is similar to those that plague Microsoft. Intel’s $1.4 billion "Intel Inside" advertising campaign involves kick-backs of up to 6% of orders to the company’s PC customers. Competitors claim that this is more than they can match. Similarly, Intel’s patent-pending Pentium II slot "form factor" blocks competitors like AMD from competing, because x86 chips are no longer socket replaceable. Does this constitute an unfair advantage? I claim it does not, because consumers still have a choice between x86 designs and x86 vendors. They don’t even need to buy an x86 machine - Apple, SUN, SGI, HP, and ARM still live. Again, it is up to the consumer.

Like Microsoft, Intel’s future is not guaranteed. It is a single-product company, it is headed for a technological sea change, and it depends almost entirely on Microsoft products and their corresponding lock-in to keep its factories running. But Intel has one important advantage - it takes billions of dollars to ramp up the next generation fab facility, and not many competitors can afford to drop a billion or two into fab. Intel’s large cash position gives it a distinct advantage, but its x86 instruction set won’t last forever. What will Intel do after x86?

Clearly, the paranoids at Intel are seeking other sources of business. This is why the company is casting about in networking, software, and flash memories. The Merced partnership with HP - which will not bear fruit until 1999 - is rumored to significantly depart from the x86 instruction set. One misstep during this transition could end Intel.

Breaking up Intel would be equivalent to breaking up AT&T. While it may satisfy free marketers, it would not level the playing field any more than the breakup of AT&T leveled the playing field for telephone companies. The telephone industry is still dominated by 2 or 3 players. It is still an oligopoly. Similarly, microprocessor competition still thrives, especially in the non-PC segment called embedded systems. Motorola may not make many PC processors, but it dominates the cellular phone business, as well as automobile and other microprocessor-based businesses.

The Global Gig

The monopolistic practices of the turn of the twentieth century have to be viewed differently in the early morning light of the turn of the twenty-first century because Intel, Microsoft, Motorola, AT&T, and nearly all other USA corporations are playing a global gig. The competition is no longer inside a single country. Instead, it is all over the map. Microsoft and Intel may look ominous when viewed from a provincial American vantage point, but they are not so big when compared with the global giants roaming other parts of the world - companies like DaimlerBenz, Alcatel, Hitachi, and Samsung. Take Motorola versus Nokia versus Ericsson as just one example. Motorola is in a run for its cellular telephone life - not against USA companies, but against global competitors. Who should decide if 95% share of the USA market is sufficient? Probably not a USA government bureaucracy.

Over 100 years ago the US government argued similar case law and created the Sherman Anti-Trust law and ICC regulations. These laws did not prevent monopolies, in fact, they created so-called "natural monopolies". Economist Richard Posner defines "natural monopoly" as a situation in which (consumer) demand may be most economically and efficiently satisfied by a single producer. In addition, natural monopolies were allowed where competition resulted in wasted investment and ultimately, lower prices to consumers.

In today’s friction-free economy, economies of scale, efficiency, and wasted investment has been replaced by the momentum of increasing returns. In today’s new economy, consumer lock-in is more important than lower prices. After all, inverse economics relentlessly lowers prices while improving quality. In her insightful analysis of the telephone industry, Amy Friedlander says, "...the utility of a given technology increases as more people select that technology... the initial selection may be a matter of historical accident, rather than the result of economic efficiencies or technological superiority... as increasing returns tend to create a positive feedback that magnifies otherwise random variation, the process of adopting technology will tend to converge on a standard." [Natural Monopoly and Universal Service, Corp. for National Research Initiatives, Reston VA. request@cnri.reston.va.us]. Microsoft and Intel have successfully ridden the increasing returns wave to their monopoly positions. They did this by playing within 19th century regulations. If we want to prevent them from gaining more wealth and power, we must do so by kicking them off of the increasing returns wave. If the DOJ wants to solve the "Wintel problem" it must understand the theory of increasing returns, and congress must legislate against the positive feedback mechanisms that lock-in consumers to a single platform.

My guess is that Microsoft and Intel are guaranteed a long and prosperous future with little intervention from DOJ. Then, about the year 2017, the Wintel oligopoly will look as out-of-date and unimportant as Standard Oil is today.

 

Table I The top ten Silicon Valley companies in terms of cash position.

Company

Billions in the bank

Intel

8.1

HP

2.8

Seagate Technologies

2.3

Oracle

1.5

Cisco Systems

1.3

Apple Computer

1.2

Sun Microsystems

1.1

Applied Materials

1.1

3COM

1.0

LSI Logic

0.74

TOTAL

21.1

 

Figure 1. Revenues and Revenue Projections for IBM, Microsoft, and Oracle.