Computer Industry Case Study

Chapter 17

USING THE TOOLS: A CASE STUDY OF THE COMPUTER INDUSTRY*

A

lthough we have emphasized the theoretical aspects of the field of industrial organization up to this point, we hope you have gained an appreciation of the close relationship between the topics we have been studying and real-world markets. Knowledge of industrial organization is essential for policymakers in the areas of antitrust and regulatory policy. The remainder of this book emphasizes public policy in these areas. Making rational public policy decisions requires the ability to incorporate the tools of industrial organization into studies of how real-world markets work. When the Federal Trade Commission or the Department of Justice reviews the antitrust implications of a merger or the monopoly power of a particular firm, the agency begins with a detailed industrial organization study of the industry. Few industries have been under closer government scrutiny over the past 30 years than the computer industry. In 1969 the Department of Justice filed a major monopolization case against IBM, and throughout the 1990s Microsoft has been under the sword of various antitrust actions. To help you recognize the importance of an understanding of industrial organization in the context of public policy, this chapter presents a detailed study of the computer industry. By the end of the chapter, we hope you will appreciate how closely linked industrial organization is to public policy.

HISTORY 1
The history of the computer industry is in part the history of two firms: International Business Machines (IBM) and Microsoft. Although the two firms compete in different sectors of the market, their intertwined histories suggest how quickly
527
*“Computers” by Waldman, Don in Structure of American Industry 10/e edited by Adams/Brock, © 2000. Adapted by permission of Prentice-Hall, Inc., Upper Saddle River, NJ.

528

Chapter 17 Using the Tools: A Case Study of the Computer Industry

economic conditions can change in this technologically advanced industry. IBM, the firm that stood for high technology and computers for decades, is no longer the leading manufacturer of personal computers in the United States based on units shipped. IBM’s one-time dominance of the entire computer industry has been lost in part to a software manufacturer, Microsoft, and in part to a microprocessor manufacturer, Intel. A decade ago IBM produced over 60 percent of the mainframe computers sold in the world; its annual gross revenues increased every year, peaking at $68.9 billion in 1990; and it employed over 400,000 people around the world.2 By 1993 IBM’s total world employment had declined 36.8 percent to 256,000 and the company sustained an $8.1 billion loss! How did this turn of events happen? And what lessons can be learned about the relationship of market structure, conduct, and performance by analyzing this rapidly changing industry? Thomas J. Watson, Sr. built IBM by emphasizing to its sales force the importance of an understanding and commitment to its customers’ needs and a dedication to both quality and on-time delivery.3 Watson Sr. was a salesman, not a technological genius, and IBM’s CEOs for decades to come followed in his footsteps. Watson Sr. had taken a small firm, The Computing Tabulating Recording Company (name changed to IBM in 1924), that specialized in scales and measuring devices and, by focusing its attention on solving the accounting problems of large corporations, had built up a highly profitable business. Watson Sr. trained his sales representatives to have one overriding objective: to solve each customer’s problems with an individual solution. He was convinced that as long as IBM’s machines helped its customers’ businesses operate more efficiently, IBM would remain the industry’s dominant force.4 IBM did not start out as the leader in the computer industry. Remington Rand produced the world’s first important commercial computer, the UNIVAC. While Rand was producing the UNIVAC in the late 1940s, IBM was concentrating on producing large electronic business calculators that operated on vacuum tubes.5 In 1952 IBM introduced the IBM 702, an electronic calculator designed to replace its punch-card machines. When Thomas Watson, Jr. took control of the company from his father, he decided to push IBM into the computer age. IBM moved quickly to gain the lead in transistor technology and by 1957 its sales had reached $1 billion. Considering that in 1946 its sales were a mere $116 million, this was a remarkable achievement.6 Once business firms purchased an IBM computer, they quickly became dependent on IBM’s software. This phenomenon came to be known as software lock-in.7 Until the point of purchase of a computer system, buyers had a choice about what brand to install. After the purchase of an IBM computer, however, the buyer became completely dependent on IBM’s software and programming; the cost of switching to another manufacturer often became prohibitively high. Once a large corporation had paid IBM thousands, or millions, of dollars to install and organize its database, there was little chance it would switch to a new system to save a mere 10 or 15 percent in monthly leasing charges.

History 529

By the early 1960s IBM was firmly established as the computer industry’s dominant firm. In 1964 IBM held an 80 percent market share of the value of installed and on-order data processing equipment. Its profit in 1964 was 19.1 percent of stockholders’ equity, making IBM one of the nation’s most profitable companies. The development of the System 360 series of mainframe computers was key in IBM’s maintaining its dominant position in the 1960s and 1970s. At the time of the System 360’s development, IBM’s market share was being reduced by the inroads of such corporate giants as GE, RCA, Rand, and Control Data. The objective of the System 360 series was to replace all existing computers, including IBM’s existing machines. The idea behind the System 360 was that all of IBM’s computers would be compatible because they would all run the same software. Machines that leased for $2500/month would run on the same software as machines leased for over $100,000/month. Furthermore, the 360 could run all the programs that were currently running on IBM’s then leading computer, the IBM 1401. Therefore, the 1401 customers’ programs could be transferred directly to the 360. As a customer’s needs expanded, IBM would simply add new devices such as a faster processor or more memory. The success of the System 360 established IBM’s dominance for 20 years. Sometime in the mid-1970s, however, IBM began to lose its strangle-hold on the industry. Two factors were primarily responsible for this relative decline: the filing of a government antitrust case against the firm in 1969 and the failure of IBM’s management to respond to the changing nature of the computer industry as it moved rapidly from a world of large mainframe computers to a world of small personal computers. William Henry Gates III, the world’s wealthiest person, was born on October 28, 1955. In 1967 Gates enrolled in the seventh grade at Lakeside, an all-boys private preparatory school.8 Gates was first introduced to computers in the spring of 1968 when Lakeside bought a teletype machine. Gates, and his friends Paul Allen, Kent Evans, and Richard Weiland, received their first computer-related job from Computer Center Corporation (CCC). CCC signed an agreement with the newly established “Lakeside Programming Group,” where in exchange for lists of bugs in CCC’s software, the boys received free computer time. They spent their evenings and weekends on the computer, carefully documenting each bug. This experience gave Gates a chance to deepen his knowledge of the operation of the minicomputer’s hardware and software. In the fall of 1973, Gates entered Harvard. During the following summer, Allen and Gates worked at Honeywell. It was Paul Allen who pushed Gates to see the vision described by the following scenario: 9
Gates and Allen were convinced the computer industry was about to reach critical mass, and when it exploded it would usher in a technological revolution of astounding magnitude. They were on the threshold of one of those moments when history held its breath . . . and jumped, as it had done with the development of the car and airplane. Computer power was about to come to the masses. Their vision of a computer

530

Chapter 17 Using the Tools: A Case Study of the Computer Industry

in every home was no longer a wild dream. “It’s going to happen,” Allen kept telling his friend. And they could either lead the revolution or be swept along with it.

In December 1974 the January issue of Popular Mechanics had Altair 8080 “World’s First Microcomputer Kit to Rival Commercial Models” as the magazine cover headline. The article was about Ed Roberts of Micro Instrumentation and Telemetry Systems (MITS), who coined the term “personal computer.” Gates and Allen promised Roberts a version of an operating system, BASIC, for the Altair and worked solidly for eight weeks to provide the promised, but nonexistent, version of BASIC. Allen and Gates recorded Gates’s 8000 lines of machine language code onto punched paper tape because there were no plans for a keyboard for the Altair. The program worked and Microsoft was born. The later development of the MS-DOS operating system put Microsoft on the path to becoming the dominant firm in the computer industry. IBM initially approached Gates in 1980 to write a version of BASIC for the permanent memory of an 8-bit computer.10 Gates, however, suggested that IBM meet with Gary Kildall of Digital Research to acquire CP/M as IBM’s operating system. Once Gates, Allen, and IBM realized Kildall was not interested, Gates and Allen decided to provide the operating system themselves, but they knew they did not have sufficient time to develop a new system. Instead, they bought a newly developed operating system from Seattle Computer Products, QDOS. Microsoft paid Seattle Computer Products $75,000 for the right to sell QDOS.11 Microsoft and IBM signed a contract in late 1980 that prohibited IBM from licensing DOS, but placed no restraints on Microsoft. The stipulation in the contract that allowed Microsoft, but not IBM, to license MS-DOS is what enabled Microsoft to push MS-DOS as the industry standard by licensing it to clone manufacturers. When IBM announced the introduction of its first personal computer on August 12, 1981, the computer included Microsoft’s MS-DOS operating system. By 1984 it was clear that MS-DOS was the operating system industry leader. By June 1986 income from the sale of Microsoft’s MS-DOS operating system accounted for half of Microsoft’s annual revenues of $60.9 million.12 In the 1980s and 1990s three other important entrepreneurs, Michael Dell, Ted Waitt, and Mike Hammond, quietly began to impact the computer industry. Almost simultaneously, Dell, Waitt, and Hammond decided that many PC computer users were becoming sophisticated enough that they no longer needed the security associated with purchasing PCs from a large company such as IBM, HewlettPackard, or Compaq. Instead, Dell, Waitt, and Hammond believed consumers would be willing to purchase custom-made PCs directly through the mail at relatively low prices. The result was the founding of two small companies, Dell Computer Corporation by Michael Dell in 1984, and Gateway by Ted Waitt and Mike Hammond in 1985. In recent years Dell and Gateway have been the most rapidly growing PC manufacturers, and the other industry leaders have had to follow their leadership and offer PCs directly through the mail. The result has been a remarkable shift in the market shares of the top PC manufacturers.

Market Structure 531

MARKET STRUCTURE
Market Shares and Concentration
The large number of business contacts that IBM had cultivated as a result of its dominance of the tabulating machine market quickly led to its control of the computer market. By 1956 IBM held approximately a 75 percent market share of the “installed-base of the electronic data-processing market.” As Table 17.1 indicates, in 1956 five major firms were competing in this industry: IBM, Sperry Rand (formerly Remington Rand), Burroughs, RCA, and NCR. By the end of the 1950s four additional firms were in the industry: Honeywell, General Electric (GE), Control Data (CDC), and Philco. Furthermore, other giant firms such as Bendix, Budd, General Mills, ITT, Northrop Aircraft, Sylvania Electric, TRW, and United Aircraft all entered and left the industry prior to 1969. In the early years of the industry there was certainly no shortage of attempts by major corporations to enter, but none could ultimately compete effectively with IBM. If “the installed-base of electronic data processing equipment” is the correct market definition for measuring effective competition in the computer industry, then the figures in Table 17.1 leave little doubt that IBM had a considerable amount of market power. In 1967, despite the protestations of IBM and its lawyers, it may have made some sense to define the industry as “the installed-base of electronic data processing equipment.” However, this definition certainly makes no sense today. Table 17.2 shows IBM’s revenues by market segment for
TABLE 17.1 PERCENTAGE SHARES OF THE RETAIL SALES VALUE OF THE INSTALLED BASE OF ELECTRONIC DATA-PROCESSING EQUIPMENT FOR THE MAJOR SUPPLIERS IN THE UNITED STATES (1955–1967) Year 1955 1956 1957 1958 1959 1960 1961 1962 1963 1964 1965 1966 1967 IBM 56.1 75.3 78.5 77.4 74.5 71.6 69.3 69.9 69.8 68.3 65.3 66.2 68.1 Sperry Rand Honeywell 38.5 18.6 16.3 16.3 17.8 16.2 15.5 12.4 11.2 11.8 12.1 11.3 10.6 — — 0.3 0.9 1.2 0.9 2.0 2.3 1.8 2.5 3.8 5.2 4.7 RCA 5.1 1.6 0.8 1.8 1.4 2.4 3.0 3.5 3.5 3.0 2.8 2.7 3.2 NCR 0.3 0.1 0.1 0.1 0.1 0.4 0.8 1.9 2.7 2.8 2.8 2.4 2.5 Burroughs — 4.4 3.9 3.3 4.2 3.4 2.6 2.2 2.6 3.1 3.6 3.0 2.9 GE — — — 0.2 0.9 2.8 3.4 3.7 3.5 3.3 3.3 3.5 3.0 CDC — — — — — 1.0 2.2 3.1 4.0 4.4 5.4 5.3 5.7 Philco — — — — — 1.2 1.3 1.2 1.0 0.8 0.8 0.4 0.2

Source: Honeywell, Inc. v. Sperry Rand Civil Action 4-67, Civ. 138 US District Court, 4th Dist. Minn.

532

Chapter 17 Using the Tools: A Case Study of the Computer Industry

TABLE 17.2 IBM’S REVENUES BY SOURCE 1992–1998 Revenue Source Hardware segment Global service segment Software segment Global financing segment Enterprise investment segment/Other Totals
Source: IBM Annual Reports

1998 35,419 28,916 11,863 2877 2592 81,667

1997 36,630 25,166 11,164 2806 2742 78,508

1996 36,634 22,310 11,426 3054 2523 75,947

1995 35,600 12,714 12,657 7409 3560 71,940

1994 32,344 9715 11,346 7222 3425 64,052

1993 30,591 9711 10,953 7295 4166 62,716

1992 33,755 11,103 7352 7635 4678 64,523

the years 1992–1998. Notice how the relative importance of the “global services” segment has increased, while the relative importance of the hardware segment, which includes both mainframe and PC revenues, has decreased. By 1996 less than 50 percent of IBM’s revenues were generated from hardware, and in 1998 only 43.4 percent of revenues were generated from hardware. Table 17.2 suggests that IBM is rapidly transforming itself from a computer hardware company into an information technology service company. Several years ago the computer magazine Datamation began ranking the top 100 “information technology” firms. The market definition information technology as used by Datamation was quite broad, including everything from computer hardware and software to computer services. The list, known as the Datamation 100, provided insight into the large number of giant corporations competing in one or more segments of the broadly defined information technology industry. Unfortunately, 1996 was the last year of the Datamation 100 list. Table 17.3 shows the top 20 firms on the 1996 list. The list includes 13 American firms, 6 Japanese firms and one German firm. The complete list of 100 firms also includes firms from the United Kingdom, France, Italy, the Netherlands, South Korea, Sweden, and Taiwan. The “global information network” is indeed global. While some of the Datamation 100 firms are exclusively information technology firms, many are highly diversified. Notice that five of the top 20 receive less than 50 percent of their total revenues from information technology markets. Tables 17.4–17.6 present more recent market share data for the following information technology segments: high-end servers (including mainframe computers), personal or desktop computers, and software. By 1999 IBM led in only the high-end server market. In software, Microsoft surged past IBM in 1998, and in PCs, IBM ranked only fifth in units shipped for the first three-quarters of 1999. With the PC segment growing much more rapidly than the high-end server segment, it is perhaps most revealing that IBM declined from the third-ranked to the

Market Structure 533

TABLE 17.3 1996 DATAMATION GLOBAL 100 Total 1996 IT Revenue 75,947 31,398 29,717 18,109 15,242 15,092 14,441 14,050 13,610 9435 9189 8914 8500 7800 7500 7000 6907 6410 6403 6390 Total 1996 Net Income 5400 2708 5300 1313 712 841 432 1533 ( 343) 2476 20 ( 867) 222 518 N/A N/A 1000 ( 537) ( 109) 447 Total 1996 Revenue 75,947 39,427 47,170 18,109 68,735 44,766 14,441 58,300 13,610 9435 9189 8914 8500 7800 7500 7000 10,430 64,102 6960 6390 IT Revenue as a % of Total Revenue 100 80 63 100 23 34 100 24 100 100 100 100 100 100 100 100 51 10 92 100

Rank 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20

Company IBM Hewlett-Packard Fujitsu Compaq Hitachi NEC Electronic Data Systems Toshiba Digital Equipment Microsoft Siemans Nixdorf Apple Computer Seagate Technology Dell Computer Packard Bell NEC Acer Canon Matsushita NCR Sun Microsystems

Country US US Japan US Japan Japan US Japan US US Germany US US US US US Japan Japan US US

Source: Datamation, July 1997.

fifth-ranked PC supplier in less than two years. Tables 17.4 and 17.5 suggest that the current computer hardware industry is ripe with competition. Many firms compete effectively in both the high-end and PC markets. The days of complete domination by IBM appear to be long gone. However, lurking behind the figures in Table 17.6 is the fact that in the market for operating systems for “non-Apple PC-computers,” Microsoft has a virtual monopoly with approximately a 90 percent market share.

Entry Barriers
According to one engineering study, in 1967 one minimum optimal scale computer plant produced 15 percent of total United States demand for computers. Furthermore, a plant operating at one-third minimum optimal scale had long-run average costs elevated by 8 percent.13 Those figures indicate that substantial production economies of scale existed in 1967. Furthermore, we have seen that econ-

534

Chapter 17 Using the Tools: A Case Study of the Computer Industry

TABLE 17.4 WORLDWIDE HIGH-END SERVER MARKET SHARE 1998 Company IBM Fujitsu Hitachi NEC Sun Amdahl Others
Source: International Data Corporation.

1998 Revenues (billions of dollars) 6.0 1.4 1.1 1.1 0.9 0.7 16.3

Market Share 36.8 8.6 6.7 6.7 5.6 4.4 31.1

TABLE 17.5 UNITED STATES PC SHIPMENTS (THOUSANDS OF UNITS) 1999 First 3 Quarters Rank Company 1 2 3 4 5 Dell Compaq Gateway Hewlett-Packard IBM Others
Source: International Data Corporation.

1999 First 3 Quarters Shipments 5343 5251 2846 2682 2579 13,736

1999 Market 1998 Share Shipments 16.5% 16.2% 8.8% 8.3% 8.0% 42.3% 4799 6052 3039 2814 2979 16,571

1997 Market 1997 Share Shipments 13.2% 16.7% 8.4% 7.8% 8.2% 45.7% 2930 5316 2219 2063 2719 16,232

1997 Market Share 9.3% 16.9% 7.0% 6.6% 8.6% 51.6%

TABLE 17.6 TOP SOFTWARE COMPANY REVENUES 1996–1998 1998 Software Revenues 1997 Software Revenues enues Company Microsoft IBM Oracle Computer Associates (billions of dollars) 16.7 13.5 8.0 5.1 (billions of dollars) 12.8 12.8 4.4 4.5 1996 Software Rev (billions of dollars) 9.2 13.1 3.6 3.9

omies of scale are also associated with research and development expenditures, which totaled several hundred million dollars for the IBM System 360 family of computers.14 The greater the number of units over which these R&D expenditures could be amortized, the lower the average R&D expenditures per machine. Production and R&D economies of scale were important in the first few decades of the computer industry, but network externalities in combination with these

Market Structure 535

economies of scale were probably more responsible for IBM’s market domination. Network externalities exist if “the utility that a user derives from consumption of the good increases with the number of other agents consuming the good.” 15 For example, the utility that a consumer derives from a telephone is clearly related to the number of other users on the telephone network. Consider a consumer contemplating the purchase of a computer. If consumers as a group make their purchasing decisions at different times, then each consumer’s utility will be affected by both the number of consumers already using a given model of computer and the number of consumers expected to use the system in the future. The more consumers currently using or expected to use the same computer, the greater will be the availability of software applications for that computer, and the greater will be the consumer’s expected utility associated with the computer. In addition, the greater the number of current users of a particular computer system, the easier it will be to obtain post-purchase service and technical support. The current demand for a given computer, therefore, is a function of the current installed-base and expected future demand for the computer. If computers are produced under conditions of economies of scale, then the first firm, in this case IBM, to gain a large installed-base will have significantly lower costs and can charge significantly lower prices. As a result of lower prices and network externalities, the market may be “tipped” to one dominant product. Once a market is tipped, it may be difficult for competitors to “untip” it. As the first mover into the broad-based computer industry, IBM gained a product differentiation advantage as well.16 Consider the market for an expensive new computer system, such as IBM’s System 360. Because consumers were initially completely uninformed about the system, they risked buying the System 360 and disliking it. As we saw in Chapter 12, uninformed consumers will be willing to pay less for a product than informed consumers; therefore, the demand for the System 360 prior to its introduction was quite low. IBM decided to introduce the System 360 at a low introductory rental rate to induce consumers to lease it. After the introductory offer ended, consumers who leased the system were informed and their demand increased dramatically. As a result, IBM significantly increased the rental rate to informed consumers. The demand by uninformed consumers, however, remained low. As other companies developed computers that were more or less identical to IBM’s System 360, they faced a serious problem: Consumers could not be certain that these other computers were as good as IBM’s. In fact, consumers could not even be sure these computers were any good. Under these conditions, firms found it difficult to enter and earn a profit, particularly because there are significant sunk costs associated with entry. Furthermore, IBM’s ability to respond aggressively and lower rental rates to capture additional uninformed consumers served as an additional deterrent to entry. Network externalities and a first mover advantage were also critical factors explaining Microsoft’s domination of the PC operating systems market. These two structural characteristics, therefore, played a very important role in the computer industry.

536

Chapter 17 Using the Tools: A Case Study of the Computer Industry

CONDUCT
As the leading firms in their respective segments of the computer industry, IBM and Microsoft adopted strategies to maintain and expand their market power. In fact, both firms have used remarkably similar types of strategic behavior to build and maintain their market dominance.

IBM’s Early Conduct
Pricing to Educational Institutions Thomas Watson, Sr. realized that the typical business computer user, as distinguished from scientific computer users, lacked any sophisticated knowledge of computer usage. IBM was determined to fill this educational void with its own sales force, so it committed to making computers available to educational institutions at huge discounts in an effort to tie students to IBM products after graduation. This strategy worked extremely well. In the 1960s an entire first generation of computer users was trained in college on IBM machines. Upon graduation these former students turned to IBM when purchasing a computer. Pricing of Service Because early computers were dependent upon vacuum tubes rather than semiconductors, they were constantly breaking down. A system crash created complete havoc in a company trying to process its monthly payroll. To avoid this, companies were willing to pay premium prices to IBM to gain access to the company’s service network.17 IBM covered the nation with service centers to take care of problems, while smaller competitors had far fewer service providers nationwide. This extensive service network led customers to rely on IBM to provide fast, low-cost service. To further reinforce its service advantage, IBM adopted a pricing policy of bundling service with its sale or lease price, so that the marginal cost of an additional service call was zero. The tying of service to IBM’s sale and lease prices prevented the development of an independent service industry to service all brands of computers. The development of a national independent service market, like the one that currently exists for personal computers, would have made entry into the computer industry easier by providing efficient, quick service for all brands anywhere in the country. Pre-announcement of the System 360 In the mid-1960s, as the time for the introduction of the System 360 approached, IBM was concerned that several competitors would introduce technically superior machines before IBM introduced the System 360 models. One solution to this problem was for IBM to “pre-announce” all five main System 360 models well in advance of when most of these computers would be ready for delivery. In April

Conduct

537

1964 IBM announced the “forthcoming availability” of five System 360 models even though several of these models were not fully developed and many technical problems remained to be dealt with before delivery would be possible. The decision to pre-announce the introduction of the System 360 line proved to be a brilliant business strategy. Despite the fact that the first machines were not delivered until April 1965 and that some of the promised models were never developed, the mere announcement of IBM’s coming new line prevented buyers from committing to competitors’ new machines, some of which were probably technologically superior to the System 360 computers. Price Discrimination One of the most important strategies was the selective use of price discrimination. IBM bundled entire systems together, charging one price for the central processing unit (CPU), related software, and maintenance, rather than charging different prices for each component of a system. By bundling systems, IBM was able to discriminate effectively between different users because the price a user paid for an individual piece of equipment was unknown. Once a customer made a decision to buy from IBM it received an entire computer system, eliminating the possibility of competition for any components from competitors. Pricing in Response to Entry The bundling system worked extremely well for IBM until customers were ready to upgrade their systems in the late 1960s. IBM charged high prices for replacement equipment in order to cover the costs of providing free service to all customers and free systems to educational institutions. In November 1967 the first of many small firms began to market individual pieces of “plug-compatible IBM equipment,” such as tape and disk drives, at much lower prices than those charged by IBM. The market for plug-compatible computer equipment grew rapidly in the late 1960s. In 1966 only three firms produced plug-compatible equipment, whereas by 1972 there were approximately 100 manufacturers.18 In the late 1960s these firms were offering IBM-plug-compatible equipment at prices well below those of IBM. As a result, IBM lost a significant share of the peripheral market. IBM responded to these inroads with a series of actions. It selectively and dramatically reduced the price of peripheral equipment. IBM also redesigned certain equipment so that it could be connected to the IBM central processing unit only through an integrated file adapter, rather than through the traditional external disk control unit. This change made the use of non-IBM equipment much less attractive to users. In 1971 IBM announced a fixed-term leasing plan, which provided for an 8 percent price reduction for users signing a one-year lease on peripheral equipment and a 16 percent price reduction for signers of a two-year lease. Finally, at the same time that IBM announced large price reductions on peripheral equipment, it announced large price increases on its central processing units to recoup lost revenues.

538

Chapter 17 Using the Tools: A Case Study of the Computer Industry

IBM’s actions forced the manufacturers of peripheral equipment to cut prices drastically. These actions resulted in a series of private antitrust actions by the manufacturers of peripheral equipment.19 One of the unanticipated effects of IBM’s new peripheral pricing strategy was entry into the mainframe computer (CPU) market. The first major entrant was Amdahl Corporation, founded by Gene Amdahl after he left IBM. In 1975 Amdahl introduced the Amdahl 470 computer, which was fully compatible with IBM’s best System 370 computer (the technologically advanced successor to the System 360), but was much faster and priced much lower.20 By 1978 several other firms were producing CPUs to challenge IBM’s machines.21 Threatened with a significant decrease in its CPU market share, IBM once again responded with a series of actions. In 1977 it introduced a more powerful computer, the 3033. Later in 1977 IBM reduced prices on its System 370 computers by between 20 and 35 percent.22 Introduction of the IBM-PC The computer world, and the entire world, changed forever in 1971 when Intel developed the first microprocessor, which made it possible to put an entire computer on a small silicon chip. Continued technological advances in the 1970s led to the first commercially successful personal computer—the Apple. It was not long before an independent software market developed for the Apple. Once the Apple could be used in combination with both word-processing and spreadsheet software, it became a low-priced alternative to expensive computers for small business and educational uses. As Apple computers became more popular, IBM was left out in the cold. Much of the blame lies with IBM’s management, which failed to recognize that the personal computer would one day be something more than a home toy. IBM’s senior management in the 1970s and early 1980s thought that the future was still in mainframes for commercial use. According to one member of the IBM Management Committee, “[t]he general attitude was that you don’t have big problems in small markets, and we thought the personal computer was a very small market.” 23 In an almost desperate attempt to catch up in the PC market, IBM broke with all of its corporate traditions and set up an independent development team in Boca Raton, Florida. The team was charged with catching up to Apple ASAP! In its rush to catch up, IBM decided to work with an open architecture (as opposed to a proprietary architecture) so that other firms could develop software and peripherals for the IBM-PC. The strategy gave software and peripheral manufacturers a great incentive to produce products for the IBM-PC. It also meant that when IBM’s first PC was introduced in 1981, it was filled with components and software produced by other companies. As a result, IBM lost proprietary control over the computer industry, something it has yet to win back. More advanced IBM-PCs were introduced in 1983 and 1984 and IBM was earning record profits that peaked at $6.58 billion in 1984. At this point IBM made a series of terrible managerial decisions that set the stage for the company’s financial woes in the early 1990s.

Conduct

539

Recall that Bill Gates and Paul Allen had little experience writing operating systems and they had little time to develop one, so they turned to a small local firm, Seattle Computing, for help and purchased QDOS. In concert with a large group of IBM programmers, Microsoft went to work around the clock to modify QDOS. The result was one of the most significant commercial products of the latter half of the 20th century—MS-DOS. IBM management immediately made the first of three large tactical errors in dealing with Bill Gates and Microsoft. IBM failed to secure the rights to the source code for DOS. Without access to the source code, IBM could not modify DOS and was completely dependent on Microsoft for any improvements.24 IBM’s second error occurred when it had an opportunity to establish its own operating system in direct competition with MS-DOS. A group of IBM researchers wrote an operating system called CP/X86 for the PC-AT.25 Later they expanded the project to include a pictorial program that worked much like Windows. By 1984 the IBM group had developed a product that was clearly superior to DOS. When IBM management had to decide what operating system would succeed DOS, they decided to develop the OS/2 operating system with Microsoft. Apparently a major reason for their decision was Gates’s threat that if IBM developed an operating system without Microsoft, Gates would never release the source code for DOS to IBM.26 This would mean that IBM would be on its own with CP/X86 in direct competition with Microsoft’s MS-DOS. IBM feared that because of network externalities it would be difficult to establish CP/X86 as the industry standard. The OS/2 partnership between IBM and Microsoft was a disaster from the beginning, in large part because OS/2 was tied to a weak Intel microprocessor.27 By the time the project began, Gates knew that success depended on Intel’s 386 microprocessor chip, but IBM management was committed to the 286 chip. Gates thought the 286 chip was “brain dead,” but went along with IBM’s decision.28 One possibility is that Gates went along with the 286 decision in order to sabotage the OS/2 project so that OS/2 would not threaten Microsoft’s Windows when it was introduced. In any event, OS/2 was a dinosaur when it was introduced along with the Personal System 2 in 1987.29 The third IBM error ended any hopes it may have had of retaining dominance over the computer industry. Because IBM and Microsoft had jointly developed DOS, IBM was in a position to negotiate with Microsoft over how to split the royalties on DOS sales and the royalties on any future operating systems (e.g., OS/2) that the two firms might develop jointly. In 1985 IBM had one final opportunity to prevent Microsoft from taking control of the personal computer operating system industry. At that time, IBM had 80 percent of the PC business, its mainframe revenues were doing extremely well, and it had earned $6.58 billion. IBM’s management was convinced that the old adage, “nobody was ever fired for buying from IBM,” would continue to keep IBM firmly established as the industry leader regardless of future competitive threats. IBM’s management had one objective in its negotiations with Gates over a royalty split: to pay as low a price as possible for the rights to DOS on IBM machines.30 Because all other PC producers combined controlled only a 20 percent market

540

Chapter 17 Using the Tools: A Case Study of the Computer Industry

share, IBM cared little about royalties on these non-IBM machines. Gates then pulled off a master negotiating stroke: He offered to give DOS to IBM essentially for free! All Microsoft wanted in return was the right to collect the royalties on DOS from clone manufacturers. IBM quickly agreed, and a contract was signed in 1985. Following this third error there would be no more opportunities for IBM to regain control of the operating system market. Without control over either the operating system or the microprocessor for the IBM-PC, IBM was unable to prevent a flood of low-priced clones from capturing a growing share of the personal computer market. Once consumers became aware that the quality of the PC clones was at least equal to the quality of the IBM-PC, buyers bought the cheaper clones in droves. Within a decade, IBM was relegated to being one of the pack in the personal computer market. Conduct in the Computer Industry Today The elimination of IBM’s monopoly in the computer hardware industry has greatly changed industry conduct. Most of the strategic conduct used by IBM to gain control of the computer hardware industry would be impossible today. Consider the following comparisons:
IBM’s Conduct 1. Bundling of custom-made systems that included all hardware software and service. 2. Pre-announcing of nonexistent systems. 3. Price discrimination by charging for a complete system. Educational institutions often received free systems. Current Industry Conduct 1. Manufacturers charge for each separate component. “Basic systems” include a microprocessor and an operating system. Buyers pay for extended service contracts or additional upgrades. Buyers pay separately for each upgrade to their systems. 2. Technological change is so rapid and system upgrades are so easy that buyers would not delay the purchase of a new computer because of the pre-announcement of a nonexistent machine. 3. Firms charge for each component of a system. If a buyer wants another piece of peripheral equipment, such as a zip drive or high quality soundboard, the consumer pays separately. Educational institutions pay for equipment, although most companies still offer some grants to schools.

The elimination of IBM’s monopoly power in the computer hardware industry has eliminated the monopolistic conduct associated with that power. Today producers of computers quote virtually all buyers identical list prices for a given system or piece of equipment. A large number of firms also produce peripheral equipment for all types of machines. Market conduct in the computer hardware industry is now consistent with that of a competitive industry.

Conduct

541

Microsoft
The Pre-announcement of Products In 1990 Digital Research Incorporated (DRI) introduced a new operating system DR-DOS 5.0 to critical industry acclaim. Most experts considered DR-DOS to be superior to Microsoft’s MS-DOS.31 DR-DOS began to make significant inroads into MS-DOS’s market share; by the end of 1990, DRI had captured a 10 percent market share of new DOS-based operating system shipments.32 Within one month of the introduction of DR-DOS, Microsoft announced the “forthcoming” introduction of a new version of MS-DOS, MS-DOS 5.0. Although Microsoft’s announcement made the introduction of MS-DOS 5.0 appear imminent, it took Microsoft over a year to market MS-DOS 5.0, which was finally introduced in July 1991. According to the Federal Trade Commission staff, in the summer of 1990 the new version of MS-DOS 5.0 was nothing more than “vaporware,” meaning that Microsoft was not even close to having a marketable product.33 To make matters worse for DRI, during the year between June 1990 and June 1991, Microsoft continuously announced that its introduction of MS-DOS 5.0 was “imminent.” The vaporware strategy worked, and the growth of DR-DOS was stopped. By 1992 Microsoft’s share of DOS-based operating systems had increased to 81 percent and DRI’s share had shrunk to just 7 percent. Exclusionary Licenses Typically a patent license is based on the number of units of the product used by the licensee. Beginning as early as 1983, however, Microsoft negotiated many, but not all, licenses based on the total number of computer processors shipped by an original equipment manufacturer (OEM). Furthermore, every Microsoft license was individually negotiated; an official list price for MS-DOS never existed. This policy ensured that Microsoft received a royalty on every computer shipped by an OEM regardless of whether the computer had a copy of MS-DOS installed. Licenses based on the total number of computers shipped came to be known as CPU licenses. These licenses gained an increasing share of Microsoft’s licensing activity over time, increasing from 20 percent in 1989 to 50 percent of licenses by 1992.34 Microsoft’s CPU licenses required OEMs to pay for a minimum number of licenses, which was typically greater than or equal to the anticipated computer shipments of the OEM. If, for example, an OEM expected to ship 100,000 computers in a year, it would agree to pay a license fee, f, on a minimum of 100,000 units. If it shipped machines with a non-Microsoft operating system, say PC-DOS or DR-DOS, it paid a royalty to Microsoft. If it shipped less than 100,000 units, it still paid Microsoft for 100,000 licenses. If it shipped more than 100,000 units, the licensee agreed to pay an additional royalty of f per unit shipped. Once again the licensee paid the fee regardless of whether the units shipped included MS-DOS. The CPU licenses set the marginal cost of installing additional units of MS-DOS equal to zero up to the minimum number of units agreed to in the license, and

542

Chapter 17 Using the Tools: A Case Study of the Computer Industry

then the marginal cost per license jumped to f for each CPU shipped beyond the minimum number. By effectively setting marginal cost equal to zero, the CPU licenses created a system whereby an OEM had to pay double to install a non-Microsoft operating system. Under such a pricing system, buyers would pay for a second system only if the system was technologically superior to MS-DOS. Furthermore, the technological advantage had to be worth the full cost of the second operating system. This is a difficult hurdle to overcome for a competitive operating system such as PCDOS or DR-DOS. The CPU licenses typically ran for two years. Furthermore, Microsoft usually tried to negotiate licenses that set a higher minimum number of units than expected shipments by offering a lower per unit shipped fee. At the end of a year the licensee typically had a credit for unused licenses, but Microsoft would only permit the licensee to carry the credit forward into the next year. This created an incentive for OEMs to install MS-DOS-based operating systems on next year’s machines. In addition to imposing the implicit penalties associated with the use of a competitor’s operating system, Microsoft utilized direct penalties for the installation of non–MS-DOS operating systems. An OEM using large numbers of non–MS-DOSbased systems might be prohibited from carrying forward unused credits from previous years, or it might be required to agree to a higher minimum number of licenses in future years. In addition, technical support services might be withheld from firms installing non-Microsoft operating systems. Finally, Microsoft was known to increase the price of Windows for firms installing non–MS-DOS operating systems on their computers. Creating Technical Incompatibilities Having access to the application programming interfaces (API) for Windows is critical for the developers of alternative operating systems and software. Without the APIs, coordination between Windows and either another operating system or applications software is very difficult. Microsoft left some of the important APIs undocumented, thus creating problems for the programmers of competitive versions of DOS, and giving Microsoft’s application software, such as Microsoft’s word processing software Word and spreadsheet Excel, an advantage over competitive software such as WordPerfect and Lotus. Furthermore, even if competitors discovered the necessary APIs, Microsoft could subsequently change any undocumented interfaces to sabotage compatibility. Because it is generally in Microsoft’s interests for application developers to develop a large variety of application software for Microsoft’s operating systems, Microsoft has made advance copies of its new or updated operating systems available to software programmers. The software developers then test the compatibility of their software with the new or updated operating system. The compatibility testing procedure is referred to as beta testing. In order to create compatibility problems for DR-DOS, Microsoft denied access for beta testing to DRI when Microsoft introduced Windows 3.1 and Windows for

Conduct

543

Workgroups. In addition, Microsoft’s version of Windows for beta testing was programmed to check to see if the program developers were using DR-DOS. If they were using DR-DOS, an error message appeared on the screen asking the programmer to contact Microsoft for a version of MS-DOS or else risk incompatibility with DR-DOS. In reality there were no incompatibility problems between Windows and DR-DOS, and if the programmer ignored the error message and continued working, everything would work fine. The error message, however, was intended to create uncertainty in the minds of the programmers, suggesting that running Windows with a non-Microsoft operating system would result in serious problems. Bundling of Products In the application software market Microsoft typically competes against companies that produce one major product, for example, WordPerfect in word processing, Lotus in spreadsheets, and Intuit in home financial management. Beginning in 1990 Microsoft bundled several pieces of application software together and sold them as an application “suite.” Today the dominant application suite is Microsoft Office. The price of the bundled pieces of software in Office has been greatly discounted. Buyers can purchase the standard version of Office, including Word, Excel, PowerPoint and Outlook, for approximately $400. This represents a huge discount compared to the retail list price of approximately $300 per individual application. Office has not only resulted in a greatly increased market share for Word and Excel, but has established PowerPoint as a major factor in a market where Microsoft had previously played only a minor role. Microsoft also bundled its Internet browsing software, Internet Explorer, with Windows. Microsoft realized that the Internet provided a potential platform for weakening its operating system monopoly. Gates feared that software could be written that would enable PCs to interact directly with the Internet without going through a PC operating system. Such software makes it theoretically possible to develop a new operating system that could be downloaded directly from the Internet. Netscape was the first company to produce a high-quality Internet browser, Netscape Navigator, that permitted users to easily surf the Internet. Netscape Navigator quickly became the dominant Internet browser. In 1995 Microsoft recognized Netscape’s control of the browser market as a long-term threat to its operating system monopoly and made the strategic decision to gain control of the Internet browser market by bundling Microsoft’s Internet Explorer with its Windows operating systems. By controlling the browser market, Microsoft hoped to control use of the Internet and prevent the development of alternative PC operating systems that could be easily downloaded from the Internet. While Microsoft may well have used bundling of software applications to gain market share for its application software programs, it is unclear that the net effects on welfare were negative. For one thing, there is little doubt that the net impact of bundling was to lower the price of many Microsoft application programs. Furthermore, Microsoft has been unable to gain a large market share in markets where its application software is significantly inferior. The best example

544

Chapter 17 Using the Tools: A Case Study of the Computer Industry

of this inferiority is the continuing dominance of Intuit’s Quicken in the home financial management market. In this market Microsoft Money has failed to gain a large share.

PERFORMANCE
Static Efficiency
Are prices in line with costs in the computer industry? Let’s examine the evidence. There was relatively little in the way of effective competition for IBM in the hardware market until the mid-1980s. Until then IBM’s net income increased every year except for a slight decline in 1979. In 1986, however, net earnings suddenly declined dramatically. The 1986 decline was only a precursor for IBM’s experience in the early 1990s when the company suffered through three consecutive years of staggering losses, culminating with an $8.1 billion loss in fiscal year 1993. As Table 17.7 indicates, the past decade has brought dramatic swings in the net earnings of all of the major desktop computer manufacturers. In 1998 Compaq saw its net earnings fall by $4.5 billion to a $2.7 billion loss. In 1997 Gateway saw its net earnings decline by over 50 percent. Meanwhile Apple sustained huge losses in 1996 and 1997, but recovered to earn over $300 million in 1998. Even the Dell miracle suffered a setback in 1993 when the company sustained a $36 million loss. The picture that emerges is of an industry characterized by great volatility in the performance of individual firms from year to year: a classic competitive pattern. The current competitive nature of the hardware industry is further evidenced by the rapid decline in prices in recent years. Consider the following recent evidence of effective price competition: 1. In July 1997 Compaq reduced the price of its corporate PCs by 22 percent, Hewlett-Packard followed with a 24 percent cut, and Digital Equipment cut prices by 21 percent.35 2. In November 1997 IBM introduced its first PC priced under $1000, the Aptiva E16. That same month, Compaq reduced the prices of eight Presario models by between 15 and 25 percent, bringing the cost of its lowest priced model down to $799.36 3. In March 1998 Compaq again cut prices by up to 11 percent on some corporate PCs, and IBM responded the next day with price cuts of up to 20 percent on its corporate PCs. 4. In the second quarter of 1998, prices were 10 percent lower than in the first quarter as IBM, Compaq, and Hewlett-Packard all reduced prices.37 5. According to a July 1999 report, “[t]he cost of computers has fallen dramatically in recent years, with the average selling price of a machine down to $1,200 from $1,640 three years ago. . . . That decline accompanies dramatic improvements in processing, speed, storage capacity, and video performance.” 38

Performance 545

TABLE 17.7 NET INCOME FOR THE LEADING AMERICAN PC MANUFACTURERS 1990– 1998 (MILLIONS OF DOLLARS) Year 1990 1991 1992 1993 1994 1995 1996 1997 1998 IBM 5967 2861 4965 8101 3021 4178 5429 6093 6328 Compaq 455 131 213 462 988 893 1318 1855 2743 Dell 27 51 102 36 149 272 518 944 1460 Apple 475 310 530 87 310 424 816 1045 309 HewlettPackard 739 755 549 1177 1599 2433 2586 3119 2945 Gateway 17 39 106 151 96 173 251 110 346

Source: Company annual reports.

TABLE 17.8 MICROSOFT’S SALES, PROFITS, AND R&D SPENDING 1987–1999 Sales (Millions of Dollars) 346 591 804 1183 1843 2759 3753 4649 5937 8671 11,963 15,262 19,747 Net Income (Millions of Dollars) 72 124 171 279 463 709 953 1146 1453 2195 3454 4490 7785 Return on Stockholders’ Equity 20.8% 33.0% 30.4% 30.4% 34.3% 32.3% 29.4% 25.8% 27.2% 31.8% 32.0% 27.0% 27.4% R&D Expenditures 38 70 110 181 235 352 470 610 860 1432 1863 2897 2970 R&D as a % Sales 11.0% 11.8% 13.7% 15.3% 12.8% 12.8% 12.5% 13.1% 14.5% 16.5% 15.6% 19.0% 15.3%

Year 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999

Source: Microsoft’s SEC annual shareholder’s filings 1987–1999.

While IBM has struggled through some difficult profit years, Table 17.8 indicates that Microsoft has experienced a series of remarkably strong financial years. Between 1987 and 1999, Microsoft’s annual sales increased by 5607 percent and its net income increased by 10,713 percent! Perhaps even more impressive are the very high rates of return on stockholders’ equity, ranging from a low of 20.8 percent in 1987 to a high of 34.3 percent in 1991. Microsoft’s monopoly power has clearly translated into above-normal economic profits.

546

Chapter 17 Using the Tools: A Case Study of the Computer Industry

Dynamic Efficiency
The last column in Table 17.8 shows Microsoft’s R&D expenditures as a percentage of sales. The numbers are astounding. Microsoft has consistently spent over 11 percent of its sales on R&D. By comparison, IBM spent $5 billion on R&D in 1998, or 6.2 percent of its net sales. In 1998 Microsoft and IBM together spent $7.9 billion on R&D or 3.6 percent of the nation’s total R&D spending; the entire computer industry accounted for $22.5 billion or 10.2 percent of the nation’s total R&D spending.39 Schumpeter surely would point to the development of the computer industry as a classic example of creative destruction. The computer industry, and particularly the personal computer industry, may have experienced the most rapid rate of technological advance of any industry in history. Most of the technological advances over the last two decades have resulted from Intel’s rapid development of new microprocessors. Table 17.9 depicts these furious advances. The first microprocessor, the Intel 4004, had only 2300 transistors, a speed of 108KHz, and 640 bytes of memory; it was only good for powering calculators. By comparison, the Pentium II processor introduced in 1997 had 7.5 million transistors, speeds over 300 MHz, and 64 gigabytes of memory! Furthermore, the new Pentium III processor operates at speeds of 1130 MHz. The power of microprocessors is commonly measured in MIPS, or millions of instructions per second. The Intel 4004 was rated at 0.01 MIPS, while the Pentium II was rated at over 25 MIPS.40 Intel expects to have a new processor available shortly that will be rated at more than 500 MIPS! The rapid rate of technological advance in the microprocessor industry has been described in Moore’s Law. Moore observed that each new chip contained approximately twice as much capacity as the previous chip and that new chips came into production about every 1.5 to 2 years. Moore reasoned that at that pace, computing power would rise at an exponential rate over relatively short periods of time. Clearly, the rate of technological advance continues to proceed at a remarkably fast pace. Until recently, Intel was a virtual monopolist in the market for microprocessors for non-Apple PCs. However, two new firms, Advanced Micro Devices (AMD) and National Semiconductors (NSM), have marketed high-quality chips priced below $1000. These firms should compete with Intel for market share in the next few years. All of the computer manufacturers have been quick to incorporate the newest microprocessors into their PCs. The rate of diffusion of new technology, therefore, has been very rapid. In terms of both static and dynamic efficiency the computer hardware industry has performed well in recent years. Most importantly, new technology is rapidly developed and marketed. Furthermore, prices based on dollars per MIPS have been declining rapidly, as have the prices of both mainframe and personal computers. The performance picture is not quite as clear with regard to PC operating systems, where Microsoft has controlled the market with a system that may or may not be the most technologically advanced. Here is where recent public policy has played a significant role.

Public Policy Issues 547

TABLE 17.9 INTEL’S INTRODUCTION OF NEW MICROPROCESSORS Intel Processor 4004 8008 8080 8086 8088 80286 386 DX 386 SX 486 DX 486 SX Pentium® Pentium® Pro Pentium® II Year Introduced Speed 11/15/71 4/1/72 4/1/74 6/8/78 6/1/79 2/1/82 6/17/85 6/16/88 4/10/89 4/22/91 3/22/93 11/1/95 5/7/97 108 KHz 200 KHz 2 MHz 10 MHz 8 MHz 12.5 MHz 33 MHz 33 MHz 50 MHz 33 MHz 66 MHz 200 MHz 300 MHz Bus Width 4 bits 8 bits 8 bits 16 bits 8 bits 16 bits 32 bits 16 bits 32 bits 32 bits 64 bits 64 bits 64 bits Number of Transistors 2,300 3,500 6,000 29,000 29,000 134,000 275,000 275,000 1.2 million 1.185 million 3.1 million 5.5 million 7.5 million Addressable Virtual Memory Memory 640 bytes 16 Kbytes 64 Kbytes 1 MB 1 MB 16 MB 4 GB 16 MB 4 GB 4 GB 4 GB 64 GB 64 GB — — — — — 1 GB 64 terabytes 64 terabytes 64 terabytes 64 terabytes 64 terabytes 64 terabytes 64 terabytes

Source: Intel Processor Hall of Fame, www.intel.com/intel/museum/25anniv/hof/hof main.htm.

PUBLIC POLICY ISSUES
Antitrust Policy
IBM Antitrust policy has played an important role in the computer industry since the late 1960s. On the last day of the Johnson Administration, the Justice Department charged IBM with attempting to monopolize the general-purpose electronic digital computer market.41 The Justice Department measured IBM’s market share at approximately 75 percent, based on the lease value of the installed base of electronic data processing (EDP) equipment. IBM contended that its market share was only 33 percent of the correct relevant market, which it defined as all data processing equipment, including computers, programmable hand-held calculators, message-switching equipment, and just about any other piece of equipment imaginable. The government charged IBM with a series of illegal actions. The complaint alleged that IBM bundled systems together by charging a single price for the central processor, related software, and maintenance. IBM was also accused of introducing new computer lines, particularly the System 360, in ways that were aimed at destroying the sales of machines recently introduced by competitors. In particular, the government complained about IBM’s practice of announcing a new line well in advance of its actual introduction. Furthermore, according to the government, IBM sometimes announced that a machine would be forthcoming even though no such machine existed or was planned.

548

Chapter 17 Using the Tools: A Case Study of the Computer Industry

On January 8, 1982 (on the same day it announced a consent decree had been reached to break up AT&T) the Justice Department announced that it was dismissing the case against IBM. William Baxter, the assistant attorney general in charge of antitrust, decided that IBM had not committed any major violations of the antitrust laws, and therefore, the case was “without merit.” Baxter also believed there was very little chance of a government victory, and given court decisions at the time, he was probably right. One of those decisions involved a private action taken by Telex against IBM. In the late 1960s Telex was one of many peripheral manufacturers offering IBM plug-compatible peripheral equipment at prices well below IBM’s prices for comparable products. As a result, IBM lost a significant share of the market for peripheral equipment. IBM responded with a series of aggressive actions and Telex sued.42 Telex claimed that as a result of IBM’s actions, Telex was forced to cut its prices drastically. Despite IBM’s price cuts, Telex was able to increase its volume of business between 1970 and 1972, but its profits declined substantially. The District Court ruled in favor of Telex on most issues, and awarded Telex $259.5 million in damages plus $12 million in attorney’s fees. The Tenth Circuit Court of Appeals, however, overturned the District Court’s decision, arguing that IBM’s actions were merely “normal methods of competition.” Microsoft The first government antitrust action against Microsoft was started by the Federal Trade Commission but concluded by the Department of Justice. This is highly unusual. The FTC staff argued that once Microsoft’s basic control over operating systems was established, it continued to maintain power through a variety of illegal practices including the pre-announcement of products, exclusionary perprocessor licenses for MS-DOS and Windows, unreasonably long-term licensing agreements, and restrictive nondisclosure agreements. The FTC staff presented its case to the full Federal Trade Commission and in February 1993 the Commission deadlocked 2–2 over whether to issue a preliminary injunction against several Microsoft practices.43 Six months later the FTC met again and deadlocked for a second time 2–2 on the Microsoft case. This marked the end of the FTC case, but in an unprecedented action the Department of Justice decided to pursue the FTC case against Microsoft. The Justice Department expanded the investigation and on July 15, 1994 proposed a consent decree.44 The case then took yet another bizarre twist—in February 1995 Federal Judge Sporkin rejected the Justice Department’s consent decree. Judge Sporkin feared that the decree would not protect the public from Microsoft’s monopoly power. The Justice Department and Microsoft filed a joint appeal to Judge Sporkin’s ruling. On appeal the Appeals Court ruled in favor of the Justice Department and Microsoft, and the consent decree was finalized. The major points in the consent decree were: 1. Microsoft agreed to stop offering large discounts for CPU licenses based on the total number of CPUs shipped instead of the number of copies of MS-DOS actually shipped.

Public Policy Issues 549

2. Microsoft agreed to end its use of long-term contracts that committed OEMs to purchasing large volumes of software in the future. 3. Microsoft agreed to end its policy of requiring nondisclosure by software developers. This ended Microsoft’s practice of requiring beta testers not to disclose details of Microsoft’s operating systems for three years after the systems came to market. The nondisclosure requirement had restricted the ability of programmers to move from one company to another, unless the programmer moved from another company to Microsoft. The consent decree did not deal directly with a much more important economic issue, the bundling of Windows with Microsoft’s Internet Explorer (IE). The Justice Department believed that the decree banned the tying of the IE browser to Windows and complained to the court that Microsoft was in contempt. On December 11, 1997, the District Court entered a preliminary injunction banning the tying of the Internet Explorer to Windows.45 Microsoft filed an appeal of this injunction, and on May 12, 1998, the Court of Appeals granted a stay of the injunction. Six days later, on May 18, 1998, the Justice Department filed a formal antitrust action charging Microsoft with attempting to monopolize the market in Internet browsers by tying the Internet Explorer to Windows. The trial received great national attention, especially when Bill Gates testified by video deposition. Much of his testimony consisted of the government’s attorneys showing Gates’s e-mail messages that he had either sent or received, with Gates indicating that he had no recollection of the messages. Among the more interesting evidence was a series of 150 e-mails provided by Netscape showing how Microsoft attempted to force users to use Microsoft’s IE instead of Netscape’s Navigator.46 For example, in a June 13, 1996 message, a Netscape employee reported that Microsoft had offered a Dutch Internet service provider free browsing software and a $400,000 marketing fund. “This was extended on the understanding that [the service provider] would NOT purchase any s/w [software] from Netscape.” In another message, an Internet service company tells a Netscape employee that “we have ceased distribution of Netscape. Your product is excellent but totally lacking in marketing support and we could never justify the $20 setup cost when Microsoft will fly a blimp with our name on it for free.” A June 10, 1996 message to Netscape from Kurt Brecheisen, president of a company called Global Telecom, said: “Microsoft gave me a deal I couldn’t refuse. Free dialer, browser, developer kit, freely distributable, etc. . . . I know Netscape is better, but $0 vs $18K is impossible to beat.” On November 5, 1999, District Judge Thomas Penfield Jackson issued his findings of fact in the case.47 Judge Jackson found that Microsoft was a monopolist in the market for Intel-compatible PC operating systems. Furthermore, Microsoft had used its monopoly power to restrict competition and harm consumers. Judge Jackson concluded:
Most harmful of all is the message that Microsoft’s actions have conveyed to every enterprise with the potential to innovate in the computer industry. Through its conduct toward Netscape, IBM, Compaq, Intel, and others, Microsoft has demonstrated that it will use its prodigious market power and immense profits to harm any firm that insists

550

Chapter 17 Using the Tools: A Case Study of the Computer Industry

on pursuing initiatives that could intensify competition against one of Microsoft’s core products. Microsoft’s past success in hurting such companies and stifling innovation deters investment in technologies and businesses that exhibit the potential to threaten Microsoft. The ultimate result is that some innovations that would truly benefit consumers never occur for the sole reason that they do not coincide with Microsoft’s selfinterest.

The Judge’s findings appeared to be the harshest possible indictment of Microsoft’s conduct. There was virtually nothing in Jackson’s 412 separate findings of fact from which Microsoft could take comfort. Consistent with his findings of fact, on April 3, 2000, Judge Jackson handed down his conclusions of law. Jackson concluded “that Microsoft maintained its monopoly power by anticompetitive means and attempted to monopolize the Web browser market, both in violation of §2 [of the Sherman Act]. Microsoft also violated §1 of the Sherman Act by unlawfully tying its Web browser to its operating system.” 48 On June 7, 2000, Jackson handed down his final judgment in the case.49 He ordered Microsoft to split itself into two separate companies: an operating system company and an applications company. Microsoft, of course, immediately appealed. When this text went to print, the Justice Department was appealing to the Supreme Court for the Supreme Court to hear the appeal directly without first having the case heard by the United States Court of Appeals for the District of Columbia. Microsoft, however, was urging the Supreme Court to send the case to the appeals court first. While the Supreme Court was deciding whether or not to hear the appeal directly, Judge Jackson delayed implementation of all of the restrictions he had placed on Microsoft’s behavior in his final judgment. By the time most of you read this book, more legal decisions will probably have been handed down in this precedent-setting case. Barring any huge last-minute surprises, the Supreme Court will eventually decide this case.

Copyright Protection for Software
A very important issue affecting the rate of technological advance in the computer industry is the attitude of the federal courts toward copyright protection in the software industry. Some issues are straightforward. For example, a firm cannot purchase a license for 100 copies of Microsoft Windows and then make 200 illegal copies to be placed on employees’ computers. Most issues, however, are complex. The copyright law with regard to computer software presents a dilemma for policymakers and the courts. If the courts take a very tough stand against the infringement of copyrighted software, they would further solidify the monopoly power of software manufacturers. However, such a tough position would create an increased incentive for other firms to invent around copyrighted materials. Alternatively, the courts might liberalize the definition of “fair use” with regard to computer software and reduce the software manufacturers’ short-run monopoly power. Such a policy, however, might create an environment that would be less conducive to technological change

Summary 551

in the future because there would be less incentive to invest in new product development. There have been several important cases in recent years, none watched with more interest than the legal battle between Lotus and Borland. Lotus 1-2-3 was the first commercially successful spreadsheet software for business accounting on PCs. When Borland developed a competing product, Quatro Pro, it admitted that it copied the basic menu commands from Lotus 1-2-3. In July 1992 the District Court ruled that Borland had infringed on Lotus’ copyrights.50 However, on March 9, 1995, the First Circuit Court of Appeals reversed.51 Borland’s defense was based on its belief that the menu structure of Lotus 1-23 was a “system” or “method of operation,” and as such was not protected by copyright law. The federal copyright laws, in fact, expressly state that “systems” and “methods of operation” cannot be protected. The Appeals Court ruled that the menu displays of Lotus 1-2-3 were a “method of operation,” comparing them to the buttons on a VCR, and therefore, the menu structure could not be legally copyrighted. This was the first time that a federal court had taken such a broad interpretation of the “fair use” doctrine with regard to computer software. Lotus appealed the ruling to the Supreme Court and the Court agreed to hear the case. However, on January 16, 1996, the Supreme Court, by a 4–4 tie vote (Justice Stevens not participating), let the Appeals Court’s ruling stand.52 The legal effect was to affirm the First Circuit’s decision and to greatly broaden the “fair use” doctrine in the software industry. The decision could have a dramatic impact on the future shape of the software market. Another important court battle took place between Apple and Microsoft over Microsoft’s introduction of Windows. Apple claimed that Windows infringed on Apple’s copyrights in 189 different ways. Microsoft’s defense was based primarily on a 1985 licensing agreement between Apple and Microsoft. In 1992 the District Court ruled in favor of Microsoft.53 The Judge found that of the 189 claimed infringement violations, 179 were clearly protected by the licensing agreement. Of the remaining 10 claimed violations, the Judge ruled that the ideas were not original to Apple, and therefore, could not be protected. Apple appealed this decision to the Appeals Court, but the Court affirmed the decision.54 Apple then appealed to the Supreme Court, which denied the appeal. If other federal courts adopt the First Circuit Court’s approach in the Lotus v. Borland case, it could create a much more open environment in the computer software industry.

SUMMARY
1. The performance of the computer industry will continue to be crucial to the growth of the United States economy. It will probably remain a market in which giant firms and small entrepreneurs compete against one another to be the first to develop major new technologies.

552

Chapter 17 Using the Tools: A Case Study of the Computer Industry

2. There are many similarities between the rise of IBM and the rise of Microsoft to power in the computer and operating systems markets respectively. 3. This industry has been characterized by conduct that includes a great deal of tying and bundling of products. 4. First-mover advantages appear to be extremely important in this technologically advanced industry. 5. Network externalities are extremely important in both the computer and operating systems markets. 6. IBM and Microsoft have responded aggressively to the entry of new firms and other competitive threats. 7. Performance in the computer hardware industry is now comparable to what would be expected in a competitive industry, but performance in the operating systems market is more problematic. 8. Antitrust policy has had a major impact in this industry and continues to play an important role.

KEY TERMS application programming interfaces (API) beta testing CPU licenses information technology industry Moore’s Law network externalities open architecture pre-announcement software lock-in

DISCUSSION QUESTIONS
1. Since the publishing of this book, what has happened in the Microsoft antitrust case? Has either the Court of Appeals or the Supreme Court made a ruling? Is so, who won the appeal? 2. Compare and contrast the factors that led to the creation of IBM’s market power with those that led to the creation of Microsoft’s market power. 3. How did luck enter into Microsoft’s rise to dominance? How did business acumen help the company gain dominance? 4. What were the major factors leading to IBM’s loss of market power in the PC market? 5. What do economists mean by the term network externalities? How have network externalities played a role in the computer industry? 6. Consider the first-mover model as presented in Chapter 12. How could you use this model to explain IBM’s, Microsoft’s and Intel’s market power? 7. Do you agree or disagree with Judge Jackson’s ruling in the Microsoft case?

References 553

REFERENCES
1. Much of the remainder of this chapter is derived from Don E. Waldman, “Computers,” in Walter Adams and James Brock, eds., The Structure of American Industry, 10th edition. (Englewood Cliffs, NJ: Prentice Hall, 2001), Chapter 6. 2. For more details of this history see Don E. Waldman, “IBM,” in David I. Rosenbaum, ed., Market Dominance: How Firms Gain, Hold, or Lose It and the Impact on Economic Performance (Westport, CT: Praeger, 1998), pp. 131–152. 3. For more on the life of Thomas Watson, Sr. see Thomas G. Belden and Marva R. Belden, The Lengthening Shadow: The Life of Thomas J. Watson (Boston: Little, Brown and Company, 1962); Emerson W. Pugh, Building IBM: Shaping an Industry and Its Technology (Cambridge, MA: MIT Press, 1995), pp. 29–36; and Robert Sobel, IBM vs. Japan: The Struggle for the Future (New York: Stein and Day, 1986), pp. 28–34. 4. Charles H. Ferguson and Charles R. Morris, Computer Wars (New York: Random House, 1993), pp. 3–4. 5. Ibid., p. 4. 6. Pugh, op cit., pp. 323–324. 7. Richard T. DeLamarter, Big Blue: IBM’s Use and Abuse of Power (New York: Dodd, Mead & Company, 1986), p. 29. 8. For more details of this history see Rochelle Ruffer and Don E. Waldman, “Microsoft,” in Rosenbaum, Market Dominance: How Firms Gain, Hold, or Lose It and the Impact on Economic Performance, pp. 153–173. 9. James Wallace and Jim Erickson, Hard Drive: Bill Gates and the Making of the Microsoft Empire (New York: John Wiley & Sons, 1992), p. 60. 10. Michael A. Cusumano and Richard W. Selby, Microsoft Secrets (New York: Free Press, 1995), p. 137. 11. The $75,000 figure is stated in both Cusumano and Selby, op cit., p. 137 and in Ferguson and Morris, op cit., pp. 66–67. Daniel Ichbiah and Susan L. Knepper, The Making of Microsoft: How Bill Gates and His Team Created the World’s Most Successful Software Company (Rocklin, CA: Prima, 1993), p. 76, state “the exact amount is not clear, but indications are that Microsoft paid less than $100,000 for QDOS.” 12. Ichbiah and Knepper, op cit., p. 93. 13. William G. Shepherd, The Economics of Industrial Organization (Upper Saddle River, NJ: Prentice-Hall, 1997), p. 182. Ferguson and Morris, Computer Wars, p. 8. Michael L. Katz and Carl Shapiro, “Network Externalities, Competition, and Compatibility,” American Economic Review 75(3) (1985): 424. For a more detailed discussion of this theory see Supra, Chapter 12, pp. 325–329. Gerald Brock, The United States Computer Industry: A Study of Market Power (Cambridge, MA: Ballinger, 1975), pp. 33–37. Telex Corporation v. International Business Machines 510 F.2d 894 (1975). See for examples, Telex Corporation v. International Business Machines 510 F.2d 894 (1975); ILC Peripherals Leasing Company v. International Business Machines 458 F.Supp. 423 (1978); and Memorex Corporation v. International Business Machines 636 F.2d 1188 (1980). “A Tyro Challenges IBM in Big Computers,” Business Week, (May 12, 1975): 65–67. “New Wave of Change Challenging IBM,” Business Week (May 29, 1978): 92–99. “I.B.M. Foresees Benefits from Price Cuts,” New York Times (April 26, 1977): 51; and “More Tumult for the Computer Industry,” Business Week (May 30, 1977): 58–66. Paul Carroll, Big Blues: The Unmaking of IBM (New York: Crown Trade Paperbacks, 1994), p. 18. Ferguson and Morris, op cit., p. 71. Ibid., p. 72. Ibid., p. 74. Ibid., p. 76. Ibid. Ibid., pp. 80–81. Carroll, op cit., p. 89. Kenneth C. Baseman, Frederick R. WarrenBoulton, and Glenn A. Woroch, “Microsoft Plays Hardball: The Use of Exclusionary Pricing and Technical Incompatibility to Maintain Monopoly Power in Markets for Operating System Software,” Antitrust Bulletin (Summer 1995): 272. Ibid., p. 272. Ruffer and Waldman, “Microsoft,” in David I. Rosenbaum, ed., op cit., p. 165. United States v. Microsoft, No. 94-1564 (D.D.C. filed July 15, 1995). “Digital Cuts Prices of PCS up to 21%,” The Arizona Republic, (July 18, 1997): E2.

14. 15.

16. 17.

18. 19.

20. 21. 22.

23.

24. 25. 26. 27. 28. 29. 30. 31.

32. 33. 34. 35.

554

Chapter 17 Using the Tools: A Case Study of the Computer Industry

36. “IBM Finally Introduces an under-$1,000 PC,” The Ottawa Citizen (November 7, 1997): F8. 37. Larry Kaggwa, “Inacom Official Says PC Prices Will Stabilize,” Omaha World-Herald (July 23, 1998). 38. Nicole Jacoby, “Buying a New Computer,” CNN Financial Network, cnnfn.com/1999/ 07/19/life/q buyingpc/. 39. “R&D Continues to Rebound in US—Forecast,” Newsbytes, www.newsbytes.com, January 4, 1999. 40. Intel Processor Hall of Fame, www.intel.com/ intel/museum/25anniv/hof/hof main.htm. 41. For two dramatically different views of the case see Richard T. DeLamarter, op cit.; and Franklin M. Fisher, John J. McGowan, and Joen E. Greenwood, Folded, Spindled, and Mutilated: Economic Analysis and U.S. v. IBM, (Cambridge, MA: MIT Press, 1983). 42. Telex Corporation v. International Business Machines 367 F.Supp. 258 (1973); and Telex Corporation v. International Business Machines 510 F.2d 894 (1975). 43. United States v. Microsoft Corporation No. 941564 (DDC filed July 15, 1994). Amended

44. 45. 46. 47.

48.

49.

50.

51. 52. 53. 54.

versions filed with the court on July 27, 1994. Ibid. See United States v. Microsoft Corporation No. 98-1232 (DDC filed May 18, 1998). “U.S. Releases E-Mail to Back Up Testimony,” Washington Post (October 24, 1998): D4. United States v. Microsoft Corporation, Civil Action No. 98-1232 November 5, 1999, District Court for the District of Columbia. United States v. Microsoft Corporation, Civil Action No. 98-1232 April 3, 2000, District Court for the District of Columbia. United States v. Microsoft Corporation, Civil Action No. 98-1232 June 7, 2000, District Court for the District of Columbia. Lotus v. Borland, 788 F.Supp. 78 (D. Mass. 1992); and Lotus v. Borland, 799 F.Supp. 203 (D. Mass. 1992). Lotus v. Borland, 49 F.3d 807 (1st Cir. 1995). Lotus v. Borland, 516 US 233 (1996). Apple v. Microsoft, 799 F.Supp. 1006 (N.D. Cal. 1992). Apple v. Microsoft, 35 F.3d 1435 (9th Cir. 1994).

INTERNET EXERCISES
Visit www.awl.com/waldman_jensen for this chapter’s web exercises. Answer our real-world economics problems by searching the Internet for solutions!