A Guide to the Microsoft Case Outakes

Navigator v. 2, #1, September 1998.
"A Guide to the Microsoft Case", an exclusive interview with Robert A. Levy of the Cato Institute. The following are outtakes from the published interview.

• The 1994 Consent Decree
• Tying arrangements
• Microsoft's potential competition
• The Chicago School and the Austrian School
• Competition and protecting competitors

Levy: In 1990, the Federal Trade Commission opened up their investigation of Microsoft, and it continued until early 1993. The FTC deadlocked 2 to 2—I think one member recused himself on whether to bring an action under the anti-trust laws against Microsoft. They took another vote later that year, in July, and deadlocked again. So they closed their probe, but turned the files over to the Justice Department and the Justice Department opened its own anti-trust investigation, in August of 1993.

Navigator: That seems almost like double jeopardy. You've been investigated by one antitrust agency for three years, and it takes no action, so the other picks it up, and they have another shot at you.

Levy: Yes, I think that's correct. One would expect that one agency having decided, the other agency would defer. But because the FTC deadlocked 2-2, it meant that there was no clearcut statement that the antitrust laws were inapplicable to the Microsoft situation. So an ambitious Justice Department took over.

Navigator: What happened with Antitrust Division?

Levy: In 1994, Microsoft reached a settlement with the Justice Department, and agreed to change some of its contracts with computer makers, and eliminate some restrictions in those contracts. As a result, a Consent Decree was negotiated and signed in 1994. Then in February of 1995 Judge Sporkin of the U.S. District Court, Washington DC, rejected that settlement as not being in the public interest. But in June of 1995, the Appeals Court in Washington overturned Judge Sporkin's decision, and indeed castigated Sporkin for relying upon evidence that wasn't in the record. Apparently, he had read a book on the beach that summer, and that was the basis for his decision. The Appeals Court directed that a new judge—Judge Thomas Penfield Jackson—be assigned. And he's still the judge on the case. He approved the settlement, the Consent Decree, in August of 1995. One of the things that the Justice Department objected to at that time was the so-called per-processor fee. The way it worked was that Microsoft offered a discount—let's say it was 40 percent—if a manufacturer would license Microsoft's operating system on a per-processor basis instead of a per-copy basis. That meant on every processor the OEM sold, they would pay Microsoft a fee, even if Windows were not installed. Looking at that from the OEM point of view: If Windows 95 was to be installed in more than 60 percent of the processors, and you're getting a 40 percent discount, then of course per-processor is cheaper than per copy. The decision was entirely up to the manufacturer and virtually every manufacturer selected this per-processor option.

Navigator: Was the Justice Department arguing that this arrangement meant the OEMs could not go out and buy a different operating system?

Levy: Yes, it meant that the OEMs were effectively put in the position of paying for something whether they used it or not, and so therefore they would use it. DOJ apparently gave little or no weight to Microsoft's contention, first, that it was up to the OEMs to decide and, second, that they would choose the per-processor option only if indeed the percentage of the computers in which they intended to install Windows was greater than the complement of the discount they were getting.
 

Outtake 2: Tying Arrangements.

For a fuller treatment, see Levy's "Microsoft and the Browser Wars," published by the Cato Institute and available at http://www.cato.org/pubs/pas/pa-296es.html

Navigator: What has been the history of tie-ins and the history of using anti-trust laws?

Levy: The first important case came a few years after the Clayton Act was passed in 1914: The United Shoe Machinery case. That involved a company, United Shoe, that controlled 95 percent of the shoe machinery market. The company refused to lease its equipment unless the lessee also bought supplies. So if somebody wanted to enter the supplies market, the competitor would either have to produce its own machinery and go up against United Shoe, which already had 95 percent of the market, or be content with scrambling for 5 percent of the customer base. The Supreme court concluded that United Shoe's arrangement was an unlawful tie-in. Over the next thirty years or so, the courts applied a "rule of reason" test—that is, the courts would weigh the facts peculiar to each situation presented to it. They would look at the business condition before and after the restraint, the nature of the restraint, and what it's likely effect was going to be. It wasn't until 1947, in the International Salt case, that the court set forth a brighter-line rule that hinged on the three-part per se illegality test: two distinct products, a monopoly in the tying market, and a "not insubstantial" volume of commerce in the tied market. Of course, if the case did not meet the test of per se illegality, that meant only that the court would fall back to the "rule-of-reason" test and apply that.

Yet even with the brighter-line rule, the court pronouncements have been inconsistent over time; there's still a great deal of uncertainty, as we see in the Microsoft case. For example, the Supreme Court held that a newspaper that required advertisers to purchase space simultaneously in the morning and evening editions was okay, because the dual editions were actually one product. On the other hand, one of the circuits, the Fifth Circuit, decided that you couldn't tie automobiles and air conditioning units. And the Ninth Circuit said you couldn't tie a computer and an operating system. So there's been an inconsistent application of the tying doctrine.
 

Outtake 3: Microsoft's potential competition

Navigator: A pro-capitalist should never say how the market would deal with a dominant firm if left to itself. But there is no harm in speculating how it might do so. And in "The Browser Wars," you speculate about some potential sorts of competition that Microsoft's operating system might face. What are they?

Levy: One is network computing technology. Software applications will be downloaded from the Internet. And that'll convert the PC from what it now is, basically a processor, into nothing more than an inexpensive communications device. In addition, we have digital TVs, set-top boxes, a whole variety of consumer electronic devices, that are going to radically diminish the importance of the operating system as we now know it, changing its very nature and scope and function.

When it comes to applying the antitrust laws, and deciding whether or not a company has a monopoly in the economic sense, you have to look at more than just existing competition. After all, the company could have 100 percent of the market simply because it is the best product available. But if there's the threat of other competitors coming into the market, with new products or similar products provided more efficiently, then that threat disciplines the would-be monopolist's behavior. I think that's the case here.

The Washington Post had an interesting article some time ago that made this point, and here's a quotation from that article: "The high-tech world is on the brink of changes just as profound as when the PC emerged. The era of a single dominant device is fading. We have a smorgasbord of new devices, from hand-held machines to digital TVs to network computers. Industrial giants Intel and Microsoft appear headed in different directions." (Elizabeth Corcoran, "Is Intel's Partnership with Microsoft Waning?" Washington Post, December 4, 1997, p. E1.) So there is countervailing power held by huge companies going at each other, like IBM and Intel and Microsoft and Sun and Oracle and Novell, and that makes short shrift of any would-be monopolist.

Corel, by the way, came out with a new product using the Linux operating system (that's a free operating system) and the Netscape browser, which is also free. They're providing the package, written in Java, for less than a hundred dollars, to anybody that buys their J-suite product for network computers. International Data Corp., the consulting firm, has described the Corel package as the single most interesting threat to Microsoft in a long time. Clearly Corel is not deterred by Microsoft's alleged predation and tying and network effects. In the software business, potential competition soon becomes existing competition, even against a company as firmly entrenched as Microsoft.

Potential competitors will take advantage of any weakness on Microsoft's part or any delay on Microsoft's part in adapting to emerging new technology.
 

Outtake 4: The Chicago School and the Austrian School

Levy: Let me make one other point, and that is the difference between the Chicago approach to efficiency, and the Austrian approach, because I do think that those two approaches are quite distinct. The economic efficiency approach of Chicago involves a cost-benefit calculus. If it can be shown that the benefits from intervention exceed the costs, thereby increasing aggregate welfare, then intervention is justified. By contrast, the Austrian approach says: Whenever you impede the market process, you cut off the flow of information that decision makers need in order to determine what to produce, where to produce, at what price, what features, what services, etc. The market itself is a discovery process, and within the market price is impounded all of the needs and desires and aspirations, both economically-based and non-economically based, of market transactors. If that discovery mechanism is shut down, decision makers will not be able to adjust production efficiently to meet the demands of consumers. That insight is critically important. Even if Microsoft has a monopoly, even if it is behaving anti-competitively in a manner that harms consumers and diminishes consumer welfare, government intervention is justified only if we're better off with political power under Bill Clinton than market power under Bill Gates. Because the very nature of intervention is to prevent markets from conveying information to producers, government would not be able to determine what actions to take in order to remedy perceived inefficiencies. There can be no doubt, for example, that DOJ does not understand the long-run implications of shutting off incentives for product development and improvement.

Outtake 5: Competition and Protecting Competitors

Navigator: The courts seem to have said, rather frequently, that the purpose of antitrust is not to protect competitors but to protect competition. And yet, when they find a company like Microsoft acting as a really fierce, drive-'em-out-of-business competitor, that seems to be seen as illegal. Is the antitrust law construed to mean that you can compete, but only if you keep your eyes on benefiting the consumer and not on driving your rivals to the wall?

Levy: Much is made of documents that are being uncovered, written by Microsoft executives. Imagine if you would issuing a memo to your staff saying, "We want to do x, y and z because it enhances consumer welfare and increases economic efficiency." I mean—having been in the position of running a software company—that's not the way things happen in the real world. Your memo's going to say, "We want to do this because it's going to increase sales and increase profits and put the guy across the street, who's producing the same product, out of business." That's what competition is all about. Indeed, that's what competition ought to be.

When it comes to software in particular, the risk of private monopoly is far less than the risk of regulatory excess, because software is different from oil and utilities and railroads and other capital intensive ventures. Software is not based on physical equipment but ideas—on human intellect, the ultimate resource. No company has a monopoly on ideas. The history of the software market is that better ideas have brought better products, and better products have won in the marketplace. When government gets involved, we politicize competition, enlisting the public sector in pursuit of the private interests of competitors.

If you read the complaint by the Justice Department plus the accompanying memorandum, it's 130 pages. And Netscape is mentioned 130 times: one time per page on average. So DOJ's lawsuit is not in service of consumers but in service of Netscape. And the computer market is the last place we need this kind of intervention. We don't need it anywhere, but particularly not in the computer market, which has accounted for 30 percent of GDP growth since 1994. It's growing at seven times the rate of the economy, it's creating millions of jobs, billions of dollars worth of GDP, and the cost of computing is declining precipitously.

What then is the consumer justification for this antitrust suit? The result is that Microsoft and Netscape and Sun—instead of devoting their resources to production and marketing, which is what they ought to be doing—spend their time currying political favor in Washington DC. Now, they've hired a half-dozen former Congressman and staffers and former antitrust officials and aides to presidents and the best law firms and lobbying firms and PR firms in DC. That's what DOJ has provoked: a diversion of our nation's most valued resources from business to politics.