IEEE Micro 0272-1732/95/$04.00 © 1995 IEEE
Volume , 15 Number 2; April 1995, pp. 6-7
Microsoft and vaporware
Download PDF for this article
If you were Bill Gates and Philippe Kahn was about to introduce a new Turbo Basic package, what would you do to keep the Borland genie from getting out of the lamp? Those of you who have ever played Whack-A-Mole can guess.
Microsoft "sticks it to Philippe"
Three anonymous companies in Silicon Valley have protested entry of the Antitrust Division's settlement of the Microsoft case (for earlier story see Micro, Oct. 1994). The Gang of Three say that Microsoft chilled the market for Borland's Turbo Basic in 1988 by preannouncing Quick Basic 3. They call it vaporware; Microsoft calls it fair competition. The Justice Department says it doesn't matter what you call it, because the facts of this case don't show an antitrust violation. The focus of the Justice Department's antitrust case against Microsoft was per-processor licensing of MS-DOS and related licensing practices, which the settlement prohibits. The case and settlement did not address vaporware, because at the time the Department felt it did not have a winnable antitrust case against vaporware. Either Microsoft's use of vaporware did not happen at all, or the circumstances did not make it illegal. For example, Microsoft says that in the 1988 incident Microsoft in fact eventually brought out QB3 a week before Borland introduced its new Turbo Basic compiler. In addition, Microsoft denies that it has market power in the market for Basic compilers, so that whatever it did had no real effect on the general vigor of competition in the marketplace. In the early 1970s, in the aftermath of Watergate, Congress passed the Tunney Act. The Tunney Act permits government settlements of antitrust cases to become final only if a federal judge approves the settlement as being in the public interest. DC federal judge Stanley Sporkin, who used to be the Securities and Exchange Commission's chief enforcer, drew the assignment of reviewing the Microsoft settlement. Judge Sporkin was troubled by a Microsoft internal memorandum that said: "The best way to stick it to Philippe is preannounce ... to hold off Turbo buyers." Arguably, the memorandum shows Microsoft's specific intent to hinder Borland's marketing by a preannouncement. Should that be illegal? Reasoning from securities fraud principles, Judge Sporkin said yes.
The Microsoft hearing
But chief antitrust prosecutor Anne Bingaman thinks the memo is just so much hot air, thinks the anonymous Gang of Three protesters are just whiners or cry babies, and wants to wrap up the case and move on. Condensed here is some of the Sporkin-Bingaman dialogue at a recent court hearing (Jan. 20), which shows their different views of the legality of vaporware:
Sporkin: Why isn't this preannouncement document a smoking gun?
Bingaman: Your Honor, there is law on vaporware. Courts generally find that product preannouncements violate the antitrust laws, not just if they're false, but also knowingly false, intentionally false, deliberately false.
Number two, you have to prove they acquired market power as a result of it. You have to prove not just that they did it. We're not in an SEC case, Your Honor. We're in an antitrust case. You need proof of impact.
Sporkin: Well, how about if it's done for the purpose of holding off other people? I don't know what more you need.
Bingaman: You need market power.
Sporkin: You know, Ms. Bingaman, I don't know how much experience you have in this field, but I spent 20 years investigating white collar crime, and I must tell you I come up with a different conclusion than you come up with.
Bingaman: You need proof of impact. You have to tie that announcement to the market. You have to prove who didn't buy something because of that.
Sporkin: Aren't you concerned when the firm that is supposed to be the number one firm in this field puts out statements to freeze out others? You say that, because it doesn't have market impact, to go ahead and do it? Is that the message you want the Justice Department to give?
Bingaman: I enforce the law, and the law requires market impact. I'm real sorry, but it's not an SEC statute.
Sporkin: So, in other words, the message to Microsoft is to continue with this practice?
Bingaman: Your Honor, my statement to you is I'm the prosecutor. You're the judge. I decide what makes out a winning case, and if I don't want to file it, nobody can make me file it. That's the role we have. Your role is to accept or reject. Mine is to decide what to file. That's the reality.
Sporkin: My role is also to give you what's on my mind, so you can address it. So you can see what I see as a problem and try to deal with the problem. I do think you got a lot of good in this decree. But I do think that if they are engaging in practices that are improper, they ought to stop it.
I think I'm having more trouble with you than I might have with Microsoft.
Bingaman: Well, you know, what can I say? Lots of people do.
Judge Sporkin says that either a purpose to exclude or a harmful effect is enough to make a false preannouncement (vaporware) a monopolistic action, at least when the preannouncer is a dominant firm. Prosecutor Bingaman says that the vaporous preannouncement both must be knowingly false and also have really diverted customers, say from Borland Turbo to Microsoft Quick Basic. Judge Sporkin seems to have a point, but precedents are sparse and the law is murky. Moreover, what the parties involved meant by the term vaporware seems to have shifted somewhat during the hearing. Some times, the term was applied to an imaginary product that one company announced to discourage customers from buying an available or nearly available product of a competitor. Thus, Microsoft's counsel, Urowsky, at one point denied that QB3 was really vaporware: "But these are not true vaporware, products that don't exist when they're announced." The judge responded that Microsoft's detractors merely said that "you put out an announcement before the product is ready in order to freeze the people out there, so that they will not go out and buy the competing product that's already available on the market." Microsoft's Urowsky then tried to finesse the judge on the definition of vaporware and make him back off a little: "The Court isn't taking the position, I take it, that it's illegal to announce products in advance of shipment?" This made the judge do some backpedalling, because he was no longer certain what he meant was vaporware. (The judge's answer, as all Micro Law readers with 20-20 hindsight know, should have been that he was taking the position that it was or wasn't illegal, depending on the intent and surrounding circumstances. For example, a one-week advance announcement is probably never obnoxious. But a six-month advance announcement that may slip another six months, if an 800-pound gorilla makes the announcement with exclusionary intent, may very well be illegal.) Microsoft then explained that preannouncement of products gives the public more information, so that it can make a better-informed choice. A public that mistakenly thinks that Turbo Basic is the only new Basic that will be available for many months, while in fact a new Quick Basic will presently become available, is thus deprived of valuable information. Because Microsoft's preannouncement actually provided a benefit to the public, Microsoft should not be labeled an antitrust violator for alleged vaporspew. That is just fair competition. At this point, Judge Sporkin could no longer restrain himself, and broke in:
Sporkin: You're making it sound like it's benign. It isn't benign. This is done with an intent to defeat a competitor. It's done at a time when you don't have a product, prior to when you were going to normally announce it. It's done with an intent to defeat these people.
Urowsky: It's done with an intent to capture sales. That's what every competitor does.
Sporkin: You take the position that it's got to be totally false to be illegal?
Urowsky: Got to be false or misleading at the time the announcement is made.
Sporkin: Is this what you're doing all the time?
Urowsky: This is what competitors in this industry do all the time.
Sporkin: You would agree, though, that the intent here was to defeat the competitive Borland product?
Urowsky: That is inevitably the intent in any effort to capture sales from a competitor. Inevitably the intent. So if the intent to defeat sales to a competitor in order to capture those sales is a form of mens rea [evil mind], then all of our free market system is guilty.
Sporkin: I gather that your point is that this is a practice that you believe is proper and will continue to do?
Urowsky: Proper, truthful, necessary--
Sporkin: And will continue?
Urowsky: --continue to do. Necessary to consumers and very much in the public interest.
Sporkin: Can you...freeze somebody for a year?
Urowsky: If you had a product that was really developed and was in testing, or ready for testing, and you knew reasonably well what the product would look like even if you didn't know every feature or form of functionality, and you made a truthful disclosure to the market that you were developing it, this is what we think it's going to look like, and this is when we expect to be able to deliver it on our best view, the answer is, yes.
Previous vaporware cases
Only three reported decisions say anything of significance about vaporware. Two past cases involved IBM and Kodak preannouncements of new products, where slippage in delivery dates occurred. Both decisions came down for the defendants, but the plaintiffs failed to establish that the defendants knew they were not about to introduce the new products when they announced them. The courts were unwilling to stifle truthful or good-faith preannouncements, considering them legitimate and procompetitive. The Federal Trade Commission in 1975 settled a case against Xerox, however, with a prohibition of any product announcements more than three months in advance; there was no court decision. The remaining court decision on vaporware is troublesome. MCI won a jury verdict against AT&T, which had preannounced a telephone service plan by well over a year. The preannouncement had a significant adverse effect on MCI, and the jury apparently believed AT&T's purpose was to discourage AT&T customers from shifting to MCI. The court of appeals mentioned the IBM and Kodak cases, saying that they established that the legal test for preannouncements challenged under the antitrust laws is whether the incorrectly early announcement is knowingly false when made. The majority of the court of appeals then reversed the jury verdict, on the ground that the FCC, not AT&T, was directly responsible for the delay or most of it, so that AT&T could not be said to have known that its announcement was premature. A dissent unsuccessfully argued that AT&T knew perfectly well that the FCC would delay regulatory approval as it did (which the ordinary public did not know). The dissent also argued that the jury found that AT&T intended to chill MCI's market with the preannouncement, and that pointing a finger at the FCC (which directly caused all but two months of the delay) should not be an excuse for AT&T's deliberately announcing the service as if it were imminent rather than vaporware. Thus, the MCI-AT&T case announces the same kind of legal rule for vaporware that the earlier cases did, but it applies the rule to the facts of the case in a somewhat peculiar way. (Probably, most observers would agree with the dissent and call this preannouncement deliberately falsethat is, vaporware.) The Gang of Three brought up another preannouncement suit, but it wasn't really a vaporware case. AT&T considered establishing a digital data network (DDN), but decided not to do so. Datran then publicly proposed a DDN service and solicited investors. AT&T responded with a crash program to develop a DDN and preannounced it by two years. This discouraged investors from financing Datran and led to its demise. The government sued AT&T for monopolization, alleging that this episode was part of a continuing monopolistic program of AT&T to prevent other companies from competing with it. The court held that the total pattern of conduct would show monopolization unless AT&T rebutted the evidence with its own contrary proof. While the US v. AT&T case involves preannouncements, it is not clear that any deliberate falsity (vaporware) by AT&T was involvedalthough two years is rather longish for a preannouncement. Also, the DDN preannouncement was only one of a half dozen other exclusionary tactics that AT&T used. Thus, the court did not say that the Datran episode, by itself, showed monopolization. The more general antitrust rule stipulates several conditions that can lead to a dominant firm being guilty of monopolization. First, the dominant firm is a monopolist if it intentionally uses tactics "not honestly industrial" to preserve its monopoly position. Alternatively, it is guilty if it uses such tactics in an attempt to gain a monopoly and there's a dangerous probability of success that the firm will gain or maintain a monopoly as a result. Most would agree that use of vaporware can be evidence of specific intent to exclude rivals from the market, regardless of whether the tactic succeeded on the particular occasion in question. But intent to exclude or monopolize, without some amount of success, doesn't make a legal case.
Unresolved vaporware issues
This suggests a number of directions in which the analysis can branch out. First, how general should the general principle be that one applies? Broad rules like that just stated for what is monopolization may make more sense in some contexts and less in others. The rule about dangerous probability of success balances sets of considerations against one another. One consideration is what the court may do to the defendant; another is the social value of the challenged conduct. Private civil actions for monopolization usually involve a project to mulct the defendant for millions of dollars in treble damages. Moreover, most government cases seek a break-up of the industry structure that resulted from the monopolization, rather than just an injunction as in the government's Microsoft case. Even when just an injunction is involved, if the conduct to be forbidden has competitive benefits to the public, when not abused, it can chill the vigor of competition to suppress the conduct. For example, truthful preannouncements usually benefit the public. Only the deliberately false ones are totally lacking in redeeming social value. Building production capacity ahead of demand can be exclusionary but can also benefit the public, and the same can be said of vigorous price wars. Therefore, the dangerous probability rule is a filter or sieve that the antitrust laws use to try to separate anticompetitive and procompetitive conduct from one another when they may look very much alike at first glance. However, the premise of the rule is that something of value is at stake. If illegal vaporware is defined at the outset as intentionally false, the definition may filter out the socially useful conduct from the valueless conduct. Further, if the only remedy is going to be a court order not to do it again, rather than damages or a break-up order, perhaps it is excessive to expect proof of actual, specific impact as a result of the deliberately false vaporware announcement. Little is gained, and possibly much is lost, when such vaporware flourishes. It harms the public and also discourages potential market entrants from creating new software. Will small software houses even try to develop new applications if they think an 800-pound gorilla will vapor-zap the product before it gets a chance in the marketplace? Another direction in legal analysis may question the application of the concept of relevant market. Here, Microsoft may be presumed to have monopoly over PC operating system software. But the challenged conduct occurred in an application marketBasic compilers. Is Microsoft still an 800-pound gorilla once it engages in conduct in some other market where it presently has a relatively small market share? Microsoft argued, and the government appeared willing to concede, that Microsoft has no dominance in Basic compilers, tax preparation programs, or any similar, specific, relevant market other than operating system software. According to one school of economic thought, without market power in the relevant market that the conduct is said to affect, no harm can be done, irrespective of intentno harm, no foul. (This is directly contrary to what Judge Sporkin was saying in his dialog with Microsoft's Urowsky.) Another school would argue that if Microsoft intends to chill the market for Borland, you should take it at its word; it probably has a better idea of what is likely to happen than economic theoreticians do. After all, Microsoft went from 0% market share in word processors and spreadsheets to formidable, or perhaps dominant, positions in those application software submarkets. Market share in a relevant market is only a convenient numerical proxy or surrogate for measuring actual market power. Normal tests for estimating dangerous probability of success in monopolization may not apply to 800-pound gorillas with a track record of prior success. Besides, are Basic compilers, tax preparation programs, and operating system software each in separate compartments, or is all software intertwined and mutually impacting? The anonymous Gang of Three argued that one cannot meaningfully examine one part of the software market as separate from the others. Moreover, applications have to run under and interact favorably with Windows. Still another approach to vaporware might be to forget about the antitrust laws and turn to unfair competition laws, which forbid unfair and deceptive acts and practices, if they harm consumers or competitors. Is vaporware conduct of the kind involved in the "stick it to Philippe" memorandum an unfair act and practice? Does it amount to a false and misleading description of the properties of Microsoft's product? If so, it is an unfair method of competition or unfair or deceptive trade practice, actionable under laws prohibiting such conduct. On the other hand, does fencing in product preannouncements too closely, under such laws, cause a risk that legitimate information distribution will be chilled? Mistakes are made, and one can never be sure that a delivery date will not slip. Would upholding vaporware suit verdicts against product preannouncements cause the public to be worse off? Would that deter entrepreneurs from providing the public with information on future products, lest a litigious school of piranhas descend on them? Or is that just a figment of the over-active imagination of advocates?
Reader Interest Survey
Indicate your interest in this department by circling the appropriate number on the Reader Service Card. Low 170 Medium 171 High 172
Since this article on the acrimonious January 20 hearing was written, Judge Sporkin issued a 45-page opinion on February 14, blasting the parties. He ruled that it would not be in the public interest for him to approve the proposed settlement. The Justice Department and Microsoft then announced that they each planned to appeal Judge Sporkin's decision and request expedited review from the US Court of Appeals in Washington, D.C. Judge Sporkin said his problem was lack of information. "Basically, other than being told that the government spent a great deal of time on a wide-ranging inquiry and that the defendant is a tough bargainer, the Court has not been provided with the essential information it needs to make its public interest finding." He objected to being "stonewalled" and asserted that he and his Court "are not mushrooms to be placed in a dark corner and sprinkled with fertilizer." Judge Sporkin said he was particularly concerned with vaporware accusations. He felt that Microsoft's dominant position in operating system software (MS-DOS, Windows) was so strong that it would not likely be dislodged, an opinion that the government's economist expert had also offered. "Given this fact," Sporkin said, "Microsoft could unfairly hold onto this position with aggressive preannouncements of new [upwardly compatible operating system software] products in the face of the introduction of possibly superior competing [operating system software] products." This could work because consumers and OEMs would be unwilling to switch operating system software at the risk of not having applications to run under the new operating system. (Presumably, these OEM and consumer fears would interact and reinforce one another.) Further, Judge Sporkin disagreed with, or at least questioned, the government's and Microsoft's antitrust law standard for what is illegal vaporware. That standard requires both known falsity of the preannouncement (that is, deliberate misrepresentation of an early shipping date) and also that the preannouncement contributes to acquisition or maintenance of a monopolistic market share. He termed that a "rather narrow view of a highly questionable practice" and use of criminal law standards where civil standards are appropriate. Regardless of what is the right standard for determining the legality of vaporware, Sporkin felt the Court should be informed about the surrounding facts. He also wanted to know whether there is any legal basis for prohibiting vaporware tactics at a lesser degree of illegality, as a means of prying open Microsoft's existing operating system monopoly. "Given the government's stated desire to open up competition," he pleaded, "why does it not want to take on the vaporware issue?" Microsoft admits it intended to defeat competitors' marketing plans, and says it will go on vaporizing in the future, because it is perfectly legal. (Here, Judge Sporkin referred to the episode involving Borland's Turbo Basic and also to another anti-Borland incident involving Microsoft's preannouncement of Quick C. Neither of these incidents, however, involved the operating system software over which Microsoft has a monopoly.) Sporkin said he cannot sign off on a decree allowing Microsoft to continue engaging in vaporization, with Microsoft saying it plans to continue engaging in the conduct, "without the government providing a full explanation as to [why it takes] its 'no action' stance." It would be like the Court giving probation to defendants who tell the court they plan to go out and engage in the same conduct again once they are out on the street. Sporkin painted Microsoft as a company that has a monopoly in a field central to US well being, not only for the balance of this century but also for the next. Microsoft is so feared by its competitors (referring to the anonymous Gang of Three) that they are afraid to disclose their identity when they inform on it. "The picture that emerges from these proceedings," according to Sporkin, is that Microsoft is so powerful "that the US government is either incapable or unwilling to deal effectively with a threat to this nation's economic well being." He said that the government "is so anxious to close this deal" that it came up with such a narrow interpretation of what is illegal vaporware that it gives Microsoft "a green light to engage in anticompetitive practices with impunity." Yet, "'vaporware' is a practice that is deceitful on its face and everybody in the business community knows it." Why else do they call it "vaporware"? he asked.
Send general comments and questions about the IEEE Computer Society´s Web site to firstname.lastname@example.org.
This site and all contents (unless otherwise noted) are Copyright ©2000, Institute of Electrical and Electronics Engineers, Inc. All rights reserved.