by David Kopel
In its battle with Bill Clinton's trustbusters, Microsoft learned some frightening lessons about how business really works in America. The government and Microsoft both gave up a little in the Nov. 2 settlement, but the real losers were computer users and justice in America.
Liberty, Jan. 2002, pp. 21-24. This online edition corrects some errors in the print edition.]
Last fall, Microsoft and the United States Department of Justice announced a proposed settlement of the antitrust case against Microsoft. While the settlement amounts to a substantial victory for Microsoft, the long-term result of the case is harmful for Microsoft, consumers, the economy, and freedom.
Certainly Microsoft came away much better off than the mainstream media believed possible at most stages of the case. Federal trial judge Thomas Penfield Jackson's order to break up Microsoft was voided by the District of Columbia Court of Appeals in the summer of 2001. The new settlement does not include many of the terms ardently sought by the Department of Justice subsequent to the Court of Appeals ruling such as restricting Microsoft's long-standing practice of giving away the Windows Media Player as part of the Windows operating system. Nor does the settlement include the most egregious terms which had been sought by the Department of Justice in the 1997 case which was the forerunner to the antitrust case -- such as a prior restraint on any changes in the Windows operating system. And Microsoft did not agree to what the Department of Justice had originally demanded as the price for not bringing an antitrust case against Microsoft: a requirement that Microsoft distribute the Netscape browser with every copy of Windows.
Yet Microsoft's troubles are not necessarily over. While the Department of Justice and Microsoft worked out a carefully-structured settlement, nine of the eighteen states which the Department of Justice had convinced to join the original Microsoft antitrust case refused to settle. Federal District Judge Colleen Kollar-Kotelly opened a three-month comment period for public comment on the proposed settlement. A court hearing is scheduled for March 4. Whether the nine recalcitrant states will be able to undo the settlement, or to carry on with their own cases under state business laws, is uncertain.
However much trouble the states cause Microsoft in 2002, Microsoft does owe the states a large debt of gratitude. In the spring of 2000, Microsoft and the Department of Justice had been led by mediator Richard Posner (a highly respected federal appellate judge) to a proposed settlement that was far more severe than the current proposed settlement. Microsoft and the Department of Justice were ready to sign; but the states refused, and demanded much, much more. The deal fell apart. Thus, state participation in the federal antitrust case deprived the Department of Justice of what would have been a major DOJ victory, a victory which, as a voluntary settlement, never could have been undone by a federal court of appeals. Now, the DOJ's strategy is again being thwarted by unrealistic state demands. Perhaps the Microsoft case will be a caution for future DOJ litigators to stick to bringing their own cases in federal court, rather than assembling a peanut gallery of state attorneys general with parasitic claims.
Moreover, even if all the state claims were eventually dismissed, the DOJ settlement leaves intact trial judge Jackson's findings of fact including the finding that Microsoft is a monopoly. These findings are plainly wrong, since they define the "relevant market" so as to exclude Apple, since they falsely conclude that Linux is not a serious competitor to Windows, and since they are premised on indisputably incorrect data about Windows' share of the market (as Bob Levy of the Cato Institute has detailed).
Even so, a "finding" is a finding. The federal court of appeals having affirmed the trial court's finding that Microsoft is a monopoly, scores of plaintiffs attorneys will be able to bring private antitrust suits against Microsoft, relying on the conclusive findings from the government's case. These cases will natter at Microsoft for years, although the plaintiffs will have a very difficult time showing that consumers suffered as a result of anything Microsoft did.
Regarding the terms of the settlement itself, some terms amount to the government intervening to settle ordinary commercial disputes. For example, one issue on which Microsoft decisively lost at trial and on appeal related to the initial start up screens--the screen that the consumer would see the very first time he turned on a new computer. Microsoft insisted that computer manufacturers ("OEMs" Original Equipment Manufacturers) not replace the Windows startup screen with a customized screen. OEMs could add as many icons as they wanted, but OEMs could not remove the icons that Microsoft included.
As a matter of copyright law, Microsoft was plainly within its rights to insist that its software display not be altered. The trial and appellate courts, however, found Microsoft's copyright irrelevant.
After the initial boot, changing the start-up screen once and for all takes only a few mouse clicks. OEMs could always include icons (along with supporting paperwork), to tell a user to "Click here to start your Gateway experience." From there, the OEM could customize at will, steering the user into the OEM's preferred Internet Service Provider, or wherever the OEM wanted to steer.
If steering from initial boot (rather than from the first screen after the initial boot) were really all that important, OEMs could have paid Microsoft a few dollars extra for each copy of Windows, and bought steering rights. Nothing prevented the OEMs and Microsoft from coming to mutually satisfactory terms. The effect of Department of Justice intervention, however, was to give the OEMs some of Microsoft's property rights, without the OEMs having to pay for it.
In a society in which we believe that people are capable of calling an 800 number in order to switch telephone companies, it also seems possible that many of them are capable of following menu instructions to change an opening screen.
Another provision requires Microsoft to charge OEMs uniform rates, with a published discount schedule. This will significantly affect Microsoft practices. For example, IBM helped Microsoft develop Windows 3.1, and accordingly received a special discount from Microsoft for purchases of Windows 3.1. Compaq helped develop Windows 95, and received a special price for that product. These discounts were alleged by the government to be illegal antitrust violations, because they gave preferential discounts. Actually, it was a sound move for Microsoft, when developing a new operating system, to find a major computer manufacturer who could help design and debug the new operating system. This made the operating system more robust, less buggy, and better-tested for every consumer. A special discount is a reasonable reward for a company that does special work. At the antitrust trial, IBM complained that it didn't get the same deal on Windows 95 that Compaq did--even though IBM did nothing help with Windows 95, and only belatedly decided to license Windows 95 when IBM discovered that IBM's own OS/2 wasn't very popular.
But under the terms of the settlement, Microsoft is forbidden to offer special negotiated discounts to companies that help develop better products. The inevitable result will be more bugs, and software that has less testing, and is inferior to what would be produced in a world without the anti-competitive mandate of antitrust.
The real harm of the Microsoft antitrust case, however, has little to do with the terms of the settlement. Rather, the case's greatest destructive effects are how the case has transformed the computer business from one that was proudly independent of the government into a business which participates in the same illicit relationship with government as does most of the rest of American big business. To understand this problem, let us look at antitrust law and the Microsoft case in political terms.
The principal-agent problem has long been recognized as one of the keys to understanding government intervention in the economy. The government is supposed to act as the agent for its master (the people). But the government's interest may not always be the same as that of the people. So when the government claims to act as the agent of consumers, the government may in fact be promoting its own interest (more government power) rather than consumer interests (better products at lower prices).
As economist Bruce M. Benson and other authors have explained, antitrust is just as subject to the public choice problem as any other form of regulation. That antitrust is enforced through post-hoc lawsuits, rather than anticipatory rule-making, does not alter the public choice incentives.
Indeed, antitrust is well-structured to enjoy insulation from the normal political processes that combat government economic favoritism. Most forms of regulation consistently harm some companies while benefiting others; antitrust, though, targets a shifting set of victims--sometimes large companies, sometimes small ones. Some industries may be left alone for decades, while others are hounded incessantly. And unlike with regulations such as federal price-setting for air or bus travel, the harm to consumers is indirect, and not felt first-hand. Further, antitrust enjoys intellectual respectability among some scholars who are ordinarily critics of government control of the economy.
Milton Friedman used to be one of those scholars, because from a standpoint of theoretical economics, a proper use of antitrust could benefit consumers--such as by deterring conspiracies to fix prices. Yet in 1998 Friedman noted that many of his fellow economists have concluded that antitrust laws "tend to become prey to the special interests. Right now, who is promoting the Microsoft case? It is their competitors, Sun Microsystems and Netscape." In the abstract world of economic theory, "Monopoly is a problem," Friedman acknowledged, and for that reason antitrust used to enjoy universal support among economists. But in real life, monopoly "tends to be transitory, to be very short-lived in most cases. The only ways in which monopoly can last is when it has government backing." In 1999, Friedman elaborated: "as I watched what actually happened, I saw that, instead of promoting competition, antitrust laws tended to do exactly the opposite, because they tended, like so many governmental activities, to be taken over by people they were supposed to regulate and control. And so, over time, I have gradually come to the conclusion that antitrust laws do far more harm than good, and that we would be better off if we didn't have them at all, if we could get rid of them."
In The Wealth of Nations, Adam Smith argued that "Consumption is the sole end and purpose of all production, and the interest of the producer ought to be attended to, only so far as it may be necessary for promoting that of the consumer." The Microsoft case was accompanied by the usual pro-consumer rhetoric despite the government's failure at trial to prove that consumers had been harmed. Indeed, the government never put on a single witness who described consumer harms. To the contrary, Microsoft's market successes in operating systems, spreadsheets, word processing, Internet browsers, and other software products were the result of dramatic price reductions and quality improvements. Consumers benefited, but competitors suffered greatly. It was these suffering competitors who were the origin of the Department of Justice case against Microsoft.
Microsoft was a better market entrepreneur than anyone else, but the company failed to realize that its competitors were political entrepreneurs. The Federal Trade Commission (in the early 1990s) and the Department of Justice (starting in 1993) did not develop an interest in Microsoft by reading computer industry news magazines. Rather, they began to notice Microsoft only after a long, persistent, and well-financed lobbying campaign by various Microsoft competitors. The acronym for this group is NOISE (Netscape, Oracle, IBM, Sun, Everyone else).
Among the most prominent companies in "everyone else," was Novell, a Utah-based company which has suffered doubly at the hands of Microsoft. Novell's small-office networking business has been eroded by the small office networking capabilities built into Windows 95, and improved in Windows 98. Novell also bought WordPerfect when it was still the leading word processor, and sold it a few years later for a loss of hundreds of millions dollars, as WordPerfect was supplanted by Word, due to Novell's miserable product management. Utah is also the home state of Senator Orrin Hatch, chair of the Senate Judiciary Committee in 1995-2001. His widely-publicized anti-Microsoft hearings helped lay the political foundation for the DOJ antitrust prosecution.
AOL was also an important member of NOISE, even before AOL acquired Netscape. While Netscape did sincerely care about Internet Explorer--as a threat to Netscape head Jim Barksdale's self-proclaimed "God-given right to a 90% market share"--the rest of the NOISE coalition did not. They had about as much genuine concern about the Internet browsers as Lyndon Johnson did for whether the North Vietnamese had actually been the aggressors in the Tonkin Gulf Incident. The alleged aggression was simply a pretext for war with a long-hated enemy.
What about the rest of NOISE? IBM sells to businesses and consumers, while Oracle and Sun sell almost exclusively to businesses. For all three companies, the model is high price and low volume the opposite of the Microsoft model. Like Microsoft, Sun integrates a web browser into its Sun Solaris operating system. Until 1993, if you wanted to buy the Sun Solaris operating system, you also had to buy microprocessors, storage, system software, and middleware from Sun, or from a small number of Sun-licensed affiliates. Sun makes two types of workstations, and one type works only with Sun servers. IBM also offered its customers limited choices in the past, although competition resulted in has looser practices.
Nobody challenged the model of making business customers buy everything from a single source. That is, until Microsoft entered the market in the early 1990s. Microsoft's Windows NT operating system for servers is sold as a stand-alone product, and works on many different kinds of computers. The NT software is also simpler to use, has a well-designed easy graphic interface, and is cheaper than the products from the dominant companies. Computer hardware to run Windows NT machines is made by many different manufacturers, and is significantly cheaper than the proprietary Solaris machines made by Sun. With Windows NT as a platform, low-cost hardware companies like Dell and Gateway (which at the time knew a lot about Windows, but nothing about Unix) could start taking sales away from more expensive machines.
Suddenly, the NOISE companies discovered that Microsoft Windows 3.1, and then Windows 95, and then Windows 98 (all made for the desktop market, not the server market) were infected with all sorts of antitrust violations. Incidentally, these consumer products also happened to be the cash cows that gave Microsoft the resources to get into the server software business. These consumer products also used a popular graphical interface that made many office workers eager for their companies to adopt Windows NT--since the worker could use an interface like the one that was already familiar from his home computer.
Did the NOISE companies really believe their fantastic warnings about Microsoft's "chokehold" on the Internet? Was Microsoft's style of competition really different from that of the NOISE companies? In fact, every weapon Microsoft used in the browser war (e.g., giving the product away, special deals with favored websites) was also used by Netscape.
Microsoft's Internet Explorer the original target of the Department of Justice's antitrust suit -- is descended from the Spyglass browser, for which Microsoft bought a license from its creator Doug Colbeth. But when Microsoft started giving the browser away for free (to beat Netscape), Colbeth's licensing agreement became worthless. The Department of Justice interviewed Colbeth as a potential witness in the antitrust case. But Colbeth refused, explaining that Microsoft simply behaves like every other Internet company, and that the CEO's complaining about Microsoft are hypocrites: Netscape's "Barksdale and [Sun's] McNealy, they've all done what Gates has done. There is nothing there I wouldn't have done."
It is strange that Americans are being lectured on business ethics by Larry Ellison of Oracle. Ellison hired a detective company (previously known for digging up dirt on women who accused Bill Clinton of sexual misconduct) which attempted to steal trash from the offices of think tanks which defended Microsoft. Ellison defended the theft and spying on the grounds that it was a "public service." If only Richard Nixon had been creative enough to claim that the Watergate burglary showed his commitment to open records and public disclosure. More recently Ellison tried to convince the Bush administration to create national ID card which would run on Oracle's database software.
Not just Oracle, but the entire NOISE coalition has lobbied, successfully, to pervert the law, and to deploy the coercive force of government against a company which has done nothing wrong except compete aggressively and cut NOISE's profit margins. To add insult to injury, the NOISE propagandists self-righteously proclaimed their devotion to "competition." It is as if Torquemada began each Inquisition by announcing his firm support for religious liberty and tolerance.
By bringing the antitrust case, the Department of Justice demonstrated to every computer maker in America that NOISE political tactics were a good way to get the government to harm one's competitors.
Sadly, Microsoft has also learned political lessons. Microsoft used to be a company which was proud to stay out of politics. Even after an FTC investigation in the early 1990s (over Microsoft's agreement with IBM to work together on the development of the next iteration of Windows and on OS/2), after a Department of Justice investigation culminating in a 1995 consent decree, and after a 1997-98 lawsuit over the consent decree (in which Microsoft's interpretation was vindicated), Microsoft's financial and lobbying involvement in Washington was puny. In 1995, Bill Gates was naοve enough to declare that political issues are not "on our radar screen." As of 1994, the company had one lobbyist in Washington. Even in late 1997, Microsoft "had zero presence on the Hill," according to Republican Rep. David McIntosh.
Incredibly, Microsoft's political non-involvement was dubbed "arrogant" by the Washington, D.C., establishment as if the D.C. political class were an organized crime syndicate to which every large company should be expected to pay protection.
The May 1998 antitrust lawsuit served as a Pearl Harbor for the company, which finally began to spend as much money on lobbying and campaign contributions as do similarly-sized companies which are under heavy political attack. In 1995, the Microsoft PAC spent only $16,000 in 1995 (on copyright and encryption issues), but now Microsoft is one of the largest corporate political donors in the U.S. Microsoft has bought itself a major lobbying presence in Washington, and begun throwing soft money at the two major parties, and hard money at various candidates.
Microsoft's belated self-defense spurred Microsoft's competitors (who got into the influence-buying game years earlier) to put out their paid consultant Robert Bork to fret that "There is so much Microsoft money flowing through the system that the danger for nonpoliticized law is very real." This was like Hirohito complaining that American submarines were threatening to militarize the Pacific Ocean.
Microsoft's enemies managed to gin up a mini-scandal in early 2000, over the fact that Microsoft and its allies were lobbying against a record budget increase for the DOJ Antitrust Division. If Microsoft's lobbying was morally wrong, then it is wrong for victims of abusive IRS enforcement to lobby against a higher IRS budget, for gun stores victimized by the BATF to lobby against BATF budget increases, and for Branch Davidian survivors to lobby against more money for the FBI "Hostage Rescue Team." In fact, Microsoft's mistake was not that it lobbied in 2000, but that it failed to lobby in earlier years to reduce or eliminate the budgets of abusive federal divisions and bureaus. When a pack of wolves is tearing into one's tent, an interest in predator control is legitimate, but belated.
Given the new national mood after September 11, Microsoft is probably safe from the political predators, at least for the foreseeable future. The Antitrust Division appears to be less interested in playing to the politics of envy and attacking successful American businesses. Yet one can only wonder how much stronger the information economy, and the rest of our economy, might be today if corporate success depended entirely on satisfying consumers, rather than on the ability to employ or resist the use of antitrust laws as a political tool against business competition.