Browser Wars Revisited : The Forgotten History of Microsoft and its Dirty Days
Before the Hololens?
Ben Thompson in his theory of aggregation suggested consumers will flock towards “aggregators” — companies delivering superior experiences (Netflix, Airbnb, Uber). Then, Thompson applies this theory to antitrust. He says aggregators will soon become monopolies, and that ancient antitrust laws no longer apply because consumers elect monopolies to emerge willingly in exchange for better goods and services.
Thompson’s intuition that the competitive landscape is shifting is correct — but consumer well-being has always gone beyond just allowing for choice and self-selection; the purpose of antitrust law goes far beyond that. And as the landscape continues to shift it has never been more important to understand how to best protect the consumer and her total wellbeing. Fifteen years later, understanding a staple case in antitrust law, United States v. Microsoft Corporation, is the key to understanding and adapting to this issue.
Ladies and gentleman, these are the facts
In 1984, Microsoft was one of the most successful companies in the world. It was basically set to dominate the computing industry. They owned the operating system market and hoped to do the same for internet browsing, Marc Andreesen and Netscape be damned. But that was when the law stepped in.
In 1993, the Department of Justice inquired into possible anticompetitive practices of Microsoft. In a settlement, Microsoft agreed not to *tie* separate Microsoft products into Windows. But the settlement still allowed for Microsoft to integrate features into the operating system. Later, the debate whether the Internet Explorer browser was a separate product or a feature of Windows would become a key issue.
The USDOJ insisted IE was a separate product entirely. Microsoft maintained it was an essential feature of Windows.
The USDOJ, twenty US States, and the District of Columbia sued Microsoft. Assistant Attorney General for Antitrust, Joel Klein prosecuted and judge Thomas Jackson presided over the case. Both Klein and Jackson were very pro-business according to their peers in the legal and political communities. All this to say the facts and law seemingly prevailed over ideological bias in the Microsoft case.
In late 1999 — Judge Jackson published his findings and sided heavily with the prosecution. Jackson established that Microsoft possessed monopoly power in the PC operating system market and used an array of exclusionary tactics to protect their operating system monopoly against competitors. Microsoft’s actions were deemed harmful to both producers and consumers. To justify his findings, Jackson pointed to three major facts.
“Viewed together, three main facts indicate that Microsoft enjoys monopoly power. First, Microsoft’s share of the market for Intel-compatible PC operating systems is extremely large and stable. Second, Microsoft’s dominant market share is protected by a high barrier to entry. Third, and largely as a result of that barrier, Microsoft’s customers lack a commercially viable alternative to Windows (Jackson).”
Keep in mind, Monopoly power is not illegal in and of itself. But U.S. antitrust law does prohibit leveraging that power to engage in anticompetitive practices that harm consumers. Jackson not only ruled that Microsoft was a monopoly, but that it had engaged in practices specifically designed to limit competition and maintain monopoly power. He identified activities which were clearly harmful to the consumer and that limited competitive practices to maintain monopoly power.
Microsoft maintained this power by creating significant barriers to entry. One barrier was the “applications barrier” — the immense ecosystem of applications surrounding Windows. Windows had 70,000 applications available, operating systems had much less. The sheer number of applications on Windows made it more attractive to new PC users than any other operating system. The cost to develop an application ecosystem was already sunk by Microsoft, yet remained a clear entry barrier for competitors. The “applications barrier” was a clear competitive advantage for Microsoft, but was not illegal in itself. It did however increase the cost for any hopeful entrant into the operating systems market. Microsoft now had high potential to engage in anticompetitive practices to maintain its monopoly.
Jackson and the government believed Microsoft engaged in anticompetitive practices by tying Internet Explorer with Windows. The court stated Microsoft leveraged market power to protect its monopoly of the operating system market. It was clear, Microsoft wanted to eliminate Netscape Navigator from the browser market. Microsoft feared Navigator’s ability to support a variety of applications, on many operating systems. Platforms like Navigator became known as middleware — software bridging the gap between operating systems and applications. Middleware became a threat to Microsoft’s operating system monopoly and the primary target of its anticompetitive practices.
The debate whether Internet Explorer extended functionality of Windows, which was legal under the previous settlement, or if it was a product-in-itself distributed with Windows was (and still is!) important. Microsoft selling Internet Explorer for other operating systems supported the idea it was a standalone product. The court made clear that the operating system and web browser markets were distinct from one another, and by forcefully bundling the two products, Microsoft hurt consumers.
Consumers can be hurt by anticompetitive practices in so many ways. Rising prices, limited choices, and lack of innovative products and services are just a few.
On June 7th, 2000, Jackson issued the remedies for the Microsoft case. He ordered Microsoft be split up into two companies, one to deal with operating systems and the other to relate to Microsoft’s other products and services. Microsoft appealed.
What the law says
I’ll try and keep this fun.
The law of the Microsoft case revolves around antitrust. Specifically, the Sherman Act. Congress passed the act in 1890 to prevent abusive monopolies. The act derived its power from Congress’s constitutional ability to regulate interstate commerce. See, abusive monopolies damaged the economy by eliminating competition — allowing them to charge higher prices, provide less choice, produce inferior products, and innovate less. Over its long history, Sherman went from rare invocation, to complete dismantlement, to re-emergence and prominence.
Sometime during Sherman’s history, the “rule of reason” emerged. The “rule of reason” basically means that anticompetitive practices employed to attain or preserve monopoly were illegal but that monopoly in itself was not.
The specific legal allegations against Microsoft were unlawful attempted monopolization and monopoly maintenance under section 2 of the Sherman act. In addition, unlawful tying and exclusive dealing arrangements under section 1 of the Sherman act (remember this, its important).
Jackson found that Microsoft was found to have violated section 2 of the Sherman act by illegally maintaining its operating system monopoly. Under section 1, Microsoft was liable for tying but not exclusive dealing. *By tying the sale of Windows to the sale of Internet Explorer, Microsoft was guilty of tying.* However, Microsoft remained innocent of exclusive dealing because as it had never successfully dominated the browser market it attempted to monopolize (but they still managed to kill Netscape).
Microsoft’s practice of tying Internet Explorer with Windows is what relates to section 1 of the Sherman Act. Legal analysis of tying law under section 1 is particularly contested in the Microsoft case. Violations of section one are “per se” unlawful as compared to the “rule of reason” violations mentioned earlier. “Per se” violations are only required to meet the fundamental conditions set out, and disregard any redeeming virtues like examining practical effects on markets or intentions of corporate leaders. When the conditions are met — the company is held liable for violating the Sherman Act.
A firm was said to be liable for tying violations when four conditions were met.
1. Two separate products are involved.
2. The defendant does not allow customers to purchase one tied product without the other tying product.
3. The arrangement effects a significant volume of interstate commerce.
4. The defendant has market power in the tying product
This is where the problem lies, its difficult to determine if two distinct products exist is an issue in the technology industry.
In Jefferson Parish Hospital District No. 2 v. Hyde, the Supreme Court ruled that the functional relation between products should be disregarded. Instead, the court should look to the respective demand for both products **individually**. If there is enough demand for one product, independent of the other, then the two are considered separate. Judge Jackson employed the Jefferson Parish decision in his own tying analysis of Windows and Internet Explorer.
In Microsoft’s appeal, the D.C. circuit upheld many of the lower court’s rulings. However, they disagreed that Microsoft’s bundling of Internet Explorer and Windows was per se unlawful. The court announced a new tying rule related to “platform software products,” outside the scope of the Supreme Court ruling in Jefferson Parish. The court held that the “rule of reason” rather than “per se analysis” was the appropriate measure for tying. The D.C. circuit returned the case to the lower court for consideration under this new holding.
The new holding determined the Jefferson Parish test could only identify when tying occurred — not whether it was good or bad. The D.C. circuit asked the lower court to balance the “benefits against the costs to consumers.”
This is where the complex and contested legal issue of the case emerges, and the problem that was forgotten. Balancing anticompetitive practices, like forcing customers to buy unwanted products, with the potential benefits of integrating technology products lies at the heart of how tying law must be applied. Tying law needs a revisit to better equip antitrust law to operate in the software and technology industries; unique industries which did not exist when the Sherman Act was first passed. This is only going to become a bigger issue as aggregators tie more and more. These issues apply to anything ecosystem or compatibility related (Kindle AZW, Android OS, iCloud ecosystem).
In the return to the lower court, Judge Kotelly heard the case. On September 6th, 2001, the government stated it no longer seeked to break up Microsoft (much to Netscape’s dismay, I’m sure). Microsoft drafted a settlement allowing PC manufacturers to support non-Microsoft software. The government did not prevent Microsoft from tying its software in future iterations of Windows. But, they did require Microsoft to share APIs with third-party companies.
Where to go from here?
Decades after Microsoft the clearest takeaway is that the rules must adapt. Its so clear that Microsoft engaged in anticompetitive practices and the judicial system was justified in imposing remedies. The all too relevant problem still remaining from Microsoft is how to conduct tying analysis in the technology sector; and how it relates to consumer protection. The decision of the D.C. circuit to adapt Supreme Court precedent for technology affirms that the sector is unique. Likewise, its relationship to tying law should be unique as well to best uphold the purpose of antitrust policy.
Why the D.C. circuit had its unique perspective on the software sector lies in the unique attributes of technology markets — they’re just different. Economists have defined some characteristics of these markets.
1. Strong economies of scale.
2. Network effects.
3. Lock-in effects.
The software industry in particular has high fixed costs coupled with low marginal costs. As production increases, average cost decreases. This adds to strong economies of scale. Network effects are clear when observing a consumer’s utility increase as more people use the product or service. Technology products are subject to strong lock-in effects as well. As a user becomes more familiar with the product, they are less likely to switch. These characteristics allow technology companies to have a more natural tendency towards monopoly. The scope and purpose of antitrust policy then must be reexamined in the context of monopolists who are sometimes emerging and sustaining monopoly naturally rather than by anticompetitive practices.
Under the “rule of reason” championed by the D.C. circuit, a cost-benefit analysis must compare the pros and cons of tying in the technology sector. Tying law must balance the danger of forced buying and loss of consumer choice with market efficiencies to determine if a business like Microsoft is violating the law. Yet, the open ended “rule of reason” does not outline clear guidelines on how to balance the pros and cons!
In The Path of the Law, Oliver Wendell Holmes defines law as a prediction. Good law is a framework for **prediction** of how courts might behave. He puts it like this: “Bad Men,” who cares little for ethics, morality, or consumer well-being, should be able to predict and avoid the consequences of the law.
By failing to provide structure to “rule of reason” tying procedures, the court fails to define good law, for the “Bad Man.” The court fails to inform good companies seeking to innovate via tying, or bad, anticompetitive companies seeking to avoid consequences, proper guidelines for accurate prediction. Because technology companies cannot predict how a trial will turn out, they will be dissuaded from engaging in tying practices. This is bad — the world may be deprived of many potential breakthroughs which emerge from tying technological products in the marketplace.
Sherman prohibits leveraging monopoly power to harm consumers via practices like tying. Yet (as Thompson laid out), due to various external economic factors, technology companies today are able to provide the greatest benefit via this very practice.
The issue is differentiating between good and bad tying. Thompson believes consumers self-select the good ties from the market — allowing only benevolent monopolies to aggregate and survive. But trusting consumers to self-select is **not enough** for the law.
A strict system on which predictions can be based is necessary. Adopting a systematic and multifaceted consumer-manufacturer-innovation matrix to differentiate good and bad tying is close to what courts should adopt.
Samuel Noah Weinstein, a scholar on the Microsoft case, proposes such a system which asks the relevant questions:
1. Was there separate demand for the integrated items at the time of the integration?
2. Does there continue to be separate demand for the stand-alone items after the integration?
3. How do other manufacturers view the product market?
4. What was the intent of the monopolist in combining the products?
5. Is there evidence that the integration is a genuine technological advance?
The general purpose is to gauge not only how consumers view tying but also consider economics, manufacturers, and innovation. Thompson suggests leaving corrections up to the market and the consumer. But when companies ultimately harm “self-selecting” consumers via anti-competitive practices, only the law remains to protect them. A multifaceted, strict framework for the technology sector is a step in the right direction. It is a legal compass Oliver Wendell Holmes would be proud of. And the best thing for consumers.
To know that they had broken the law to do it, to know that they were essentially taking food from the mouth of my children really made me angry. They set out specifically to kill our company. I think that would make anyone angry.
— Lou Montulli — early member of the Netscape team in response to Microsoft’s illegal activities.
Brought to you by Netscape