April 21, 1998

Microsoft Versus the World
by Jim Surowiecki (TMF Cinder)

Antitrust law represents an attempt to deal with one of the paradoxes of free markets, which is that market forces sometimes produce a situation in which one company owns a monopoly of the supply of a given product. Oddly, but importantly, monopoly itself is not necessarily illegal. Rather, it's the creation of a monopoly by predatory behavior or, alternatively, using monopoly power in one market to create an anticompetitive situation in other markets that tends to bring the Justice Department sniffing around. And it's the second aspect of monopoly power in particular that is at issue in the ongoing tango between Justice and Microsoft (Nasdaq: MSFT).

That Microsoft does, in fact, own a monopoly on the operating system market for personal computers is probably not worth debating. Although Apple Computer (Nasdaq: AAPL) still makes a qualitatively competitive product, its minuscule 3% of the market speaks for itself, and competitors like OS/2, which Microsoft actually helped design for IBM (NYSE: IBM), fell by the wayside almost as soon as they appeared. An interesting challenge is being offered at the moment by Java, the programming language designed by Sun Microsystems (Nasdaq: SUNW) to work on computers regardless of their operating system, but right now Java appears more likely to become an adjunct to the Windows world rather than a replacement of it. In any case, at this moment, more than 90% of the world's PCs run on some form of Windows software, and that percentage is increasing rather than decreasing.

One of the problems with thinking about whether Microsoft has violated antitrust law in earning that market share is that it's difficult to point to any one decision or one deal that ensured its success. To be sure, Apple's decision in the mid-1980s not to license the Mac OS software was probably crucial. Similarly, Microsoft's aggressive incorporation of the look of the Mac OS, coupled with its continued improvements in successive versions of Windows, has made the product qualitatively better. But on some level the real key may simply have been Microsoft's early recognition that personal computing was going to gravitate toward common standards, and that the company that could get its operating system into as many PCs as possible early on was going to improve dramatically its chances of getting its operating system into every PC later.

That's a crude version of what is now well-known as the role of "network effects" in facilitating Microsoft's success. Network effects refer simply to those situations in which the value of a product to users increases exponentially the more people have it. One phone is useless, two phones are more valuable, a thousand phones and you're on the way to a different world. If everyone else in the world is using Windows, and if everyone is making software programs for Windows, it makes more sense for you to use Windows. And if you're a software maker, and all the world's PC users are on Windows, then it makes sense for you to program for Windows, which in turn makes it less likely for people to use Macs, thus making it less likely that people will program for Macs, and so on.

Looking at that picture, one can understand why people are concerned that Microsoft's domination of the operating system market will hinder innovation and take away from the dynamism of the U.S. software industry. The odds against gaining even a foothold in the OS market are so huge that it's hard to imagine anyone even trying. And if, as people argue Microsoft is already doing, the Redmond, Washington giant uses its control of the operating system to gain a similar position in the Internet browser market, which could give it in turn a tight grip on the e-commerce market, then competition in that arena might be stifled, too. Implicit in this argument is the idea that at some point Microsoft will begin to use its monopolistic power more aggressively to raise prices. In some sense, the company is already doing that in the OS market, where Windows 98 will be twice as expensive as Windows 3.1 was. Though it's certainly a superior product, everywhere else in the industry price-for-performance ratios are falling sharply. Only at Microsoft are they standing still or even rising.

It's certainly true that Microsoft has engaged in behavior that can be described as anticompetitive, most notably when it has used PC makers' dependence on Windows to accept exclusionary deals, including its requirement that Internet Explorer be included with every PC and that icons for Explorer and the Microsoft Network be displayed on the opening screen. And though Microsoft recently abandoned an arrangement with a variety of Internet service providers whereby they would push Explorer on people who accessed their service from Windows, it did so only as part of its attempt to soften its public image and to assuage antitrust concerns. In any case, now that Explorer and Windows have been seamlessly merged into Windows 98, one suspects that the fight for the browser market is over, and that within a couple of years Netscape (Nasdaq: NSCP) will be to the browser world what Apple now is to the PC world, a romantic vestige of a time when people didn't use a product owned by Bill Gates.

It's also true that Microsoft's public posturing has not reflected well on it. The recent debacle involving Microsoft's secret plan to pay journalists to write op-ed pieces in its favor and to plant positive stories confirmed the company's opponents' worst fears, and the company's early stance of almost flippant rejection of the Justice Department's concerns was at best a strategic error. While one can understand the company's desire to control what goes into its products, its almost willful naivete about the reality of the power that it derives from Windows is unconvincing, as are arguments from those who suggest that if PC makers don't like the way Microsoft does business, they can just refuse to license Windows. Dell (Nasdaq: DELL) certainly could refuse to put Windows on its computers, but doing so would be equivalent to suicide, and it's not necessarily in society's best interest that Microsoft be unrestricted in its ability to make PC manufacturers choose between its best offer and suicide.

Yet, for all of this, it's hard to see how antitrust law is really well-suited to the Microsoft situation. The real problem is that antitrust law is simultaneously a too blunt and a too fine instrument. If used to break up Microsoft, which is after all only a $10 billion company, the bluntness of the blow would do nothing but take away from Bill Gates' profits. Even if the company were forced, as Business Week recently suggested it might be, to divide itself into an operating systems company and an applications company, much as AT&T (NYSE: T) was separated into a long-distance division and the Baby Bells, that would have no effect on Windows' status as the OS standard. It might help companies in the office suite, financial services, and perhaps even browser software markets, but it's not exactly clear why the Microsoft applications company -- whatever it would be called -- could be expected to lose the dominant position it currently enjoys.

On the other hand, antitrust law would be too fine an instrument if it meant that a judge had to take over running Microsoft, ruling on a case-by-case basis about the acceptability of contractual arrangements or the integration of products into the OS. That kind of oversight would be both legally difficult to justify and would wreak unnecessary havoc on the operations of the software market. It would most likely have the effect of stifling innovation far more convincingly than Microsoft's control of Windows does.

There is unquestionably a certain amount of cynicism in Microsoft's persistent claims that its success is the product of the computer industry's ceaseless quest for innovation, and that getting in its way will only hinder the pursuit of excellence. Microsoft's products have never been considered the best from the standpoint of quality, and the company's real genius has not been the creation of new products but rather the acquisition and incorporation of other people's innovations into its releases. Nevertheless, the simple truth is that each generation of Microsoft's products has been better than the last, and that its ability to recognize where consumers were going and to answer their needs -- however crudely -- has been far superior to that of its competitors. As with Intel (Nasdaq: INTC), Microsoft's huge war chest and commanding market share have allowed it to recover quickly when it has made mistakes, but the fact is that its ability to understand its own errors and correct them has been a huge part of its success.

In that sense, in fact, it's easy to construct a narrative of the past ten years in which Microsoft did not become the dominant software company that it is today. Gates once thought that pen computing was the wave of the future. He did not understand the importance of the Internet. He thought that the "TV PC" would quickly replace the desktop. At any of these turns, Microsoft could have invested heavily enough in technology that was going nowhere, and had it done so it would have paid the price. In that sense, Microsoft is not exactly like the old AT&T or like the trusts of the late 1800s. It is not impervious.

It's interesting to ask yourself if you believe that ten years from now everyone will still be running Windows on their desktops. That strikes me as a difficult question to answer, because while on the one hand the PC is a singularly useful device, the advent of the Internet -- which is still, remember, in its infancy -- does suggest a future in which applications will be easily downloadable and in which networking will make the isolated PC obsolete. After all, you don't need to own a Mac to talk to someone using a Mac on the Internet. From the other end, the gradual incorporation of microchip technology in everyday devices may also be working toward the PC's obsolescence. The point, in any case, is that Windows is not necessarily the operating system for the next century. Though it may, of course, very well be.

To generalize wildly, Americans have what seems like an ingrained distrust of centralized power, whether in the form of the government or business monopolies. And in a world in which PCs are increasingly ubiquitous and increasingly important, Microsoft's stranglehold on the PC operating system makes it an obvious target of that distrust. But the negative consequences of that stranglehold are very hard to identify, no matter what Scott McNealy or Jim Barksdale say. More importantly, overestimating the durability of that stranglehold is very easy in a world in which technological paradigms come and go with remarkable speed. It's undoubtedly appropriate for the Justice Department to inquire into and take action against practices -- like the exclusionary browser deals -- that seem openly anticompetitive. But the case for broader antitrust actions against Microsoft has not in any real sense been made.

Agree? Disagree? Check out the Microsoft message board!

Other Links of Interest:
Information on Tuesday's Court Action (WashingtonPost.com)
A Q&A with Bill Gates (Washingtonpost.com)
Tuesday's Microsoft Press Release