The Power of Independent Thinking


Stay Connected
Get the latest updates straight to your inbox.

The Microsoft Monopoly
December 9, 1999
Peter W. Huber, Thomas M. Lenard, Stan J. Liebowitz


Judy Pendell, Center for Legal Policy, The Manhattan Institute

May I have your attention please? May I have your attention? Hello. Welcome. I’m delighted to see such a fine crowd here today. We think we have a program that you’re going to enjoy very much.

I’m Judy Pendell, I head the Center for Legal Policy of The Manhattan Institute.

I’d like to first thank the New York City Chamber of Commerce who were helpful to us in organizing this program today. I’d also like to mention that we have some handouts on the table. As you leave the room please take note and take one with you.

Without further ado I’d like to bring to the podium someone who really needs no introduction to you. Peter Huber, Senior Fellow at the Manhattan Institute. [Applause]

Peter Huber, Senior Fellow, Manhattan Institute

When I arrived as a freshman at MIT in 1970, some three hundred of us shared time on an IBM 1130. I remember the machine well. My daughter in her Dell, she’s eleven years old, has, I’d say, somewhere between a hundred and a thousand times as much power on her desktop today. We communicated campus-wide on a system called Multix which as I recall ran about 60 bits per second. My DSL line is now 1.6 megabits per second. My first Fortran program was 110 lines; I still have it. It’s 110 lines of code. I was very proud of it. And I’m told that Microsoft Windows is now topping thirty million lines, though of course, it depends how you define Microsoft Windows. Whether you include the browser or not. [Laughter]

This raises many questions. Among them is 30 million too many. Where does it stop, if ever? How will we know when it ought to stop? Who will decide and by what process? These are not easy questions. Some years ago, not thirty years ago, but some years go I did a Crossfire debate on Microsoft’s venality. I’ve forgotten most of the details of what we all said except that our interlocutors at the time were Bill Press and John Sununu, and I was told I cannot say, this isn’t necessarily true, but I was told when I arrived at the studio they had had to flip a coin to decide which one would be on the right and which would be left on that one. And it was with some reluctance that one of them had taken one side, but I’ve forgotten which.

Anyway, enough from me. We have today for the prosecution and for the defense, the authors of the two books you have on your table. I will introduce them very briefly, leaving more time for them to speak.

For the prosecution, or the plaintiff, we have Tom Lenard who’s Vice President for Research and Senior Fellow at the Progress & Freedom Foundation, which is a counterpart to the Manhattan Institute based in Washington. Dr. Lenard previous served in government at the Office of Management & Budget, the Federal Trade Commission, the Council on Wage & Price Stability. He has an extensive and distinguished career on competitive matters, particularly those relating to the economics of high technology.

I will also introduce Dr. Stan Liebowitz for the defense, or the defendant though of course, neither of our speakers today actually represent these parties. Dr. Liebowitz is Research Fellow at The Independent Institute and Professor of Managerial Economics, the School of Management, the University of Texas at Dallas. He is the author of the other book on your table, Winners Losers & Microsoft, published by The Independent Institute. He has an extensive list of academic publications. Among other things he is one of the few people in the world who actually knows, because he’s researched this with care, where the QWERTY keyboard came from and it’s not where you think it came from, suffice it to say. Read his book, you’ll find out. [Applause]

Thomas M. Lenard, Vice President for Research, Progress and Freedom Foundation

Thank you Peter, and I’d like to thank the Manhattan Institute for arranging this talk, and Peter, who is a long-time friend of The Progress & Freedom Foundation, has participated in a lot of our events. I was thinking when he was talking about the prosecution, and left and right, I guess I’m here for the prosecution. But it seems to me in this particular issue, the lines between left and right are not as clear as in some other instances. I usually probably identified with being on the right, on this side. Let me just say that I hope I am speaking in the interest of competition in this particular industry.

In any event, the Microsoft case obviously is going to, whatever the result, is going to be a landmark case in the history of antitrust and it is going to essentially answer the question of whether there is a role for antitrust in the digital economy. And if there is one, what that role is going to be. Obviously the information technology industry is the information technology sector, is one of the great economic success stories of our time. It’s a fantastic engine for growth, productivity, increasing standards of living.

However, the emergence of one company in an overwhelmingly dominant position in that industry is problematic as far as competition in that industry is concerned. The concern from an antitrust perspective is not simply that Microsoft has a monopoly in the operating system for personal computers. That is not illegal in and of itself. But rather, that it uses that monopoly to raise barriers to entry both in the market for operating systems as well as in complementary markets. That’s a legitimate public policy concern, even for those who would generally favor a rather cautious antitrust policy.

The recent findings of fact that the District Court in Washington issued, Judge Penfield Jackson issued, show that what we have here in the Microsoft case is not just a business practice or two that have strayed over the line, but rather a broad anti-competitive campaign detailed in two hundred pages of the judge’s findings, that was undertaken by Microsoft to protect and extend its operating system monopoly. So given the magnitude of the expenses the remedy should be a serious one. Not punitive, because there’s obviously no point in that, but one that really does try to promote competition.

But obviously in crafting such a remedy we need to be concerned about the dangers of regulating the software industry which I don’t think anybody is in favor of. I’m going to argue that a breakup, the right kind of a breakup of Microsoft, a restructuring of Microsoft is both the most pro-competitive and at the same time the least regulatory of the available remedies.

Obviously the issue of remedy has to be seen in the context of what has been proven in court. And, you know, not withstanding the claims of Microsoft’s defenders I from time to time have been on podiums with various Microsoft defenders, and one of the first things they all say is well, Microsoft doesn’t even have a monopoly. Well, for those of us who’ve been in a computer store recently, I would challenge any one of you to buy a computer that doesn’t have Windows preinstalled. I mean, and the proof of Microsoft’s monopoly in court has been overwhelming.

Every year for the past decade Microsoft has had over ninety percent of the market for PC operating systems. For the last couple of years the market, they have had over ninety-five percent. So this is a durable monopoly which has persisted and expanded over time. There are no significant competitors. There are no substitutes. Even its large customers, Gateway, Dell, IBM have no bargaining power with respect to Microsoft. And there are very high barriers to entry.

Entry, this is not as some people believe, these high-technology markets, notwithstanding the rapid pace of change, this is not a market where all you need is a couple of bright programmers in a garage to enter. It is very difficult to enter markets with these type of, that have been dominated by a single firm where you have all these types of network effects. Once a technology like Windows becomes dominant, a potential entrant finds itself faced with a large installed base of users who have a large supply of applications that have already been for this program and the assurance that programmers are going to continue writing programs for this operating system. Obviously the incentive, especially given the cost structure of writing software programs where the major cost is basically the first copy cost, the incentive to write programs for a new entrant operating system that may have a very small fraction of the market is not very significant.

So a potential entrant would face a very high hurdle. There are significant switching costs to anybody who wants to switch to a new program. Entry is clearly not easy, and this in fact is why software firms have such large, at least the successful ones, have such large market capitalizations. And this is what Judge Jackson found. And in his findings he said the entry of the new operating system was severely constrained by this technical chicken and egg problem. Which he called the applications barrier to entry.

Now given this general scenario, once Navigator and Java, the computer browser in Java, a programming language, started to appear, Microsoft saw correctly the immediate threat that they were. Because essentially these were new technologies that could erode its monopoly, could erode the applications barrier to entry, because they are both examples of middleware, where you can write the programs to the middleware and then those can be placed on top of a lot of operating systems. And so the applications then become operating system neutral; they’re not specific to a particular operating system and that dramatically would erode Microsoft’s market position and obviously they saw that threat.

And Navigator was released first in December 1994, and in May 1995 Bill Gates wrote a memo that warned Netscape was quote “Pursuing a multiplatform strategy where they move the key APIs into the client, the APIs being the Application Program Interfaces, the hooks under which the applications program are written for, key APIs into the client to commoditize the underlying operating system.”

So Microsoft had a clear incentive to engage in exclusionary practices to protect the operating system monopoly against these threats even at the cost of large foregone profits. And it also had the ability, by leveraging its operating system monopoly to do so.

After initially trying and failing to gain a market division agreement with Netscape, essentially to keep Netscape out of the Windows market, Microsoft followed a strategy of leveraging its operating system monopoly to block Netscape using a variety of means. Pricing, contractual and technical, essentially to block Netscape from the two main avenues through which virtually everybody gets their browser. They either get it, either the OEM market, the Original Equipment Manufacturing market, or the Internet access market. Those are the two ways that almost everybody gets their browser, either preinstalled with the machine that they buy or from their Internet access provider like AOL.

Obviously with respect to the equipment manufacturers, the OEMs, they all need Windows. They can’t be in business without a Windows license, without having Windows installed on their machine. So it’s pretty easy to say if you want our license you have to take Internet Explorer and you have to not sell Navigator. It’s essentially what happened. Although the judge went on for many, many pages in his findings detailing how Microsoft forced the OEMs to take the Internet Explorer with Windows, forbade them to remove it or obscure it, imposed a variety of technical restrictions to increase the cost of promoting Navigator. Threatened to penalize OEMs that insisted on preinstalling and promoting it.

On the other side, on the Internet access, Microsoft gave valuable considerations to Internet access providers that agreed to distribute the Explorer either exclusively or virtually exclusively. For example, they made a deal with America Online where they would distribute AOL in Windows. They gave AOL valuable placement on the Microsoft desktop in competition with its own access provider, Microsoft Network. Essentially sacrificed Microsoft network to this effort. They even gave them direct payments. All of this to promote a product that they were giving away for free. Internet Explorer.

Now, this strategy was successful before 1996. Navigator had substantial and growing presence on the desktop of new personal computers. By January 1998, of the sixty original equipment manufacturer subchannels that the judge talks about in his findings, fifteen OEMs, each of which, and four subchannels reach, Navigator was being shipped on only four. Mostly with the icon not on the desktop. For example, Sony featured Navigator in a folder rather than on the desktop, and Gateway shipped a separate CD rather than preinstalling it. So by 1999 Navigator was present on only a very tiny fraction of the computers that were shipped.

So Microsoft clearly won the browser war. They now have a share of over sixty percent. As people continue to upgrade Windows that will go rapidly upward.

And these are not the only threats that Microsoft dealt with. They were able to convince companies like Intel. For example, Intel was developing software that had provided directly to the OEMs and other purchasers of Intel chips. Microsoft was concerned about that; that that again, eroded its position. They were able to convince Intel to stop doing that.

IBM was in competition with Microsoft on the software side of the market and Microsoft punished the personal computer arm of IBM with higher prices for the Windows license, a delayed license for Windows 95 and withholding of technical and marketing support.

In the latter half of the 1990s, IBM along with Gateway paid significantly more for Windows than other major OEMs like Compaq, Dell and HP that were more compliant with Microsoft’s wishes.

Now the cost of all this to consumers, there are a variety of costs of all this to consumers. First of all, prices are higher. Microsoft acknowledged essentially that its pricing decisions were basically unconstrained by competition. They have extraordinary profit margins even compared to other firms in the computer and software industry. The evidence of a trial was that their profit margin on sales was close to forty percent. The most recent quarter was over fifty percent. There was a memo that was produced at trial saying they could have profitably sold Windows at forty-nine dollars but they charged eighty-nine dollars instead.

But I think probably the most important costs to consumers really are in the effect that this entire arrangement has had on innovations. Obviously harder to quantify. But there is now no threat of competition in the personal computer operating system market, either directly or from middleware, like a competing browser or a Java. There is a lack of differentiation between the offerings of personal computer manufacturers, due to contractual restrictions imposed by Microsoft. You buy a computer from virtually anybody and you go home and you boot it up and it looks pretty much the same.

Microsoft devotes a considerable amount of its efforts, its research efforts and its marketing efforts to maintaining and extending its monopoly. And obviously that’s a rational thing for it to do. Those profits are really pretty hefty. And they devote a considerable amount of their efforts to maintaining that, rather than meeting customer needs. Because customers really have no place else to go. Microsoft dictates the innovation schedule for both applications and hardware. And the judge found that that results in some innovations never occurring. And of course, the market is distorted by the fact that firms stay out of areas where Microsoft is likely to be dominant.

Well, all that brings us to the question of what to do about this, which is obviously a difficult issue. I’m going to propose a remedy that is consistent with two statements. One by Judge Posner, hopefully consistent by Judge Posner, and one by Bill Gates.

Judge Posner said, in his well-known book on antitrust, that the proper purpose of the antitrust laws is to promote competition as that term is understood in economics. Bill Gates has said that Microsoft can never accept any remedy that limits its ability to design software.

Well, let me move to the remedy. I mean, there’s a whole range of remedies, ranging from conduct to structural remedies, and the big disadvantage of the conduct remedies is that it is hard to see how you could have a conduct remedy that actually addresses the findings of the court without basically having the government looking over Microsoft’s shoulders every step of the way, both with structured business practices and with respect to its, even extending to its design decisions.

Now structural remedies, and the structural remedies are of various kinds. The one that has been talked about a lot is kind of an AT&T solution where you kind of cut the firm in half with, you know, separate the applications part of the firm from the operating part of the firm. There are two problems with that. One, is it doesn’t address the issue of the operating system monopoly because the operating system monopoly remains. And secondly, for a variety of reasons you’d have to supplement that with conduct remedies just as you did in the AT&T case so you wouldn’t overcome the problem with conduct remedies.

The one solution that suggests, and I can’t say it’s original with me, it’s been suggested by several people, is kind of a hybrid solution where you first divide the, and the other structural remedy is kind of a vertical solution where you just cut the firm in thirds, three equal firms, but that’s a little extreme, given the fact that the case has only been about the operating system, not about the rest. So that leads us to a hybrid solution where you separate the applications from the operating system and then you divide the operating system in thirds and essentially give all, you’d have three identical operating system firms that would be competing with each other. And you’d have the applications firms as a separate firm.

And the nice thing about this is that it will require very, very minimal supervision. You’d have to have a transition period that essentially, to assure that the firms don’t recombine in some way or another during the first three years. That they don’t just undo the effects of the decree. But even after a few years you wouldn’t even have to have that because the industry would have changed so much. So you’d have to have virtually no supervision by the Justice Department or the court. And it would really kick-start competition in this area. You would have three operating system firms competing with each other and you’d have a big, you’d have price declines, you’d have much more service to customers, you’d have, I think, probably a big wave of new innovation that would get started. And without the downsides of having government supervision over the software industry.

And why don’t I stop there. [Applause]

Stan J. Liebowitz, Research Fellow, The Independent Institute; Professor of Economics, University of Texas at Dallas

Well, I don’t usually carry a computer up here when I give a presentation, but I figure I’ll give it a try. And then I‘m told it’s not a good idea to do things for the first time when you have a big audience like this. But you’ll see why, in fact, I actually planned a little multimedia ten seconds for you.

I’m actually going to be able to talk about what I planned to talk about, because on the particular proposal I was apparently very prescient on thinking what it might be. The first page of your handout, which actually was out of order, my secretary sent it somewhat randomly, is a cost of the proposal that was just recommended. Which is an estimate of what the cost would be to developers of having three versions of Windows to write for. Now those are not the only costs; those are just the costs of development. I’ll be able to talk about that in more detail. But you can take a look.

And it’s approximately thirty billion dollars over a three year period. And this is based on asking developers how much extra it cost them to produce different copies of Linux, of Unix and other costs of porting programs from one operating system to another.

But let me get on to the main point that I want to talk about, which is, what is it that we should do about Microsoft, if anything. And to some extent, what both of us are doing is rehashing the case, which is what I thought we might do. I was afraid I was going to have to back into it, but I can probably be a little more direct now. But I started this talk on the assumption that we were supposed to talk about what the remedies would be, which of course assumes some guilt. I do think that the lower court will find Microsoft guilty.

But the way I thought I’d back into rehashing the case was to talk about using the aphorism, the punishment should fit the crime, and then talk about what the crime might have been.

Now one of the things we need to know is who are in the injured parties, okay? And the judge goes on and presents a well-written document for quite a few pages, four hundred and something paragraphs. And in there he gives a litany of various nasty activities that Microsoft engaged in, some of which were just mentioned. One of them was IBM. Microsoft was bullying IBM. Was nasty to IBM. And what Microsoft was basically doing was threatening to withhold Windows from IBM, we’re told. Unless IBM stopped promoting its competitor in the applications market, which was the Lotus Suite, which was a word processor and a spreadsheet and what not, and Microsoft has Microsoft Office, which most of you probably use, which also has a word processor and a spreadsheet and some other programs.

And it went right now to the last minute. And Microsoft kept saying to IBM we’re not going to let you have Windows unless you guys behave better. Presumably IBM said no. Tough. And they got Windows. Maybe Microsoft shouldn’t have been so nasty. Maybe they shouldn’t have been so tough. But nothing really happened there.

Another example that was brought forward of Microsoft’s behavior was with Apple. And Apple actually comes up twice in the finding of fact. And again, when you read it it’s hard to know exactly what Microsoft is threatening. But apparently it was threatening to compete really hard against Apple, if Apple kept developing QuickTime, which is an audio/visual program for both Windows and the Macintosh. And Microsoft said gee, we’d really like it if you didn’t create this program. Because we have our own version. And we’re going to work really hard on our version; we’re going to be really tough if you don’t stop creating yours. And Apple said no, we’re going to keep working on QuickTime.

Microsoft now is presumably working on its own. And these are supposedly indications of the nasty behavior that Microsoft engaged in because the judge went on for some length talking about these. I’m not sure these really are such terrible behaviors in there that we can talk about. But this is a case where Microsoft appears to be bargaining in a way that was not as pleasant as it might be.

But normally when people bargain, which may not be the subject necessarily of antitrust, we don’t have E-mails talking about it. That’s sort of the one novelty in this case that’s been the biggest problem for Microsoft, is their E-mails. But, you know, if you bargain you can often say lots of things like, you know, I’m going to go to someone else, I’m going to stop dealing with you, I’m going to do this, I’m going to do that, I’m going to shoot myself, I’m going to hold my breath until I turn blue and die. I mean, you know, kids do that all the time. But the bottom line is what you wind up doing, and in those cases Microsoft didn’t do anything.

Now the other set of potential harms, and there are several of them, and the one that should be the most important I think is the harm to consumers. And the judge has a very hard time coming up with any concrete harm to consumers. He has five paragraphs he devotes to that. The harms are almost laughable. He talks about how there’s wasted hard drive space. Because if you don’t want to use Internet Explorer it’s still sitting on your hard drive, and that’s taking up all this space and that’s a waste. That’s true. But that’s not much of a harm because I don’t think there’s a single program that you have on your hard drive that’s not wasting some space because you’re not using every part of that program.

Then there’s the cost of removing Internet Explorer for the people who want to remove it. It takes time and effort to do that. Well, that’s a cost. Well, it’s true, but it’s not really a very big cost. It’s not the end of the world. I don’t think it’s something you break a company apart over.

And then there’s the confusion that people might have when Internet Explorer is called by some other program and you don’t know what’s happening on the screen. Well for people who are really confused by that and are really upset, they’re probably confused by a lot of things that are happening on their computer. And that’s a cost, but I don’t think that’s the type of cost that antitrust should be dealing with either.

Now the reason the judge has so much trouble finding harm to consumers is because there’s virtually no harm to be found. That the evidence, if you look at the broader picture, which is beyond the scope, perhaps of the case itself, is that Microsoft’s had a terrific impact on consumers.

If you look at page 3 of the handout, and I guess I should look at it, my notes say it’s page 3, and sure enough it is page 3. Yes. Somewhat out of order. You can take a look at prices in some real markets over a period of time. This is not hypothetical, this is not theoretical; these are actual prices. And it’s really one chart that I’ve just split in two. And it’s a chart of word processor and spreadsheet prices. And I’ve broken it at 1991-92. So you can see what was going on in the first half and what you see what’s going on in the second half.

The first half, Microsoft is not the dominant player in word processors and spreadsheets. Most of you probably know, WordPerfect was the dominant word processor; Lotus 123 was the dominant spreadsheet. And if you take a look at the left-hand side of the chart the prices are two things: they’re high and they’re not going down. They’re just sort of sitting around.

The second half is when Microsoft became the dominant player in that market. And you see a very, very steep and very, very large decline in software prices. Now I think actually we should give Microsoft some credit for that; they’re probably responsible for that.

Now there has been a claim made, based on this evidence, that well, maybe there was just some uniform decrease in prices going on in software that just happened to occur at the same time that Microsoft was becoming dominant in the applications.

So if you go to the next page, on page 4, you can take a look at the average prices of every application to consumers that DataQuest follows over a period of time where they had contiguous definitions which ended in 1995. They changed the definition of application markets that year. And we broke the applications into two categories: those that competed with Microsoft and those that didn’t. The diamond line on the top are the applications that don’t compete with Microsoft. And yes, there was a small price decline over this period of time. Something on the order of twelve percent.

If you take a look at the markets where Microsoft competes, and this will be true whether it’s weighted by how large the market is or whether it’s unweighted, so it’s not dominated by word processors and spreadsheets. The average fall is approximately sixty percent in price, okay? So that there’s no general, overall decline in prices that you could say is responsible in the markets where Microsoft competes. That the Microsoft markets that are different.

Now it could be that there’s some other factor going on, but no one has suggested any. And what I suggest is that Microsoft is in fact more competitive. It’s like the Wal-Mart of software. And that’s why prices have come down. And that’s also why it’s facing the antitrust case. Because there’s some people who are threatened by Wal-Mart. Just as you know, Wal-Mart’s being prosecuted in certain states where they’re trying to pass legislation that says you can’t have a grocery store that’s more than a certain square footage, because the grocery stores in those states don’t want to have to compete with Wal-Mart because Wal-Mart’s going to have low prices.

I’m actually going to skip out of order here, but the actual history of this case goes back beyond before this case started officially. I started talking about it in 1995 based on research I had done, by the way, with my co-author. All of this is with my co-author, Steve Margolis who’s sitting here somewhere. And the reason we got into it was not because we’d done any work on Microsoft, and not because we’d been doing anything on antitrust per se. But we’ve been doing work on something called network effects. And in particular something called lock-in, or what we might call these days the applications barrier to entry.

And there was a theory that we had been criticizing for a long time as being very unrealistic. And okay, my computer just shut off, so- One of the problems was that this theory had no empirical support. Just like the claims that have been made about software prices have virtually no empirical support. And we were asked to look because at that time there was a fella named Gary Reback who was a lawyer representing certain anonymous software firms that were so afraid of Microsoft they were afraid to be mentioned.

And he was using these theories to attack Microsoft. And at that time it looked like AOL was going to be the story, but Microsoft was including Microsoft Network with the operating system just as it later included Internet Explorer. And they were thinking they might be able to bring a case with the Microsoft Network. The Microsoft Network never went anywhere and never got any big market share. AOL did great. So that potential case fizzled out.

But then they brought the same theory to this case with Internet Explorer and Netscape. Netscape did get beaten. Internet Explorer is a better browser now and it certainly has a bigger market share. And so now they have a case and that’s the case they brought.

But the case was really brought at the behest of Sun and Oracle. Why are they so interested in this case? And the answer is that they’re afraid of Windows NT. They really don’t care very much about Windows 98; that’s not their market. But Windows NT goes right into where their bread and butter is. And if they could get the government to bring a case that would punish Microsoft in its major market, that might slow Microsoft down in its progress in competing with Sun and Oracle in the high-end work station market where Microsoft’s been going all along, and made it very clear that’s where NT was going to go. That’s what this case is really about, okay?

But the fact is that that’s all I’m going to say about that, but that’s sort of a little bit of the politics and we were involved in those theories in 1995.

Now the harm that’s been mentioned, I’ve mentioned two which I’m not very enthusiastic about. There’s harm to Netscape. And undoubtedly Netscape has been harmed. Its market share is lower than it was. But it’s been harmed because Microsoft produced a better product at a lower price. Even the judge admits that. If that’s all it was there’s no harm to consumers and it’s not clear that there’s anything terribly wrong, unless he wants to say there’s predation which it’s not clear he wants to say.

So then what harm can there be? And what the judge says is well, there’s a harm to competition. And that’s because Netscape wasn’t just a browser, but Netscape is really a potential operating system. That it’s the Trojan Horse for Java. Where Java is this language that’s going to essentially turn into an operating system and people are going to write programs for Java, and then Microsoft will have a competitor so Microsoft needs to kill Netscape so that therefore Java is hurt. Java, by the way, is Sun’s program, okay? So the same participants keep coming back into this.

The problem with that is that it is based on this idea of the application barrier to entry which is this lock-in idea. It’s an interesting theory; it could be true. There’s just no evidence that it’s ever been true. That there’s ever been a case of lock-in. All right. This is a theory with no support whatsoever. It’s a theory that has been tested where evidence has been put forward and then found to be incorrect. This is a theory that’s been spun out of thin air. The logic is somewhat sensical, but, you know, that’s what the book is about. The book is about lock-in and how it doesn’t occur, and how markets work. And the better products actually win even when we thought that might not have in the past.

We spend a particular chapter or two talking about software markets. And what we find there is that even though the theory predicts that we might get stuck with a single firm and it would be very hard to dislodge an entrenched incumbent because of these network effects, and in fact the opposite seems to happen. That yes, the incumbents have very big market shares. That is a function of the network effects. But they get overturned very quickly. Much more rapidly than happens in normal industries and quite frequently. And they get overturned, the key turns out to be whenever somebody comes along with a better product. But when you come along with an inferior product and you don’t overturn the incumbent, is that any big surprise? Okay. And that’s what we’re talking about in some of these cases.

Okay. Now what I think is going on is that there’s something that my dissertation adviser Harold Demsich used to call the nirvana fallacy. Which was at the time that the government could do things better than the market could, so if there was some market failure we should have the government take over. There is here what I think is a different type of nirvana fallacy, which is that yes, the government’s being used but there’s some sort of ideal competition that’s going to occur in this market if in fact we could just jump-start it and get Microsoft out of the picture.

Now this is where, and I don’t know how much time I have left, I was going to bring my little audio/visual in, anybody can take flights to fancy and the particular one I wanted to take was that Christmas movie, “It’s a Wonderful Life” with Jimmy Stewart. Imagine, it’s hard to imagine, really, Bill Gates being the Jimmy Stewart character here, but imagine that we had a world without Microsoft. Where is it? After all this talk I can’t find my little icon. It must a problem with Windows. It must be something worth suing.

At any rate, I’m afraid I may not have it, but I had a little clip where Jimmy Stewart says, you know, the world would have been better off without me. And then they go to Potterville, if you remember, and Potterville was the evil banker who is now controlling everything because Jimmy Stewart hadn’t been there to help. The version of that that I wanted to give you was well, let’s say Microsoft hadn’t existed. What would a real market for operating systems tend to look like? Not some fantasy we have of lots of great competition going on, because that’s not the way these markets tend to operate.

Well, there’s one little story that I should tell you which is about IBM, which I’m not going to say we should do anything about, but IBM has something called OS2. The judge talks about OS2 and talks about how it should have been able to do better against Windows, but that nasty Microsoft just kept it down with its nasty behavior. What the judge doesn’t tell us is anything about the history of OS2.

When it initially came out there were going to be two versions of OS2. And this might have been the real alternative to Windows. One version that would run only on IBM PCs. You may remember something called the Microchannel, a special architecture that IBM had that no one else had because they had patented it and it was idiosyncratic to them. And it wasn’t clear there was any improvement in that technology, but they had this operating system sitting in the wings that was only going to run on machines that had that architecture. And nobody else could have it because it was patented.

They had another version of OS2 that was going to run on everyone else’s computer. The version that was going to run on the IBM machines was a better version of OS2. It was going to have built-in networking and a few other communications abilities. And then there was the OS2 Light which the other producers were going to have. Now, why was IBM doing that? Well, they were doing that because they wanted control again of the business PCs. If you had a PC that was going to be networked you were going to have to buy their version of OS2 and therefore their machine because that was the only one it was going to run on.

Now that never came to pass. And it didn’t come to pass because sort of like the Jimmy Stewart character. Bill Gates was there with his cheap version of Windows and they kept Potter from his evil ways. And so we don’t have only one business computer running, we have Dell and Gateway and lots of others out there besides IBMs. But this is the way the industry really works, is that’s what would have happened if IBM weren’t there.

We can talk about what the world would have been like if Apple had been in control, where we’d only have one brand of hardware to choose from. We’d have high prices because Apple is very famous for high prices. We might still have black and white computers for all we know. Not that I have anything against Apple per se, but those are the alternatives out there; not some world where we have hundreds and hundreds of operating systems and everyone is dancing around with flowers all over the place. That’s not the way these markets work.

Now there are some remedies, I can come up with some novel remedies as well. This is the old version somehow. One thing would be that I would say that Microsoft executives should not be allowed to write E-mail unless they take a course in both antitrust and charm school. I also think, and this is a little bit more serious but still fairly facetious, that the price of Windows should be automatically set to the difference between a Sun work station, a low-end work station, and an equivalent powered Intel machine without an operating system. All right. And then after you saw the high price you would have to pay for that you would be thrilled to get Windows back at the old price. By the way, we’re talking about forty dollars, forty-five dollars is the price of Windows. We’re not talking about an enormous amount of money.

One of the authors of the book that is out there because that’s a book from a conference, is named Ben Klein. He was also on my dissertation committee. He didn’t know anything about computers before you guys spoke to him about coming to the conference. And he says well, what are they talking about anyway? And I said oh, you know, Windows. And he says so how much is that? I said well, the cost of Windows is about forty-five dollars. He started laughing hysterically. He said forty-five dollars? That’s what they’re upset about? That’s a monopoly price?

At any rate, also, I would say Microsoft would have to make its operating system to at least as many OEMs as Apple and Sun, a weighted average of those two. Which of course, is zero. And that they have to give the same discretion over screen icons to their OEMs that Sun and Apple need to give to theirs. Which of course, is none. Because there are no OEMs.

The alternatives to the current system are not likely to be a world that’s a whole lot more competitive. Probably a world that’s less competitive, and with higher prices. And as I said before, the real reason that this occurs is because it’s, I believe, the politics of this case is that it’s due to another market entirely and the Justice Department got pulled into something which I’m sure they’re very happy about now because it’ll be great for the careers of all the lawyers involved. And certainly I’m very impressed with David Boyce. He’s a fabulous lawyer. I would hire that guy in a second. But it’s a bad case for consumers.

And that’s what everyone’s forgetting. That’s the biggest weakness in the Judge’s finding of fact. There’s no harm that they’re going to be able to find. It’s a good thing for the states that they weren’t able to get their case going because their case was going to bring in applications. And if you bring in applications you have that chart that I showed you and that chart shows prices did not go up, they went down. Even though Microsoft is not always the lowest one, but it’s whose watch are we under?

The book has detailed a lot more cases than that. Lots of details as to what’s going on. If you really want to know what’s going on in those markets I suggest you read the book.

The last page of the handout is just for those of you who like to reminisce about the good old days. And it’s from a 1986 copy, I think, of Creative Computing. So you can see what software prices were like when it was three hundred dollars for a word processor, three hundred dollars for a spreadsheet, as opposed to buying an entire office suite for two hundred something or three hundred dollars. Things were a lot more expensive in those days.

The remedies that have been proposed, this three integrated Baby Bills is whether you just do it with the operating system or whether you say that the Baby Bills all produce the entire set of software, whether it’s applications and operating systems and whatnot, would be very expensive. It would be very tough on consumers because now you have to decide do you want to buy Windows A, Windows B or Windows C. They have to become incompatible with each other if they’re really going to be separate products. And so there’s confusion on the part of consumers. Developers have to decide are they going to develop for A, B or C. If they decide to develop for all three, that’s where the costs come from in that first page of the charts I gave you. They’re very high costs for no benefits that we can really measure.

Except that we’re not in the ideal world of competition. That’s really the cost that the judge has pointed to.

If we had three specialized Baby Bills, which is not one that was suggested just now, where you have one application, one operating system and one Internet, which has been suggested by others, then you have three monopolists instead of one. And it’s not clear how that benefits anybody. And the reason that Sun and Oracle have been the most vociferous in their speaking, they want to break it up three ways. The way I first said. Because that’s the most costly. What they really want to do is drive a dagger through Microsoft’s heart. They want to impose costs that are as high as possible on Microsoft.

And the way you do that is to break it up vertically and then say well, person A goes here, person B goes here, person C goes there. It’ll take a very hard time for Microsoft to recover from that. It will take a very long time to set that up. And while all that’s going on they continue to clean up in the high-end server market and the work station market. And that’s what they really want.

There’s been some claims that we should auction Windows. That doesn’t seem to make a whole lot of sense to me. Then we’re just going to have two versions. Who’s going to pay for it? The minute it’s auctioned off what happens to it? I don’t know whether that does anything. And I think, yes, then there are some behavioral remedies, one of which I actually think is a good idea. And I think Microsoft should voluntarily agree to it.

And that’s to publish a pricelist for Windows and stick to it and not give any deviations. Only because it’s such a headache. It sounds like they do all this crazy stuff. And it sounds like they’re already engaging in price discrimination if it’s true, and I don’t know why that charge wasn’t brought up against them by these companies that they’re supposedly discriminating against. But if they published a pricelist, if they were forced to do that and charged people what it said, I think that would be a good idea. That would fix what I see as the only behavior here that has some evidence behind it, and which may be illegal. And so I’m all in favor of that.

But the idea of not being allowed to add items to the operating system I think is a terrible idea. All right. That’s just, it’s hard to enforce, it’s hard to know how to monitor. And the operating system should have things added to it. They do have things added to it. They almost always, whether we’re talking about Apple or the Windows, whenever they add something new it’s almost always competing with somebody who was producing something individually before. And it’s almost unavoidable. Because there’s a market opening, someone sees that people want it, they start producing it, then it’s something that goes in the operating system and then those people may be out of business. But that’s something that, that’s the way the world works. And it’s useful to have it.

Let me give you an example. Voice recognition. I’m not out of time yet? I am. Okay. Voice recognition is something that all programs ought to be able to have access to so that the creators of the programs don’t have to write the voice recognition themselves; they just call the voice recognition module that’s in there. It’s clearly easiest to do that if everyone knows what the module is, and if it’s in the operating system. So that’s something that should be in the operating system. I think most people would probably agree, when it gets good.

But there are companies out there selling voice recognition products right now. So if Microsoft then tries to put it in, we enter this realm again of gee, they’re going to be putting these guys out of business. And I bet you those voice recognition guys have some APIs that they have there, that people would be able to write to to use their products. And so it’s this Netscape thing all over again. Those were not APIs that anybody was going to be able to write an entire program to run on. You weren’t going to be able to run, and the judge even admits it, the APIs for Java don’t allow you to write serious applications.

The reason nobody writes applications on Netscape and Java in any significant number is because nobody wants those applications. There are fifty percent of the people, almost, who have Netscape on the IBM right now. And then there’s a whole bunch of them on the Macintosh. And then there’s a whole bunch of them in the Unix world. Surely that’s a big enough market that if somebody wanted to use one of those programs for Netscape’s APIs the writer of those programs would have a plenty big audience. Plenty bigger than the Macintosh market which has twelve thousand programs which the judge tells us is insufficiently small to make it a viable competitor.

Now I always thought that if I didn’t want to buy a Windows machine I could just go buy a Macintosh, and it wasn’t such a terribly hard thing to do. But apparently that’s not true. Macintosh is not part of the market definition. In fact, according to the judge Macintosh has a monopoly. By itself it has one hundred percent of the market and it therefore should be liable to the same set of rules of antitrust as Microsoft because they don’t compete with each other. Microsoft only competes with itself. Apple only competes with itself. Sun only competes with itself, maybe the other Unix boxes. We’ll have lots of monopolies out there according to this ruling.

At any rate, I think that pretty much finishes up where I am, and as you can tell, I think we should just tell Bill Gates to be a little bit sweeter and tell his guys to be a little bit nicer, and slap around a bit, and then we’ll actually all be better off. Thank you. [Applause]

Thomas M. Lenard

Well, I’m glad we finally got Stan to say that Microsoft doesn’t have a monopoly. I’m going to have to be a little selective. Let me just focus a little bit on the empirical stuff that Stan has presented, because I have looked over some of that stuff, and not in enough detail to really check the numbers. The thing that kind of comes out at you, he’s come out with two numerical estimates. One is the costs of fragmentation, like it’s thirty billion dollars, and the other is the fact that markets where Microsoft competes have lower prices.

The striking thing about both of these analyses is that, you know, wherever the numbers come from, they really don’t take into account the economic logic that kind of analyze [Phonetic] these markets. If you take the proposal that I was talking about, about the kind of the hybrid Baby Bill solution where you separate off the applications and from the operating system, and then you divide the operating system into three competing companies, what Stan was talking about, there would be large porting costs to write for all those three supposedly incompatible versions of Windows. But obviously on day one there’s no porting costs.

And then as you go on, what you’d have, I mean, the entire argument, when I read it the first time I thought was, you know, it kind of brought me back to kind of Galbraithian economics which I thought I had left in my economics one course where it said well, why do we have more than one brand of soap. You know, it’s wasteful to have more than one brand of soap. You know, they advertise against each other, they market against each other. I mean, you know, soap is soap; we only need one brand. Well, I mean, that’s kind of antithetical to the premise of competitive economy, it’s antithetical to the premise of the antitrust laws. And it ignores, in this particular case, the incentives on the part of the participants in this market.

Because each new company would start out by sharing a large installed base of users who would have a very strong incentive to maintain compatibility both with the existing stock of programs, backward compatibility, and with each other. I mean, consumers want to have access to a large supply of programs. Program developers want to have access to a large supply of consumers, okay, and if you’re one of these three companies and you decide to make yourself incompatible with everybody else, that’s a pretty risky decision. You’re going to risk losing one or both of those things. So there’s really very strong incentives to maintain compatibility and compete essentially within the Windows standards.

Now obviously over time you may not have one hundred percent perfect, complete compatibility, and there may be some porting costs, but you would also have a lot of competition and a lot of improvement in the quality of the product and these porting costs which, I mean, Stan’s estimate of the porting costs, and I really didn’t get into the numbers to be able to evaluate it but it seems to me that it assumes three entirely incompatible operating systems. And you know, there’s nothing in the economic logic of the situation to think that that would develop because there are very strong incentives to maintain that compatibility.

Now, with respect to the price. The price of software being lower in the markets that Microsoft has entered, that also ignores the economics of the software industry. The economics of the software industry, first you have these large network effects, okay. Secondly, software is a declining cost industry. There are large fixed costs upfront to develop the software, to produce the so-called first copy costs, and then after that, the marginal costs of distribution and marketing are very small.

And so Microsoft being a smart company, understood all of that, okay, so they went into rapidly expanding markets, okay. They tried to make those markets expand even more rapidly than anybody other has by pricing their products competitively and so they’re in very large volume markets where the average cost of the software is very low. So the causation, I mean, to say that the causation of these low prices is because Microsoft entered the market just misses the point of the economics of those markets.

Another point that Stan made is that the benefit of this whole thing is that Microsoft’s competitors, the Suns and the Oracles of the world are concerned not about Windows, but about NT. Obviously Windows and NT are converging basically to Windows. And I’m sure that those companies are concerned about the potential for Microsoft to monopolize the NT market on the server end, and then strengthen its position in the enterprise market, perhaps do a similar thing there that they did in the applications market for Windows.

But the proposal that I’m promoting would essentially create several new competitors for the Suns and the Oracles of the world. So that, that can, it’s hard to, so you have that would be an unhealthy thing for the marketplace.

I don’t know, Stan said a lot of things and he started out by saying well, this behavior was really, you know, may have been in bad taste but really didn’t hurt the market. I mean, the fact of the matter is I think that nobody really denies the fact that Microsoft was successful in shutting off the avenues by which this browser was distributed, and basically closing it down to a competitor. Now, you can engage in exclusive arrangements of that sort if you’re in a competitive market, but the rules are different for monopolists. I mean, they always have been. And those things are just, because even if they get the monopoly fair and square we don’t want them to be using these types of practices in order to be able to maintain it forever, you know, even in the face of competition from a potentially better product.

Questions and Answers

Peter Huber

Well, we have a few minutes for questions or comments from the audience. Any takers? And perhaps if you want to direct them at one or another of the speakers.


Either one. Is there any chance that something good could come out of this in that public opinion would be increasingly in favor of Microsoft [inaudible] the government and you’d have no more of these cases brought for twenty years.

[Transcriber’s note: Speakers in audience are off-mike]

Stan J. Liebowitz

Well, I think if you adopted my rule that the price that consumers have to pay for the operating system is essentially the price that is being paid for Sun’s operating system, that there be enough of an outcry that that might happen. The reality is that these things tend to move slowly, the consumers have a hard time knowing exactly what’s happening. I don’t know that they’d be so outraged that they would sort of march on Washington and sort of try to destroy the Justice Department or anything.

Thomas M. Lenard

I mean, I was actually interested that this issue has risen to the level of being a Presidential debate question, in one of the recent Presidential debates. I’ve heard conflicting evidence on what public opinion is. I see some polls that say Microsoft is very unpopular, some polls that say the government is unpopular in this, so I don’t really know.


The Microsoft [inaudible] I think it’s more than just a hassle, as you were saying, to [inaudible] computer, but isn’t Judge Posner a pretty well-respected economic judge, isn’t his opinion likely to [inaudible] the case one way or the other, do you think?

Stan J. Liebowitz

Posner as the arbitrator in the settlement talks between the two parties is likely to bring about a settlement.

Thomas M. Lenard

I mean, I claim no inside information, but the people, whenever I talk to people who may be slightly closer to it than I am, they say they think the parties are pretty far apart, so I don’t know.

Stan J. Liebowitz

One of the things I don’t understand is exactly how much discretion or power Judge Posner has in terms of when he took the position was he actually given any sort of impact on the case? And I don’t know. Because just being there as an arbitrator, I don’t know what he can make them do. I think if he had some power over them they’re like to come to a solution. Certainly if the person who wins is the one who’s smartest, he’ll Windows. But offhand, I don’t know what that does to the chances of a settlement.


Mr. Lenard, you said, I believe, Microsoft took active measures to prevent specific innovations to occur. Which one is genuine?

Thomas M. Lenard

Well, I mean, for example, Intel was developing software to give directly to the OEMs and put it directly in the chips. That’s one. And I think there’s lots of, I mean, this whole effect of innovation is very hard to quantify. There’s lots of, I mean, if you have a market that is so, you know, influenced by one firm and potentially distorted by one firm it’s hard to, you know, there’s lots, I mean, it’s obviously hard to quantify the lack of innovation. I think it’s reasonable to assume that if you had more competition you’d have more, if you had to be more, if your customers had leverage, that you would have more innovation. Because that’s the basic logic of competitive markets.

Stan J. Liebowitz

Yes, that one particular case is a bit ironic, because if you think about it, what Intel was thinking about doing is exactly what Microsoft was doing in the way it was adding in functionality to its chip. That was going to eliminate the need for some other producers of some other chips. So in a way it was like adding something to its operating system that was going to put a competitor sort out of business. In this case, people who were making DSP chips which they were adding the functionality in.

Now, it’s unclear to me exactly what the reason was that Intel stopped doing what they were doing. It is certainly true that Microsoft appears to have told them that they didn’t like them doing that. But it’s not clear to me that Intel is going to sit, you know, down and, you know, kiss Microsoft’s feet and do what Microsoft asks it to do. My guess is they did it out of their own self-interest. But I don’t know for sure.

Thomas M. Lenard

Well, a lot of, I mean, if you read the 207 pages of the judge’s findings there’s a lot of people deferring to Microsoft out of their own self-interest.


You made a statement that there was no harm to consumers basically, but it seems to me that the consumer was harmed in being deprived of the opportunity to choose between Netscape and the Microsoft browser. I mean, they just didn’t have the chance to make that choice. You obviously think, by your saying that Microsoft was a better product, I’m agnostic on that, but clearly, the consumer never had the choice of Microsoft [unintelligible]

And the other question is, that I was intrigued by shortly after the findings of fact were published, and The Wall Street Journal published some responses of people who thought, their opinions of what the remedies should be, and one of them was the chief executive of Novell who said that what you said was that Microsoft should have a uniform posted price for Windows. And the other was that they should provide equal access to the applications interface. And I was wondering why you didn’t add that one to your suggestions.

Stan J. Liebowitz

Well, for one thing Microsoft says that they do give everybody equal access to the APIs, as they’re called.


[inaudible] IBM?

Stan J. Liebowitz

Yes. That’s public information, what the APIs are. And there’s a debate about whether that’s true or not.


I understand it was delayed in the case of IBM.

Stan J. Liebowitz

That was, they hadn’t, the new version of Windows coming out apparently was delayed because they hadn’t reached an agreement about what price they were going to charge and whether they were going to give it to them or not. And they did give it to IBM. I don’t know exactly how much of a delay there was or how much of a problem. They usually give out pre-release versions and I’d have to go into that. But your earlier question-


I’d be interested in Mr. Lenard’s comments on the punishment of [inaudible]

Stan J. Liebowitz

Tom, go second. He may know more about that. The first thing that you said was what I wanted to respond to most. Could you remind me what that was, I’m sorry. [Laughter] Oh, yes. That in fact isn’t the case. Microsoft told OEMs, the computer manufacturers, that they could not remove Internet Explorer, but they didn’t say that they could not put Netscape on there. So if they wanted they could have put both browsers on there.

And another little historical fact was that they included Internet Explorer with the early versions. The early versions of Internet Explorer with Windows as well. And no one used Internet Explorer in those days. It was only when Internet Explorer reached parity in the product reviews that people started using it, and then later on. Because one of the things we do in the book is we take a look at product reviews and see how they relate to market acceptance, both in markets where Microsoft is competing and where Microsoft is not. And what we find it that when the product review say a product is better, its market share starts going up very quickly. And that’s what happened in this market.

Now Tom’s interpretation is it’s because Netscape wasn’t allowed to get their product out there. Our interpretation would be well, at that point Internet Explorer was getting better reviews and it followed all the other markets we looked at which meant its market share went up a lot.

Thomas M. Lenard

Well, I mean, I think there’s so much evidence in the record that Microsoft felt that they, you know, notwithstanding the fact that they did a lot to improve Internet Explorer, that they just couldn’t do what they wanted to do on the merits, okay? They weren’t going to accomplish their goals on the merits and they had to shut down these avenues of supply if they were going to, you know, accomplish their goals.

I guess what I’m surprised about is why Stan endorses these kind of little conduct remedies like opening up the APIs and nondiscriminatory pricing. If there’s no problem, why do that? I mean, those are just, and then if you’re going to do that, why not do the other conduct remedies that people talk about. Because those are just two out of a list, you know, that you would get to if you were going to go the conduct remedy route.

The attractiveness of the approach that I’m talking about is you don’t have to worry about all these things. You don’t have to police whether the APIs are being public, whether they’re secret APIs, whether they’re being released, you know, to one party before they’re being released to another party. You don’t have to, you know, decide whether price discount to this manufacturer is justified and this one isn’t. You just basically create three companies and let them go at it as hard as they want.


I’d like you to say something, Peter, about this topic. [Laughter] And figure it all out.

Peter Huber

Right. Does Judge Posner settle it? No. Does Microsoft get broken up? No. Do they get a consent decree? Yes. And it won’t be as bad as a lot of people think.

Thank you very much. [Applause]


  • Catalyst
  • Beyond Homeless