The Power of Independent Thinking


Stay Connected
Get the latest updates straight to your inbox.

Technology Innovation
April 16, 1998
Timothy F. Bresnahan, Stan J. Liebowitz, John E. Lopatka, Stephen E. Margolis, Janusz A. Ordover


Introductory Remarks by David Theroux

My name is David Theroux and I am the President of The Independent Institute. I would like to welcome you to our breakfast program today on the subject of the economic and legal issues concerning antitrust, monopoly and competitive markets, especially as they pertain to high-technology innovation.

For those of you not familiar with The Independent Institute, we were founded over ten years ago as a non-profit, non-politicized, non-partisan, scholarly public policy research institute. We annually produce numerous, peer-reviewed books, our quarterly journal, The Independent Review, and other publications, and we conduct various conference and media projects based on the results of this work. Currently, we have about 120 research fellows at different universities involved in Independent Institute research projects ranging from such issues as crime, health care, international finance, urban problems, taxes, and many other topics.

The Institute is funded by over 800 supporters, including foundations, individuals, businesses, unions, and other organizations. We don’t do consulting work for private or government institutions; we don’t do contract work—in short, we’re not a consulting firm. Instead of focusing on issues as they are temporally debated in Congress or in other public forums, we concentrate on developing a more in- depth analysis of major economic and social issues based on the cause-and-effect nature of public policies. The conclusions of such work must be based solely on the integrity of the scholarship involved and hence the peer-review process to ensure this outcome.

Our involvement in the issues of competitive markets and antitrust goes back to our inception and includes our classic book, Antitrust and Monopoly: The Anatomy of a Policy Failure , by the economist D. T. Armentano. Beginning in the late 1980s, a number of economists became interested in the topic of “network externalities.” Two of them are here today: Stan Liebowitz and Stephen Margolis, who began work that led to the publication of numerous articles on high-technology markets, competition and innovation in various technical economic journals, with the first article appearing in 1990 in the Journal of Law and Economics . They are also completing an important new book for the Institute on these matters, Winners, Losers & Microsoft. This work was part of a growing scientific research process on the theory of “network externalities” that began about ten years ago and has also included such economists as Brian Arthur, Carl Shapiro, Michael Katz, Garth Saloner and many others.

Several years ago, Upside magazine then commissioned us to produce an article for them by Stan and Steve (which you have in your packet) in order to bring their analysis to a broader, more popular market. In addition, we have sponsored articles in such publications as the Wall Street Journal, Christian Science Monitor, Investor’s Business Daily, San Francisco Chronicle and many other newspapers and magazines.

Since we began this work many years ago, these issues have now become centerpieces in the raging debate over high technology and government policy. As a result, only now are we beginning to attract the attention of such firms as Microsoft, Netscape, Intel, and others. And, we would respectfully welcome their participation as Institute members in order to utilize the results of our work, as we have to firms ranging from the Washington Post Company to Freedom Communications. And as we would to any of you here today.

As you may well know, “network externalities” pertain to the potential harm to consumers from different high-technology business behavior and market phenomena, and whether antitrust in some form or other regulations are appropriate to create a more fair, competitive system.

Our panel today consists of five excellent scholars. Each will make a presentation of about 12 minutes or so. And then at the end of all five presentations, we’d open the discussion up to Q&A.

Please note that our panel represents only a portion of the economists and legal scholars in the academic world involved in this topic. For example, unfortunately, Brian Arthur and Garth Saloner were unable to adjust their schedules to join the panel. So the first person I wish to introduce is Professor Stephen Margolis from North Carolina State University.

Stephen Margolis

Thanks, David. It’s a pleasure to be here. The Independent Institute has been a contributor to the debate on antitrust for quite sometime and their involvement pre-dates any particular issues and also pre-dates the participation any of the current interested parties.

The work I’m going to talk about today I’ve entitled “Lock-in to Inferior Technologies: A Proper New Foundation for Antitrust?”. At issue is whether “lock-in” is a problem—by which we mean “lock-in” to things that you would rather not be locked into. This is work that Stan Liebowitz and I have been doing for about a decade. The issue that has been raised is: Does lock-in provide a kind of new reason for antitrust authorities to be aggressive in monitoring our current business practices?

Some theoretical writings in economics argue that it’s likely that free-market choices of products will result in lock-in to inferior technologies. However, the body of work that Stan and I have done make two points: First, theoretical arguments that support this are severe abstractions. That’s fair enough—most modeling is abstract, most economic modeling has to be. But nevertheless they don’t prove themselves. The abstractions that we make in these models require that we look at the real world and see whether in fact we’ve captured something fundamentally important about the world in building these models, and so you have to look at the world to see what works, to see if these models are in fact telling us something important.

Our second point (and this is actually where we started publishing things), is that the empirical evidence that has been offered in support of lock-in is incorrect. There are no clear demonstration of cases in which we are locked into things that are demonstrably inferior. So we’ll get into each of these two claims briefly.

Here are some special terms that will show up in my talk and some other talks today and some vocabulary you’ll be encountering in this debate: increasing returns economics, network externalities and network effects, path dependence and lock-in.

The idea of increasing returns in economics can be simply stated as: bigger is better. If you double the size of a firm—double everything about it—you’ll more than double its output. That’s increasing returns at the firm scale. With regard to new technologies, it may be that the payoffs for participating in the new technologies increase as the volume of activity in that technology increases. For example, if there are only a few fax machines, they’re not very useful; when there are lots of them, they’re very useful. Internet participation becomes more valuable to each of us as the Internet expands.

This gives us “network externalities”—a closely related topic. Our fax example, in fact, is a “network externalities” example of increasing returns economics. We have argued for calling these “network effects.” That is, you are better off when more people have telephone service, provided you have telephone service. But “externality” has special meaning in economics it concludes a kind of market failure and we have argued that we should not prejudge that case. Just because you and I might affect each other, that may or may not be an externality.

“Path dependence” is the idea that a head start, the first technology, the first manifestation of the technology on the scene, dictates what might ever occur. I’ll get into that a little bit.

And then finally “lock-in” is the special case of being trapped in some sort of inferior technology. Let me give you a simple finger exercise of lock-in. This is one of the cases that’s been alleged and the way I’m going to tell the story doesn’t actually fit the facts but it’s a simple of way of thinking this through. Most everyone here remembers the Beta/VHS story. We all went shopping for videocassette recorders and we had to decide whether we wanted to buy a Beta machine or a VHS machine. We each went to the store armed only with our own tastes and perhaps information about the market shares of Beta and VHS. Here comes the lock-in part of the story. It could happen that we each went to the store, played with each of the machines, liked Beta better, but then said to ourselves, “I hear that everybody else is buying VHS, so I’d better buy VHS.” So we each buy VHS and after the fact we’re feeling smug because we predicted correctly and we didn’t get stuck with one of these orphan machines. But in fact we are worse off because if we all did prefer the Beta machine, we could have been better off if we somehow had managed to coordinate our actions. This would be lock-in.

Well, the question then becomes: what’s wrong with this story? What’s left out of this story? It captures what is in some of the theoretical arguments that give us lock-in. What are these models abstracting from, or at least what is this story abstracting from? And you can kind of get to it on your own. If we all thought Beta was better, if it were so obvious that Beta was better, wouldn’t we suspect that other people were buying the Beta machines? Or if it where so obvious to each of us that Beta was better, wouldn’t magazines like Consumer Reports , and audiophile and videophile magazines be telling us that Beta is the thing to get, that everybody is going to be getting Beta? If so, we wouldn’t go to the store expecting that if you bought a Beta machine you’d end up with an orphan.

So what’s left out of the story? Communication among people; anticipation among people; consumer magazines; entrepreneurs who could provide money-back guarantees; leasing perhaps, for technologies that are more expensive (if the product turns out to be an orphan, we’d just terminate the lease or not renew it); discounts, if Beta where better and you needed to get a toehold in the market; product giveaways, to establish the format and so on. So there are lots of real world behaviors that need to be taken account of in forecasting which way these outcomes would go.

This ultimately leaves us with an empirical question: What empirical support is there for this idea that we get stuck with the wrong technologies, or we get stuck with the wrong formats? And this is where a good bit of our writing has been.

There is a famous story in the literature; you’ve probably heard some version of this, although perhaps not our version. That’s the contest between the QWERTY keyboard and the Dvorak keyboard.

Here’s just a very quick outline of that story: In 1867 Christopher Latham Sholes gets a patent on the typewriter. It wasn’t the first patent, but it is related to the first modern typewriter. One of the problems designers faced in getting the first typewriters to work was that the keys tended to jam. You would type two key simultaneously and the hammers would run into each other and jam up. The designers approached this problem in various ways, including experimentation with configurations of the typewriter keyboard arrangement itself. One of the allegations in the story as it is often told is that designers were trying to slow down touch-typists and so they adopted the QWERTY keyboard arrangement. (“QWERTY” is from the letters in the first row.) Most everybody is probably familiar with this.

Then, the story continues, in the 1930s August Dvorak developed a keyboard that used scientific approaches to human engineering, was much more efficient and you could learn it faster, and once you learned it you could type faster. There are various claims going around that the Dvorak keyboard is 20% better, 40% better even 80% better. (I recently read somewhere that said it was infinitely better!) There was a study attributed to the Navy that is sometimes quoted, saying that every 10 days after you retrained in Dvorak, you would recover the cost of retraining. If that were true, it would be a phenomenal rate of return on investment.

Well, Stan and I where working on the general economics of standards and we bought this story. We factored it in to some modeling we were doing and then it dawned on us that we really should dig in and find out how much better the Dvorak keyboard was. And as we dug in we found that there was no credible evidence that the Dvorak keyboard was any better. A lot of the claims that are around in the literature are from Dvorak’s own writings.

But one study done in 1956 by a professor in ergonomics from Penn State named Earl Strong done for the GSA was very careful and concluded that you would never recover the investment in retraining in the Dvorak keyboard. There have been a lot of ergonomics studies of this since then and they conclude that Dvorak offers little or no advantage over QWERTY that keyboard.

After a time, we also discovered that the Navy study was done during World War II and was supervised by a Lieutenant Commander August Dvorak, who was working for the Navy at the time. This was not published as a Navy study; it’s not in the Library of Congress or the Navy archive. We found it in a private collection, it was stamped as a war time materials that were declassified, but there was no other official recognition of this document.

Why is this important? Because this example is used over and over again to demonstrate the possibility of lock-in. The fact that it is used over and over again—and has continued to be used since 1990 when we published our piece showing that there was no advantage to Dvorak—says something about the abundance of other examples of this sort of thing.

Now there are a few others. There’s the Beta/VHS story, but Beta had a two-year head start on VHS. If lock-in were important, VHS—which offered some advantages over Beta format—would not have been able to displace Beta. In fact, VHS displaced Beta very quickly. Consumers preferred the fact that VHS allowed a two-hour taping time while the Beta format only allowed a one-hour taping time.

There are a couple of other examples. There is the Macintosh/DOS story. The end of that story is: we have the Macintosh standard today. It took us a while and it is a different company, but in fact there are standards that can evolve. And when I say we have the Macintosh standard, we I don’t mean that we buy Apples. We buy all kinds of IBM clones and we buy operating systems from Microsoft. But we use the graphical user interface, which was the innovation of the Macintosh operating system. (Actually “innovation” may be the wrong word; Apple took it from Xerox.) That main feature of the Macintosh operating system is something we have today, and operating systems did evolve to include that. We can be disappointed if we’re fans of Apple the company; but the actual standard that we used did evolve to what in fact is advantageous.

Finally, what we should note that when DOS was important, DOS did have some advantages. DOS didn’t gobble up memory and processing power the way that Apple machines did, back at a time when memory was very expensive.

So our conclusion then is that the theoretical arguments for lock-in leave out important features of the real world that may tend to overcome lock-in. And finally, the empirical evidence that’s been offered in support of lock-in is incorrect. And we would say therefore that theories of lock-in to inferior technologies are an unsound foundation for a new aggressiveness in antitrust enforcement.

David Theroux

Thank you, Steve. Our next speaker is Professor Timothy Bresnahan. Professor Bresnahan is Professor of Economics at Stanford University and he’s also the Director of the Stanford Computer Industry Project.

Timothy Bresnahan

Thanks, David. There are two open debates here and I’m on the same side in the policy debate. That is to say, over the last 7 years I have routinely counseled antitrust enforcement agencies not to pursue Microsoft. There’s also a positive debate about whether or not you should use this network-externalities/lock-in theory to understand the computer business and especially the software business. And in that debate I’m on the other side. I’ve written a series of histories of the computer business which emphasize the use of that theory as the basic background story one uses to understand competitive strategy and technological evolution in the computer business.

I can summarize the schizophrenia on my part by my reaction to the first antitrust case that the FTC was considering bring against Microsoft. One of my really good friends was a FTC commissioner and he called me up and said “Should we go after Microsoft?” I said, “Well, you know you’ll win and you aren’t going to be able to think of much of a remedy that you should really want.” Since it was a friend I went on to say, “You know, they’re a great big, red, noisy fire truck running down the middle of the street, and you’re a dog. The issue is: what are you going to do with it when you catch it?”

Now, let me take the other side of the lock-in debate. I think that Microsoft indeed does have a valuable monopoly on the client side of the computing business and that they exploit and attempt to extend that valuable monopoly very aggressively. But I’m not sure that there’s an antitrust remedy which can do anything useful to change the future structure of the computer business to one where there’s not a client-side monopoly, or where there is a client-side monopoly but one that doesn’t extend it’s influence to the rest of the business. Let me go into that in a little bit more detail.

The key thing about these lock-in stories is this. Microsoft invests in the Windows platform and in the associated applications that run in it; so do we all. We all invest in that too. We invest our time and talent, and in training collaborators. Other information technology companies write for that platform, so there’s tremendous leverage for Microsoft because they have everybody else making investments right along with them in things that only work with their products.

Now, they’re not the only people who do that. Plenty of other information technology companies are trying to get in that same leverage game. WordPerfect, a former network-externality-based monopoly in the word processing business, did exactly that. Legal secretaries all over country had invested substantial amounts of their own training and capital in working with that particular product.

Now, here comes the part where I think the network-externality story is true, but it’s not necessarily bad. What happened to that WordPerfect monopoly in word processing? It got taken out by a competitor—Microsoft in this particular case. If there’s somebody nearby that’s got some muscle, some technological capability, they’re a potential entrant to breakup those monopolies and to move us forward into a new arena. A new arena will probably involve—as it does with the office suite—substantial amounts of network externality and substantial amounts of lock-in. The point is that there are some big agents out there who might well be in a position to break that and move us forward.

For example, people say we should make much of what Microsoft does—exclusive contracts they like to have, pressuring of their compliments and things like that—more illegal because they have a monopoly. That strikes me as a bad use of one of the antitrust shortcuts. In this business, you’re suppose to have a monopoly, and the example I always use is Netscape Communications Company and the browser. What were they trying to do? They were trying to establish another network-externalities- based monopoly, one layer over from Microsoft. Here is one of those terrible contradictions: I think the industry would be more competitive if Netscape had established a monopoly. This is an industry in which the important intermediate-run technological competition is between competing providers of these local monopolies in particular segments of the industry. So the shortcut which says that everything should be more illegal—that lots of business practices should be more illegal because there’s monopoly power—seems to me to be a mistake.

What should you think about when you think about the long-term structure of the computer industry and it’s changes? The real antitrust policy problem in my mind is that we might be right at the end of the interregnum between IBM and Microsoft. It’s really good that the structure of the industry now is vertically disintegrated and that there are plenty of outside-change agents to go after everyone of these local monopolies. Lately, the same one has tended to win a lot. The IBM PC used to be a valuable monopoly. They got taken out by—guess who?—WordPerfect, which was taken out by—guess who?—Lotus Development Corporation (probably a suicide rather than a murder victim), which was taken out by—guess who? So there’s this reconstruction of the IBM-like structure of the industry, in which there’s one firm that has control of a lot of these areas, and the long-run tendency might be towards an industry structure which you don’t like, where there aren’t outside-change agents to go after each individual monopoly. That’s sort of the theory you hear from the interventionist in the valley. That we’re re-converging; it’s sort of the last change to prevent reemergence of a vertically integrated monopoly on the old model.

The thing which counters that in my view—and which makes me cautious about antitrust intervention—is that out of existing technologies we’ve seen a considerable amount of re-integration. We now see the office suite provided by the same company that provides us the client-side operating system. But even though we see that, there are still some very substantial surprises which gives entrants into the business and opportunity. In the early- to mid-nineties, it looked like we were converging towards that stable structure—at least on the PC side. And along comes the Internet. All kinds of good things happen. I made a 50% capital gain on my house. Lots of people entered the computer business in new layers, a gigantic injection of vertical disintegration takes place, and many new companies are created that have substantial muscle and that are potential entrants into all of those existing monopolized markets. It just seems to me that we have a picture in which there is vertical competition, competition between monopolists trying to take each other’s monopoly out, and new monopolies—good ones—keep getting injected into the picture by fundamental technical progress, such as the commercialization. not the invention of, the Internet.

Now comes the really hard part for the interventionists. First, let me say why the antitrust authorities along with me would be tempted to go after Microsoft. There are competing technical standards much younger and less well-established than the Microsoft standards, and you might well want to give them a fair crack at the existing Microsoft position—Java and the associated interface standards that go with that. And it would certainly be extremely useful to potential sellers of Java, which might well be a better way to organize invention of IT in the out years, because it will be more open. They might have trouble getting off the ground because of this bottleneck on the client side. So for example, you might want to break the usual rule of antitrust enforcement, which is: you don’t want to protect competitors, you want protect the competition. You might say that antitrust authorities should tilt towards potential entrant Java by, for example, keeping an independent browser layer alive for awhile so that the new standard has a chance to get its act together and get in and take out existing monopolies.

The thing about making a policy decision like that is: you want to either have it be cheap, or you want to be very sure you’re right. And both of those seem to me to be tough calls to make in the present instance. The cheap policy decision would be something like the 1995 consent decree against Microsoft; make a lot of the stuff they do that is obviously bad or pretty bad more illegal that it is. And, I submit, that would be exactly as efficacious in changing the future structure of the computer industry as the 1995 decree has been in the last three years. Remember, the purpose of that decree is to give OS/2, the IBM operating system, a chance to take out Microsoft. I predict that if we do a practices case, it will be to just leave things in place.

The other policy decision: to break them up?! Now you would have to face the question of the last talk—and not in the historical examples where we can go back and find all of those documents, but on a future example where we have no faintest clue—you have to look it right in the eye and say, “I am pretty sure we are going to be right to break them up; that that is going to get us the better industry structure.” And I would be much more comfortable with the current antitrust case if the feds had said that that’s what they are trying to do—if they had established a goal. Thank You.

David Theroux

Thank you, Tim. Our third speaker is Janusz Ordover, Professor of Economics at New York University. He is former Deputy Assistant Attorney General of the Antitrust Division of the U.S. Department of Justice.

Janusz Ordover

Thank you very much. I should disclose that I am consulting for Microsoft at present, and I did in connection with the 1995 consent decree. And I share Tim’s assessments of the enforcement of that time. It was misguided and it had no effect and I am glad that it did not.

I have about four hours worth of slides, so I hope we get lunch here, as well. I’ll try to rush through as much as I can and I’ll try to lay out what I think are the key analytic tools for looking at the competitive developments in the software market and to start by offering a few cautionary notes on the application of antitrust to high technology and software industries in particular. The reason is that, at least from the press reports, I believe that the Justice Department may be heading towards some kind enforcement action or actions against Microsoft and I would caution my friends at the Antitrust Division and support what Tim said—that such an enforcement is likely to do more harm than good.

First of all, I will begin my talk by saying that so far I have seen no evidence that Microsoft has behaved in a manner that’s harmful to competition and in particular to consumers. And I think that unless that is shown, to intervene would be highly undesirable. Let me first also say quickly that I think there are strong reasons why one should not intervene in such markets as software unless there are very overwhelming rationales to do so. I also submit that the Arthurian model [named after economist Brian Arthur, a leading proponent of “lock-in” as a basis for antitrust—ed.] provides a somewhat inadequate foundation for intervention. In any case, let’s just go quickly through the rationales for why one should be cautious.

First of all, high-tech markets move too rapidly and unpredictably for government intervention to succeed in achieving its intended goals, whatever those intended goals may be. Second, the critically important performance of high-tech markets is dangerously vulnerable to the delays and uncertainties that are inevitably introduced by the threat or the fear or the actuality of government intervention. All of us recall the good old days of IBM and AT&T decrees and the kind of uncertainty in the marketplace that was occasioned by fears that the government made about Windows 95, which certainly disrupted the marketplace to some extent. Short-run competition found this more important in high-tech markets than in markets in which antitrust is most frequently applied. One should not look at high tech industry through the lens of models that are applied to steel, to plate glass, or to extractive industries, for example. So there are special features which caution against intervention based on looking at “snap- shot” evidence, which is something the Antitrust Division is good at.

Furthermore, I would like to join in the debate around the Arthurian models as to how this market should look in fact. High-tech products are especially prone in my mind to exhibit extremely significant economies of scale and scope due to high R&D investments and first-copy costs. Most of the development cost is in writing the stuff as oppose to its dissemination.

There are also the learning-by-doing effects which make it important for consumers to stick with what they have learned, or as Margolis and Liebowitz called them “network effects” from the needs for compatibility and protocols. Again, intervention based on static views is likely to be detrimental.

And finally, what I think is a key point which is really not featured prominently: in software markets entrepreneurship is the driving force. These are not layered, management-driven companies in which people believe in sitting on their haunches trying to get rich by waiting. These are markets driven by entrepreneurial spirit. And I think the more we inject antitrust policies into such markets, the more we run the risk of stifling entrepreneurial spirits even in large companies such as Microsoft. I would hate to see that happen, given the critical importance of software markets for the growth of the American economy.

Having said that, I do need to say that I spent two years enforcing antitrust laws, as I thought proper—many times disagreeing with my bosses but sometimes on the same side. The point I want to make is that to the extent that there is a role for antitrust, it has to be in making sure that the next rounds of competition for the market are not stilted and not stifled and not distorted in a way that harms consumers, as opposed to competitors. Remember that the role of antitrust is to protect competition and protect consumers from undue exercise of market or monopoly power. And that is the role that antitrust should maintain and therefore it should try to ensure that the competition for the market is going to be full-fledged. Again, based on what I know, there is been no evidence that Microsoft has engaged in practices that would stilt and distort the competition for the market, which is the kind of competition that is likely to be emerging in these software markets as Tim described.

Let us just quickly look at the key economic features of the software industry. I have listed some that I think are critical. One that I want to especially feature is that antitrust enforcers—and rightly so—are frequently locked into thinking about market share as the critical indicator of monopoly power. But in the software industry that is possibly an erroneous place to start. Market share—even large market share—need not be an indicator of market power. Even if a firm does have a significant market share, it still does not follow—and again this is a very important point—that that market share was obtained by any nefarious or anti-competitive means. Indeed, as Tim pointed out clearly, in software markets it is likely and it is quite natural that the dynamics of the industry creates a circumstance in which one firm or maybe two will be continuously vying for the position of leader.

We are going to see circumstances over and over again in which markets, particularly software categories, are going to have one clear leader and maybe a couple of other firms co-existing, which maybe serve more niche demands because there is still heterogeneity in consumer demands and therefore niche players possibly can hang on and offer the kind of products which consumers desire.

Now, the fact of the matter is that even if a firm has currently a large market share, it does not mean that it is in an invulnerable position. Word processing, spreadsheets, browses, the ability to reach out to the Internet—all of those have been the playing fields of very intense competition between firms in which leaders were displaced. But they might not have been displaced as quickly as this year’s novelty—Beenie Babies—which are likely to be displaced next year by something else. Well, the network effect in Beenie Babies is not as profound or significant as they are in WordPerfect, although they are important. It is important for my daughter to be able to talk about Beenie Babies with her girlfriends. That’s a network effect. Are we going to intervene in the Beenie Baby market because there are network effects? No. These effects maybe slightly more profound in software, but they are not of the sort that would create the danger that somehow the marketplace is going to get locked into any nefarious standard and be invulnerable to competitive assault.

I have also given you a couple of interesting quotes on changes in word-processing and spreadsheets.

What do we conclude from all this? What I conclude, primarily, is that even though the Arthurian models and the models which proceeded Brian Arthur—after all the idea network economies and network effects did not originate with him, path dependence models did not originate with him—these models, however interesting, are not yet sufficiently developed and understood to provide us with a firm foundation on which to build antitrust enforcement in the software industries. Moreover, whatever the theory says, we also have to ask “Where’s the beef?” And that is what this powerful duo [Liebowitz and Margolis] has been doing over the years. Where’s the beef? Is there any empirical support for the key prongs of the lock-in theory? The answer that has already emerged from my quick slides is that there isn’t. The fact of the matter is that the leaders do not always emerge overnight and if we miss enforcement today, we’ll be locked into something tomorrow. It takes a while. (I will have another slide later on pointing out the acceptance of Windows, for example.)

Secondly, category leaders persist—but only if they keep on innovating, if they keep on pushing the envelope on product development. If they do not, they get taken out—as Beenie Babies will be taken out in the future because tastes will change. The same competition—but only profound and more intense—is taking place daily right around where we are talking today. While we are chatting, other people are actually contributing to the social welfare. They are developing new software. They are developing new products aimed precisely at dislodging the existing incumbent, who may be locked in to some extent. I don’t want to paint you a vision of the market which can tip from one product to another over night; it’s not that simple. But these are not markets which are likely to persist locked into what appears to be an inferior technology. And mind you: what is inferior to one person might be quite superior to another. Moreover, as you know, looking back we know how many mistakes we have made. If I had only done 20 years what I know I should have done, I would be somewhere else. I would be Bill Gates. That is the point that comes out.

Furthermore, these models fail to realize that in these markets are agents with substantial interests in steering the marketplace in a direction that benefits them. And it’s this vying for leadership between firms and individuals with interests in steering the market that in general tends to produce outcomes that are in consumers’ best interests, which should be our goal in accessing policies. (I have two minutes and forty slides to go.)

We have a couple of slides that I don’t need to dwell on very much. They try to illustrate to you why I think Microsoft’s success in the market has not been due to allegedly anti-competitive behavior but rather is part of the natural dynamics of the industry. Microsoft has been taking risks in developing new products. It has driven prices down substantially it has—even with its most innovating products, such as Windows—taken quite a while and significant R&D to gain market acceptance. And indeed, there was a period of time between Windows 3.1 and the DOS-based system when the marketplace was quite open for an alternative platform to develop. And OS/2 was one of those hopes. It didn’t go anywhere, but that is not because of Microsoft. It just didn’t go anywhere. In any case, the fact of the matter is that whatever leadership Microsoft may have is essentially, unless proven otherwise and I have seen no evidence to that, due to its continued commitment to investing in R&D and producing better products for the consuming public.

The last set of slides, which I will not have a chance to address in any great length, speaks to the issue of the Internet browsers, which seems to be on the minds of so many people nowadays. The battle between Netscape and Internet Explorer has attracted so attention in the press, on Capitol Hill, and in the building at I used to sit in at the Department of Justice. But the fact of the matter is that the browser controversy has really not been, in my view, featured in the right way.

My next slide is my best description of what I think the world is all about. What the world is all about is the user being able to access information. In the olden days information resided on those horrible floppy disks, which were a great advancement over my scattered notes. We moved down to a hard-disk drive to CD ROMs to Windows for Workgroups and now the Internet. As the sources of the locus of information is expanded, so will the insides of the computer. The operating-system software is going to have to accommodate and be able to provide a user with efficient means of accessing information which is now everywhere. And unless a software provider is able to deliver a product that efficiently enables the consumer to reach out for information wherever it is, it’s not going to maintain its leadership for very long. I believe that the product Windows 98, for example, is going to be a breakthrough in the ability of consumers to efficiently reach out for information wherever it is located—in a uniformed manner and as effectively as possible.

Having said that, I will conclude by saying that based on the evidence I have right now—and I have seen only what is reported in the press—Microsoft has not engaged in any activities that would prevent other browsers from functioning effectively on top of the underlining software which Microsoft has developed. There has been no what I would call “effective” or “willful” foreclosure. But there is a battle for the market of enabling consumers to go out, grab information and process it efficiently and cheaply. And that battle will continue. One hopes the Department of Justice will not distort that battle in a way that harms the American public and the economy. Thank you.

David Theroux

Thank you, Janusz. Our fourth speaker is Professor Stan Liebowitz. He is Associate Dean and Professor of Managerial Economic at the University of Texas at Dallas.

Stan Liebowitz

Thanks, David. I don’t have any slides. I should probably state that I am not working for Microsoft, not getting paid by Microsoft. I have at one time in my life—for a short amount of time in 1995—been a consultant to Microsoft when it appeared that there might be a new case having to do with network effects. That came to an end, I must say, all too soon.

At any rate, there are one or two things I wanted to say before I got into my talk and that is that Tim has mentioned how there are two sides to this issue and how he is on our side with respect to perhaps the empirical or the public policy side, and he is not sure he’s on our side with respect to the theory. I don’t think the distinction in the theories is really quite the way he put. We don’t have any problems with the concept of network effects. We don’t deny that network effects might be important and we don’t think that the claim that the network effects might lead to monopolies is necessarily wrong. We have minor disagreements in the nuances, in terms of whether network effects necessarily drive these things relative to, say, normal production economies, because there are large fixed costs in software. But either way, it might be that you have monopolies.

Where we tend to disagree with, say, Brian Arthur and some of the other people in the network-effect literature is with idea of lock-in. I just want to be very specific about we mean and why it’s different from what you might think. In a way you’re all locked-in to eating and breathing: if you didn’t do it, you wouldn’t be alive very long. But there is no negative connotation. That’s not what we mean by lock-in, and that’s not really what Brian Arthur means by “lock-in.”

Lock-in has a negative connotation usually. And the connotation is that you are locked in to doing something that you really would rather not do. But something, some force, some set of circumstances, has gotten you in the situation that you’re in. That’s what we mean by “lock-in.” We have a paper that refers to three types path dependence. An example of the strong form is the case in which you prefer to use the Beta video recorder but in fact we’re all using VHS. That’s what we mean by lock-in. If some firm gets a monopoly because there are strong network effects but it’s a good product and everyone is happy to use it and it maybe very hard to knock them out, then that may be like eating and breathing. If so, it would make no sense to knock them out. While in some sense they’re locked it, depending on how you want to define it, that’s not how we use the term “lock-in.” So I’m not sure if there is a disagreement between Tim and us on that.

When we looked at a lot of the literature on lock-in, we found a lot of interesting theories but not very well done empirics and very little test of the theories. And a lot of our work has been in trying to test it and concluding, as Steve pointed out, that there’s really no support for the idea that we are going to get stuck with a product that everyone knows is a bad product—a product we’d all like to leave but we can’t. This is lock-in.

On the other hand, when we take a look at Microsoft’s case in particular, we see some of this theory being applied. This is coming out of news reports. I really don’t know what is going on in the minds of the Justice Department. But also more recently we’ve seen minor variation of that and I’ll go into some of those in a few minutes. The overriding theme seemed to be, again, very similar to what we saw with lock-in: a set of theories that in this case aren’t that well developed, with almost no evidence to back up the conclusions that they want to draw.

Let’s go back to the networks effects first. We don’t believe that there is any evidence that we get locked into inferior products. Even if we did—and here I have to admit there is a little redundancy with what Tim said—it’s not clear what we could do about it. The theory says we’re going to get monopolies if you believe network effects are strong. If that were the case and even if we had wrong one, what could we do about it? Assessing damages is one antitrust remedy that is often mentioned. I can’t imagine what imposing financial penalties on the monopolist would do. It would make him less profitable, but it wouldn’t get rid of an inferior technology. You could try breaking up the firm, but then it’s not clear what comes out of that. Is there any reason to believe that the competing remnants of the large firm will somehow have technologies superior to the one the original firm had? Again, you can impose very large costs, but it’s not clear that there’s anything useful that is going to come out of it. You can have a government agency that sits around deciding which is a good technology and which is a bad technology and somehow awarding a franchise—if you can imagine two particular technologies and saying that this is the one that should exist. That’s a possibility and on paper it may not sound too bad if you’re still in college. But I think if you have had any experience with government agencies, you know that by the time they had their first meeting, the market would have already moved on to something else. It’s just hard to imagine that that would be a reasonable type of response. So it’s very unclear what you would do about these things anyway, even if we were wrong and there were lots of lock-ins.

Let us assume that we are right and there are no lock-ins to inferior technology. The theory still says—and we don’t have any problem with this part of the theory—that we are going to have monopolists, perhaps, in a lot of these parts of the market if there are strong network effects and/or strong economies to scale in normal production. Let me just say as an aside, one of the difference is that I’m not convinced that there are network effects in all of these instances. Take the case of browses. As they’ve existed in the last few years, it’s not clear that you have the traditional network effect there. It’s not clear that you really care how many other people use the same browser that you use. And if there is compatibility, where you can go to any web site you want and you can read it whether you have Netscape or Internet Explore, it’s not clear that there are in real network effects there. There may well be increasing returns in that there’s a large fixed cost in creating the product. But it is not clear there are very strong network effects.

So we have this world were we may have lots of monopolies either because of the production costs or the network effects. And what do we want to do here? Could we try to promote competition in this type of world? The answer is: probably no. This is what use to be call natural monopoly. Those are your public utilities, the natural gas company, the electricity company and whatnot. The basic story there was that we knew it would be inefficient—at least we thought it would be inefficient—to have more than one producer. And we thought that if we had a lot of small producers, they’d all have costs that where too high. The analogy to the software market would be that we’d have lots of small, incompatible standards running around, and the users be worse off because they couldn’t coordinate very well between each other. It’s not clear that that gives you any sort of reasonable solution, and it is not clear that we want to have that type of traditional competition.

Economists have a model that we teach in all of our classes. The model of perfect competition has lots and lots of small firms who are all so small, they don’t even know the impact that they have on the market. They take prices as given. That’s a nice model, but it is not clear that you need that or whether that is a good description of a lot of competitions. Tim talked about competition that occurs between large monopolists and different parts of the market. That follows from a long tradition. Schumpeter used to talk about competition—and certainly this will seem very apropos to technological markets—these gales of “creative destruction.” This is the idea that you have some technology, you may have a monopoly on it, but someone out there is working on something else and they are going to just completely wipe you out in a few years. If that is really the nature of competition here, then our standard textbook model of lots of small producers is not the right way to look at the world, and not the type of remedy that we want Justice or anyone else to try to provide. So that is sort of the standard story.

More recently what I have been reading about Microsoft had to do with the innovation in the industry. I am not sure exactly what the theory is, because it is hard to get good information on this from little snippets that appear in the newspaper. But the claim is that somehow there is going to be less innovation than there would have been, if we allow Microsoft to remain a large monopolist in the industry, if we leave them with control of the operating system and let them take over other software markets.

There are a couple of explanations that I can come up with why someone might think that way: The first one is that you might think that there is lock-in. And if that were the case, so be it. As I said, it is not clear what the remedy would be. But we don’t believe that it is likely. So let’s go on to some other alternatives.

One is that you might believe that big monopolists are not very efficient, that they are not very creative, and that they don’t come up with many innovations. That seems to resonate pretty well with a lot of consumers. I am not sure why it sort of resonates with me. It seems like competition should be good for innovation and monopolies should be bad. But the fact of the matter is that this has been investigated to death, mainly in the 1960, ’70s and ’80s by empirical economists who specialize in industrial organization. And there is no clear relationship between the innovativeness of industries and the size of the firms in the industry and how monopolistic the industry might be. So it’s not clear that we have a particularly strong or compelling set of evidence there.

Also, I think that one of the reasons people think that these firms are less innovative is that in general these big firms that they know of the be monopolists are generally regulated monopolists. And we do have some pretty good evidence that regulated firms are not as efficient as unregulated firms. And you might not be surprised if regulated firms were less innovative that non-regulated firms. But no one is talking about regulation here, they’re just talking about the fact that the firms and the industries are monopolistic, perhaps.

Now there is another theory that you might come up with if you wanted to say why might monopolies be less innovative. That has to do with the separation of ownership control.

Big firms often have millions of stockholders and those stockholders don’t necessarily pay attention to what’s going on within the firm from day to day. Small start-ups are run by the people who own the firm, and so they’re very concerned. So a dollar profit to a monopolist is a dollar profit, just as it is to a competitor. And as long as they are interested in maximizing profit and getting that extra dollar, their behavior should be the same with regard to innovation. But if we have these big firms where the managers are not controlled by the stockholders and who are interested in doing things other than maximizing shareholder wealth, in that circumstance you can imagine monopolists not working as hard. And that should include being less innovative.

But the problem of trying to relate a theory like that here is that we have a set of firms that are very large, but we don’t seem to have the separation of ownership control. In fact it’s even stronger than that to some extent. We find that if you look at Microsoft, the largest stockholder (everyone knows his name) owns something like 25% of the company and the other two or three largest stock holders own another fairly large percentage. We have what’s called “concentrated ownership” here. The owners should be very concerned with what the firm is doing—if they were purely stockholders, and the concentration of ownership might be sufficient to the make sure the firm acted in a fairly efficient manner. The fact of the matter is that not only are they big stockholders, they also manage the firm. They are involved in the day to day operations. So it is a much tighter integration than in most instances, and so there is no reason to think there wouldn’t be innovation in that industry.

Okay, I am almost done. I have a note that says “One Minute”; my watch says that I have gone less than most of the other speakers, so I am going to push it another two to three, but no too far.

There are a couple of threats to innovation that I think are out there. I don’t think it is Microsoft per se. One of the things we’ve heard speaking to people is that the Justice Department has contemplated fixing the operating system in a particular way, and not allowing add-ons. I think this would be a terrible idea. Number one, it is not clear that it should only apply to the operating system. You can imagine applying it to a dominant word processor or a dominant spreadsheet.

Number two, I think it really misunderstands the way this industry works. There are a couple of unique aspects. One unique aspect is that the software doesn’t wear out. In principle, it could last forever. Now we know there is a tremendous amount of obsolescence, but that’s only because the new software is getting better. And the improvements really are to a large extent due to add-ons. Imagine a world where you couldn’t add on any new functionality to a new piece of software. All it could do is perhaps make it run faster and become a little more intuitive. Well, you can generate that by practicing more, which makes it more intuitive to some extent (that is the way the old WordPerfect used to work) and just by buying faster hardware, which keeps coming along the pike all the time. If you couldn’t change what is in the software—functionality—people might just stick with what they have forever, and the producer could basically just go home and shut up the shop. The only sales would be to people who never bought a word processor before or never bought an operating system before. That is a really strange world. Think about word processors without spell checkers, without grammar checkers, without draw programs built in. We’ve come to get used to all of these things. But that is the type of thing we are talking about when we say we are going to fix something, because it has somehow reached a stage of evolution where there are not going to be any more improvements. That, I think, would be a true danger to innovation.

There is one other point that I wanted to make: the current controversy over the browser icons, between Netscape and Microsoft, and whether the OEMs [original equipment manufacturers] should be allowed to change it. We have a paper out there that explains why we think that is a total red herring, it had nothing to do with the issue of a monopoly or even profits. We have a paper that we just finished for Cato, and you can read this if you want or come and speak to us, but I can’t explain this now because I am out of time.

David Theroux

Thank you, Stan. Our fifth speaker is Professor John Lopatka. John is a professor of law at the University of South Carolina. He is former Assistant Director of the Federal Trade Commission.

John Lopatka

What I’ll do is concentrate on the legal standards that apply in this area—specifically about predation.

The first point I want to make is that “monopoly” is not a four-letter word, and that’s true in economics and it’s true in law. There are standards that govern when a monopoly becomes unlawful. Whether this is preceding under Section 2 of the Sherman Act (which is single-firm monopolizing conduct) or Section 1 of the Sherman Act—and the only real difference between the two in this area is that Section 1 requires some kind of agreement between or among firms. What is unlawful is having some monopoly power and doing something wrong with it: doing something wrong to get it or doing something wrong once it is possessed. That leads us to ask is: when does a firm have monopoly power? When is that condition satisfied?

Monopoly power assumes a market. It assumes we’re talking about some kind of relevant antitrust market. Within that relevant antitrust market—and it has both product and geographic dimensions—the idea is that a firm has some unilateral power over price. It can reduce output and increase price. That raises the question: when do we have that kind of a market? How do we define product markets in this kind of area?

Now this has given the courts a great deal of trouble, and maybe because the issue is not always obvious and sometimes it’s really quite hard. In the case against Microsoft initially, the government alleged that there is a market in operating systems for IBM-compatible personal computers. If that is how you defined the market and if Microsoft has virtually 100% of the operating systems for that market, then there is at least the possibility of market power. As Janusz said, “market share is not necessarily a good surrogate for market power, but at least we know what we are talking about.” But immediately you think to yourselves: “Wait a minute. From the standpoint of consumers, I don’t have to get an IBM-compatible PC, I can get a Mac; I can get some other kind of computer. Should we include those in the market?”

That is a good question, and the issue becomes one of substitutability. From the standpoint of the consumer, is one product or potential product substitutable with another? And if they are reasonably substitutable (and that’s about all the Supreme Court has told us—this idea of reasonable substitutability), then we have to include them in the same product market.

Now you say: “Are they reasonable substitutable?” And wagging it’s ugly head comes the concept of lock-in. And we see this again and again. The correct product definition of “lock-in” is IBM-compatible PCs, because once you’ve got one of these things, you can’t easily switch to a Mac. So lock-in defines what the market is going to be.

Now in other cases that leads to claims that a single firm—who has by definition a 100% of the sales of its product—has a monopoly. Of course it does. But the question is, is that a relevant antitrust market? Then the question becomes, what about lock-in? What do we do with lock-in?

Everybody has talked about lock-in, so let me just spin it in a little different way. If you turn lock-in on it’s head, you can talk about it in the sense of switching costs. You’re locked-in because it would be costly to go to switch to a different product. Now if switching costs are positive, then do we always see relevant antitrust markets?

The problem is, we have all sorts of situations where switching costs are positive. For example, suppose you need a transmission job on your car and you call around to garages and you find one located 15 miles away and it offers you the best price. You take a morning off from work to drive your car over to have the transmission done there, and you show up at the garage and the mechanic said, “I told you I would do it for $500. Now it’s $800 or I won’t do it.” Well, there are switching costs. You can’t costlessly go to another mechanic. You’re stuck with that mechanic. So to some extent there is lock-in there. There is a positive switching cost and the mechanic with whom you initially thought you had contracted has some power to exploit you. What do you do with that kind of situation, where there are positive switching costs?

The idea of lock-in was given credibility in the law in a case decided by the Supreme Court in 1992: Eastman Kodak Co. v. Image Technical Services [112 S. Ct. 2072 (1992)]. There, we had micrographics equipment and photocopiers made by several companies, and once you buy a particular kind of photocopier, for instance, then you’re locked into that brand. After that, you may need service, you made need repair parts. And you can’t buy a Sharp repair part and put it in your Kodak copier, so you’re locked in. Does that give Kodak the ability to exploit you once you are locked into Kodak? And what you ask yourself is “Well, at the time I’m choosing Kodak, if this is a competitive market, I could pick Kodak, I could pick Sharp, I could pick some others. At the time I’m making that choice, why is it that I’m going to be locked-in in such a way that I can thereafter be exploited in the after market, so to speak?

The Supreme Court said that there are a couple reasons and the biggest reason is information costs. When you pick Kodak, you don’t know that you are going to be exploited two years down the road when you need service for the machine. You ask yourself then, “Well if I’m making this very expensive decision, why don’t I find out? Why is this going to come as a surprise to me three years later that repair parts for Kodak machines are much more expensive than repair parts for Sharp machines?” The Supreme Court never really answers that. It says that information costs are positive, it’s costly to acquire information and you may not get it. That’s a very questionable assumption that the Court is making.

The second thing you would ask yourself is: “Why can’t I protect myself against that?” It is true that we don’t know everything about the future. But at the time I’m committing myself to one company rather than the other, aren’t there things I can do to protect myself against post-contractual opportunism or post-relationship opportunism? Generally, we protect ourselves against this all the time and we do it contractually. We do it through voluntary agreements. When we are choosing the Kodak machine we enter into a contract that is going to prevent Kodak from exploiting us later. And the Court never really explains why there aren’t measures of self-help that prevent this later exploitation and in particular if that exploitation then does take place, the remedy isn’t an antitrust remedy; the remedy is a contractual remedy. In a lot of cases what we may see are people who didn’t bother protecting themselves and want to be bailed out after the fact.

Let us get beyond the first issue, and let’s say we find monopoly power. Is that enough? As I just mentioned, it is not enough; you need something more. You have to prove that the monopoly was acquired illegally or that the monopoly is maintained illegally or, closely related, that the monopoly is extended illegally—in essence, meaning extended to another market. “I had a monopoly over A and now I have a monopoly over B.” So now the question is, what does it mean to say that this market power has been maintained or acquired or extended illegally? This is where, once again, the issues of network externalities become important.

One way to think about acquiring the monopoly is, for instance, if we think about Microsoft and operating systems, and we assume that they have monopoly power. How did they get it? There are three possibilities. One is that they get it by historic accident. They happen to be the one that made the first operating system; thereafter network externalities come into play and they have to keep it. Another possibility is that they acquired it through anti-competitive means, and more about that in a moment. These are means that we’re going to call exclusionary. The third possibility is that they got it through good, hard competition, through innovation. They created the best operating system and they continue to hold this monopoly because they continue to innovate and so forth.

Curiously, in the government’s original case against Microsoft, the theory was not really that Microsoft had acquired the monopoly illegally. They more or less said, “You acquired the monopoly because of the happenstance that IBM decided to choose you to produce the operating system for IBM computers.” The theory was, “Well, you’ve entered into these licensing restrictions that are going to allow you to maintain this monopoly illegally.”

This is an area where Tim and I may disagree a little bit. It isn’t clear to me; I never understood how these licensing arrangements were going to have that effect. The constellation of licensing agreements more or less had the effect of exclusive dealing. That Microsoft and a particular computer manufacturer, an OEM, would deal exclusively with Microsoft.

Now there is a good reason why they might want to do that. We are talking about intellectual property here, and intellectual property is notoriously difficult to protect. The federal law, the patent law, copyright law can define intellectual property rights. But it’s left up to the firm for the most part to protect and enforce those rights. The government said, “Well, we don’t like these exclusionary arrangements, these exclusive dealing arrangements, because they are going to prevent someone else who comes up with an operating system from having any OEMs out there to whom to sell the operating system.” But why would that be the case? If you go to McDonald’s, you’re going to be able to get a Coke. You’re not going to be able to get a Pepsi. Coke deals exclusively with McDonald’s. Does that mean that Pepsi is excluded from the soft-drink market, if we defined that to be soft drinks at fast-food restaurants? And the answer is no. Pepsi enters into an exclusive dealing arrangement with Burger King, with Wendy’s or someone else. And it’s never been clear to me why it is that some entrepreneurial, some innovating operating system manufacturer could not have acquired an exclusive dealing arrangement of its own with some OEM that decided to use that operating system exclusively. In any event, the government thought that that’s what was keeping the monopoly and that that’s what they’re going to prevent.

The latest incarnation, the latest threat to Microsoft (and this was also the allegation in a recent case you might have read about in the Northern District of Alabama with respect to Intergraph’s complaint against Intel): The idea was that Microsoft is going to take its monopoly in operating systems and extend it into a new market. It’s going to acquire a monopoly over browses. In the Intergraph case, it was to take its monopoly over high-performance CPUs. Intel was going to do that and extend it to a new market in graphic subsystems. The idea of this monopoly leveraging has a long history in antitrust. It’s never been very well thought of. There are serious problems with it, the essence of which is: if you have a monopoly in one product market, what you want to do is exploit that monopoly. And you generally can’t increase your monopoly power by attempting to leverage into another market. You can get into the other market but you have to do that by reducing, in effect, your returns from the market in which you have a monopoly. And would that be a profitable strategy for monopolist? There are very narrow conditions under which it can be, but in general it’s not.

David Theroux

Thank you, John. Also in your packet there are a number of papers. I mentioned the one from Upside . There are also some others, including paper by Tim Bresnahan that’s appearing in a forthcoming book and the presentation by Janusz at the front table. You’re welcome to take that on your way out. Tim has to run to a class in a few minutes, so I’d like to open it up to questions. And if anyone wants to address some specifically to Tim, first please do so.


First, I want to thank the panel. The points of view were very interesting and very stimulating. In many cases I’ve seen government wait until something happens before trying to see how to solve the problem. I think if you analyze the symptom while it’s going on and start to prevent it, that would be a better approach to this kind of scenario. One thing I’ve been observing in the industry is this IP situation—intellectual property. My thinking is that we have to consider that and we also have to consider the international impact because we are dealing with the global market. The question is: how can we prevent this before it happens, rather than wait until it happens then try to figure out how to deal with it? There are a lot of things that can be done to change the law of intellectual property. I’d like to know what the panel thinks about that, starting with Tim.

Tim Bresnahan

People raise two intellectual property issues. Normally, antitrust and intellectual property law are in conflict, and so people often raise in this case the idea that we should let firms in the information technology business earn a return on their invention. And that is a reason to not have intellectual property in invention. It seems to me lately that people have made so much money in this business that you could reduce the return to dominant firms’ innovation by and an order of magnitude. Maybe by an order of magnitude and a half. Take people down from billions to hundreds of millions and I don’t think it would dry up the supply of inventors. So one of the IP issues that people raise is just irrelevant. There’s so much growth.

The other one is this problem of how non-dominant firm/inventors find that their stuff, for which they have less than complete protection under the patent law, ends up in other people’s products. That strikes me as a somewhat more real IP protection problem. The first dominant firm IBM, and then dominant firm Microsoft, was accused of taking other people’s ideas and imbedding them in their own products. Too much of that could actually be a real problem

John Lopatka

In general the idea that sometimes remedies are more effective when they are imposed early on rather than later is, I think, absolutely true. The classic example of that is in the merger area, where we assume that once the merger has taken place it’s very hard to unscramble the eggs. If we stop the merger from taking place, then we have an effective remedy. That’s true, but I’m not exactly sure how it applies in this context. It is true that we may be able to have a more effective remedy if the government intervenes early. But that assumes that what would happen if the government doesn’t intervene would be anti-competitive in a legitimate sense. And that I’m not sure we know. In fact, I’m quite sure we don’t know. That’s the problem I have in applying it here.

One point I’ll make in response to Tim or along with Tim. I think the idea that we know how much protection that the inventor of intellectual property needs in order to stimulate the socially optimum amount of innovation, heaven knows. I don’t think economists have given us a clear answer to that; I think there is a lot of dispute about that. But even if we assume that the intellectual property rights given to, let’s say Microsoft, are grossly beyond what we need in order to stimulate the inventive activity, we do you do about it? If you took patents and you said, “You don’t have a 17 year monopoly, you’ve got an 8 year monopoly.” Well, we could do that; there’s nothing magical about 17 years. If we don’t do it that way, then it’s very hard to see what the standard is going to be when we intervene. So I’m open to the idea that maybe we should be adjusting our response to intellectual property, but I’m not convinced yet that we have a better way of doing it.


Professor Margolis, you where trying to convince us that the QWERTY keyboard was not necessarily the inferior technology, but the only evidence you offered—this is getting a little picky, but you are trying to prove a point—was this study by Strong that said you’ll never recover the retraining costs. But that implies that the QWERTY keyboard was already locked in and the study looked at the switching costs. But that doesn’t really have anything to say about which technology is inferior or superior.

Stephen Margolis

That’s fair enough. I appealed to the 12-minute limit that we were under.

We have quite a detailed discussion of this. It appears in a couple of places. The original academic article is in the Journal of Law and Economics (April 1990). It’s called “The Fable of the Keys.” There we go into detail—probably more detail than you want.

There are a number of ergonomic studies that are essentially mathematical simulations of typing. They take super slow motion pictures of typing and then you can actually model the time it takes for any key to key interchange. Once you have done that, you can simulate typing speed on any conceivable keyboard. That’s one technique that has been used, and there are variations on it so you don’t get exact duplication in these results. But that seems to indicate that QWERTY is a pretty good keyboard. There maybe advantages to Dvorak on order of 1 or 2% or it maybe disadvantageous to a similar degree. There are also experimental studies where people have trained new typists side by side. Again, there is a lot of detail in the paper. But these experimental studies don’t show any advantage for Dvorak.

Now, we don’t say that it is worse. We say that it is not a market failure that we didn’t switch. John talked a little bit about switching costs. That is an important thing to note. There are fixed investments, there are sunk costs. The fact that we stick with the houses that we have as they’re configured even though circumstances change and we might have done it differently if we had to do it over; that’s not a market failure. The market failure in the keyboard story would be if in fact this other keyboard where clearly better—enough to make a switch worthwhile—and somehow we ignored that opportunity. By the way, at the time of the Remington machine, which embodied the original QWERTY keyboard came out, there were other typewriters available competing with the Remington typewriter that used other keyboard configurations. And typing contests became a popular kind of demonstration and there was one important competitor called the Calligraph keyboard. There were competitions between the Calligraph and the Remington typewriter. Other people have mentioned this. There was a Hammond typewriter keyboard that had some of the features that the Dvorak keyboard subsequently had. This is back in the 1880s and 1890s when there really was not a standard established.

It is not just by dumb luck we got a keyboard that works okay. The keyboard that we have is one that has survived and so it shouldn’t surprise us terribly that it’s not a bad keyboard. Furthermore, there is this other ergonomic and experimental kind of research that does cover this matter. Again, it is in “The Fable of Keys.” A little easier to find is a popularization that we wrote of that in Reason magazine. I think it is on their web site, isn’t it? If you go to Reason Foundation’s website, I think you can get it. In the press packet, one of the articles there also has some information on the typewriter history.

David Theroux

It may turn out also with voice-activated systems that the keyboard will become a thing of the past too.

Janusz Ordover

It strikes me as a very confusing line of questioning in some way—and I don’t mean to be offensive—First of all Remington had no incentive to pick a wrong keyboard to the extent they where competing against other manufacturers of typewriters. They had no incentive to pick a wrong keyboard, given what the definition of the optimal keyboard was at the same time. In other words, somebody who has market share to himself—and Remington was a fairly leading manufacturer of these typewriters—had a very strong incentive to adopt a standard which would advance the acceptance of its product. To the extent that it mattered at that time that the keys do not jam, that keyboard may have been superior even though for the typists who where so good that they would not jam the keyboard, another configuration might have worked a little bit better.

The big issue is: right now everybody is typing on their computers, where there is no jamming problem of the same magnitude. Is there anything that we should have foreseen in 1800 or 1950 that would have steered us toward another keyboard? What is the government suppose to do?

This bears on the Microsoft matter. What is it that we are suppose to do about the Dvorak keyboard? Let’s say there was a mistake, especially given where we are right now. I was trying to bring us back to the issues of today as opposed to the keyboard itself, which is a historical specimen. What I’m trying to suggest to you in thinking about these issues is that we may want to worry about the role of the government ex post intervening and trying to undo the market forces which at that time where operating quite well and moving the standards toward the one that combined perhaps optimally the convenience versus the speed versus the jamming problem and so on. And I think that’s the point as to the way the market works if there’s any particular type of market failure.

Stan Liebowitz

For those of you who might be interested about the keyboard: The jamming problem can be solved by making the keys go from left to right. And so what QWERTY does is move your hands alternately: left, right, left, right. It turns out that that is one of the factors of fast typing, which is why QWERTY turns out not to be a bad keyboard. But it also turns out that the real constraint in typing is your brain not your fingers. That’s why there’s not that much of a difference.


You were mentioning some of the cases that might be relevant to today. We talked a lot about antitrust enforcement in the software industry. There also appears some concern on the part of antitrust officials that companies such as WorldCom will own too great a share of the Internet backbone. Can you address that concern? Should or can the government do something about that?

Stephen Margolis

The research that we’ve done does not very directly confront that. I guess I have two things to say. One is: this may look like a more traditional antitrust issue, which is that you have a firm with large market share, no particular hint of lock-in, no particular network effect or externality. For example, I really don’t care whether you call me on MCI or AT&T when the phone rings. So those issues don’t seem to be present.

On the other hand, this maybe a nice illustration of how the current technology modifies our perspective for antitrust. WorldCom, because of their new combinations of communications technologies, looks like it’s going to be the player. But at the same time it’s starting to look like other kinds of communication—whether it’s how we use the shielded cable that comes into our house for cable TV or how we use the Internet. It just stands ready to overturn a whole lot of market positioning in telecommunications. The new technology is so vibrant that we have new technologies coming along and turning things over. These are the gales of creative destruction that Stan brought up, and it strikes me as though, again, we have a case that may alter how we think about antitrust and whether (and this is consistent with what Tim was talking about) in this environment we have competition from the new generation of technology and therefore need to take perhaps a dynamic and not a static view of market share and things like that.

Janusz Ordover

Just one comment. That deal is not that different from many mergers that the Antitrust Division has analyzed in the past. An interesting issue is going to come up. If the technology is developing that way, it is going to relate to what always matters in these markets: the terms of access or interconnection. So if there’s a competing network that will require some transmission backbone from the merged company, will the merged company be able to deny access on terms which are pro- competitive? As often happens, it is the question of the price, the terms, under which an entrant can perhaps put together or try to assemble a competing network to the extent that the entrant cannot configure the network—a ubiquitous network—from the start. Issues of pricing are the ones that are most vexing in these cases.


I take it from the comments that most of you don’t think there’s anything the government should be trying to do on the basic issue of integration of new products into the operating system or separation of operating system and applications—all those big issues. But is there anything that you think would be beneficial at the margins of licensing terms or contractual arrangements that might serve to reinforce some of these other effects that Microsoft has enjoyed?

John Lopatka

It’s a hard question to answer to some extent because when you start talking about doing thing at the margin, it depends highly on exactly what we would do at the margins. I don’t have any specific things in mind, so it’s hard to evaluate that. In general I don’t oppose the idea that in theory there maybe things we can do at margin, but I would also say that in general—especially in high-technology markets that are evolving so quickly—by and large, government intervention is much less effective than the market itself. So I would approach those kinds of problems by saying, “I’m going to rely on the market unless you can prove to me in a specific case for a specific remedy that government intervention is a superior solution.” And so it’s another way of saying that in general I’m skeptical that antitrust has a valuable role to play both in terms of the direct cost of antitrust—antitrust isn’t a costless commodity—and secondly, and more importantly, the potential indirect costs of dampening creative, innovative activity. I’d rather let the market take care of that.

Stan Liebowitz

People talked a little while ago about intellectual property, and I don’t really know if there is a problem. I hear stories from people who claim that, for example, Microsoft is taking someone’s idea and putting it in their software, and isn’t that terrible. My answer is: if in fact people are working on new ideas and the ideas are being appropriated and therefore the creators are not getting the rewards—if that is true, that would be a problem. But of course there is no reason to think Microsoft would be anymore guilty of this ability to take the idea than anyone else. Any other big firm, if the idea can be appropriated, would be able to appropriate it. In fact, any other small firm would be able to do it. Certainly, Microsoft shouldn’t be able to steal anything. If there is free stealing going on, then there is a problem with intellectual property and we might want to look at the intellectual property laws. But there’s no reason to think that if there is a problem per se then there is something specific to Microsoft that wouldn’t be a much more general problem.


I’m wondering if you can expand a little bit on a point that Mr. Margolis hinted on. Which is to look beyond the current Microsoft case and elaborate on whether the existing antitrust laws really apply to this fast-moving high-tech industry. In other words, is there any reason for lawmakers or the government to think about rewriting the antitrust laws so that they could potentially be applied to this fast-paced, highly innovated high-tech industry, or are the laws in place applicable and efficient enough to react?

David Theroux

Can I interject one question also? This relates to this intellectual property rights question and to the previous question. One of the dimensions of this that might also be relevant to his question is: what is the relationship of firms developing contracts as a way to police these property rights among themselves and also among consumers, and how that might relate to these questions of antitrust?

John Lopatka

I’ll give you my perspective. I think it’s useful to keep in mind that antitrust really speaks of two things. What we’ve been talking about today is the predation side of antitrust. The other side of antitrust is collusion. By and in large you can break antitrust into those areas. Now, I think collusion is possible and is potentially a real threat in high-technology industries, as well as any other industries. The antitrust laws are certainly supple enough and they are certainly adequate to attack collusion, and I have no problem with that. When collusion is actually found, I think that using antitrust laws is fine.

I think that antitrust law with respect to identifying, correcting or preventing exclusion or predation is another matter. There, I’m not sure there’s any lack of legislative authority. The antitrust laws go back to 1890 and they’re sufficiently open-ended that they can be changed. They evolve and they have evolved over the years. I don’t think there’s any legislation that’s needed to allow the authorities or private individuals to attack predation. I just think that it is an extraordinarily difficult task to identify what really is predation and then come up with an effective remedy. I just want to make one point about the intellectual property too. Intellectual property is this very curious character. When we identify an intellectual property right, we give you a property right over this information. That may or may not give you a monopoly in an economic sense. You can have an intellectual property right over a new mouse trap, but if it doesn’t catch mice any better than old mice traps, that patent isn’t going to do you any good at all. The curious aspect of recognizing something like patent or copyright is: we’re going to give you the property right even though it may give you a monopoly. And that’s different from what we do with other kinds of property rights. The reason we do that is because of this curious economic property of intellectual property: it’s a public good. It’s very hard to exclude people. It’s very easy to copy. It’s very easy to have non-rivalrous consumption.

So I’m afraid to some extent we’re kind of stuck with this. If we are going to identify and define intellectual property—and we have to do that in order to create some or what is perceived to be an adequate incentive to create it—then we don’t know what to do other than allow you to have a monopoly once you have got it. If it happens to give you monopoly power, then the question becomes whether we should stop or limit the duration. But those are very difficult questions.

Stephen Margolis

In a couple regards, some of this may less new than we are making it. What we are saying is that we ought to take a dynamic view of this. Perhaps we always ought to have taken a dynamic view.

Some of the things John was alluding to, such as anti-collusion aspects of antitrust law, are on firmer ground than the anti-monopoly aspects. In the early days of the Sherman Act, there was a lot of fumbling around with the monopolization cases, the Section 2 cases, before we began to think, or the courts began to evolve doctrines, about those aspects of the law. For a long time we have been fumbling about what ought to constitute monopolization, what is predatory and so on.

I would also point out that in the early days of antitrust law the law was itself was a response to changing technologies. That is, there were interest groups that where being harmed by the new technology. The railroads had allowed big factories to emerge and serve the entire country in some industries, or a meat-packing region like Chicago to emerge and serve the eastern part of the country. And this was displacing people and they were being hurt. Chain stores like A&P began to displace “mom and pop” grocers and local wholesale jobbers for groceries. And a lot of the impetus for antitrust was to protect special interests. We had that going on then, when we defined antitrust doctrines around the turn of the century, and at the turn of a new century we have some of that same thing going on today.


Am I hearing that there is nothing fundamentally different about the high-tech industries that asks for different thinking today?

Stephen Margolis

One issue about technology that I don’t want to address so much is: How different is innovation today than was innovation at sometime in the past? There is a matter of generational conceit that our breakthroughs are enormous but that figuring out germ theory or something like that wasn’t such a breakthrough compared with what we are doing now.

Janusz Ordover

The FTC spent several days trying to think about what the antitrust for high- technology industries ought to be and having gathered luminaries and less (such as myself), they basically came down to the view that we’re doing okay but that we need to get some more understanding of how these new industries—to the extent that they are as new as some people claim they are—require a somewhat different thinking.

There is a tendency in antitrust not to fashion enforcement narrowly for a particular industry, although health care has its own sets of guidelines. The general principle is to try to paint with a broad brush and let the enforcers—the courts, the Justice Department, the Federal Trade Commission—evolve the doctrine based on litigation, enforcement practice, and antitrust thinking and economic thinking. You’re going to see a refinement of the doctrine. But I do not expect anyone to say, “Gosh, we’ve got to throw out Section 2 of the Sherman Act because we don’t know how to define predation in a network industry.” I think what we are going to see perhaps is a judge—mistakenly or not—ruling that, in a case of a network industry with very low marginal costs and a very strong race towards a sole winner, predation involves x , y , or z . And they will first look at price versus marginal cost,. as we are always taught. Then they will say “well, maybe that’s not the right standard” and they will try to modify it and it will be appealed and back and forth. But I don’t expect—and I would be very fearful of—industry- specific rules. Because that only creates confusion and also calls for a special pleading, and nobody is receptive to the kind of stuff.


I wanted to go back to the issue of suppression of innovation. I guess this first goes to Stan. One of the cynical observations you hear around Silicon Valley often is that start-up companies study the market and will only start up a company that is either a niche that Microsoft doesn’t care about or is a company that eventually Microsoft will buy. I don’t know what legal standing that has in terms of antitrust laws. But it is something you hear often around here and I wonder what you think of that type of suppression of innovation it exists.

Stan Liebowitz

I think it comes back to the issue of intellectual-property protection. One thing I think it’s telling us is that Microsoft is not a fat, lazy monopolist. People aren’t afraid of fat, lazy companies too much. I think that there’s probably some truth. If you are going to come up with a new innovation and you’re looking for a market niche and you know that one market has tough players and another market has less tough players, you would want to go to the market with the less tough players.

Let’s give a hypothetical. If you were going to introduce a new automobile and you were in Australia and you go back 15 years. Let’s say there were no Japanese cars in the U.S. and we only had the American automobile market and you wanted to introduce your automobile. Would you want to introduce it into the Japanese market or into the U.S. market? Given the two markets, particularly where they were 10 years ago, it would make a lot more sense to go into the American market because there was less competition here. The firms just didn’t seem to be as aggressive and know how make cars as well as they did in Japan, at least at Toyota and a few other companies. And I think that’s what’s going on today. Microsoft may not get it right the first time, but one thing they seem to be good at is getting it right eventually. That is what happen with the browser. The first versions of Internet Explorer were not very good and they really didn’t get any market share. But Internet Explorer 3 sort of matched Netscape and most people think Internet Explorer 4 is better. Microsoft has a record of coming out with a product and eventually making it better than its competitor. Why? You could come up with lots of hypotheses. One hypothesis you could have is that they have very good engineers, and they must pay them well, and they work them very hard, and they wind up coming out with good products. Why does Toyota come out with good automobiles year after year? The answer would presumably be something like that.

Now you can come up with stories that are somewhat different, that are somewhat more sinister, which is that they have some special secret that they don’t make available to everyone else. I have heard stories to that effect; that there are parts of the operating system that they know about and no one else knows about and if you have access to those special calls into the operating system, you can make your software do better things. I’m skeptical of that—in part because I took a look and what is it that makes Internet Explorer better than the old versions and relatively better than Netscape. And a lot of it seems to be the design, not some inherent feature that no one else could have access to. So I think they are a sort of lean and mean company. There are such things. And the way they turned themselves around on the Internet was fairly remarkable. They where sort of pooh-poohing it and saying they weren’t going to pay to much attention because they didn’t think it was that important. Then they marched the whole company around and made a very major change in direction in a very short period of time.

There aren’t very many companies that can do that type of thing. General Motors picked Saturn to turn the company around. Ten years later, GM’s not much different. It’s a hard thing to do. We have to give Microsoft some credit; they’re a pretty good company.


In terms of competition and innovation, what’s your take on this whole idea of network computing? Is it more of a case where they fumbled the ball, or was it more the case of these Wintel monopolies getting into the act and saying, “If the price point is where the competition going to play—which is below $1000—then let’s go ahead and build machines under $1000.” Because for years the price point for the PC was never under $2000 for a typically configured machine. What happen in the network computing game?

Stan Liebowitz

I think this is a hardware story because the software is not a big part of that cost. I own some Seagate stock and one of the things I have noticed in the last year or two is that the price of the hard drive has plunged, and so have the profits of the manufacturers. I think that when people started with the network computing, they thought the price point would be such that there would be a place for them to enter, that we were going to have a computer without a hard drive or not much of a hard drive and so there was going to be the large savings. Then the hard drive turned out to be dirt cheap, so there wasn’t much savings there. I think it caught people by surprise, and I don’t think there was any conscious decision on the part of some malevolent controller of the computer world to bring the price of computers down under $1000. Cirrus came out with the GX chip and that put some pressure on Intel and the hard-drive market went into a crazy dive in prices. And the network PC people took a look and said “We’re stuck. There’s no place for us to come in and undercut these guys with a better product.” There are very few savings. That’s what I think happened.


So it’s not a case like the airlines, where when the smaller airlines come into the market, the bigger ones start cutting their prices, drive out the smaller ones out of the marketplace, then raise the prices again?

Stan Liebowitz

I don’t think Seagate, Conner and the others sit around thinking “Let’s lose lots and lots of money now to keep the net PCs from taking over.” I don’t think there’s any conscious decision to do that.

Janusz Ordover

There are too many players to effectuate such a dire conspiracy. Even if it were true that airlines might be engaged in such anti-competitive or predatory conduct, usually it’s one airline facing an upstart. Here we are talking about a very large and diffused industry—in hardware suppliers and software suppliers—who have to somehow coordinate their actions without that ever becoming public. It’s not an issue of the price but also an issue of convenience. I like to be able to save what I worked on at the office and take it home or be able to keep stuff available to me. So there are many applications in which network computing may actually be a solution. Airlines seem to be quite fond of that. I believe that’s going to possibly be a challenge—not a frontal challenge but a challenge on some broader margin to the standard desktop. It’s a hardware story, it’s the plunging prices, which are quite astoundingly low compared to what you were getting even a couple of years ago.

Stan Liebowitz

It’s not clear that it was a better idea, this idea of wanting to have control. When I first heard about these net PCs, it sounding like these guys who used to run the mainframes and had everybody on terminals trying to get control of the world again. So it wasn’t clear that users who had gotten used to a certain amount independence and were able to have things the way they wanted it on their machines were ever really going to go for that. Then there was also this other problem. They missed the prices. Their predictions were wrong.


I find it a little bit strange, this defense of Microsoft, because I think you’re actually looking at it from a very academic point of view. I don’t know how much you’re out talking to really small companies that. I’ve talk venture capitalists who say, “We will not fund a company that is even ancillary to what Microsoft could do or potentially could do. I think all of us have gotten a little tired of hearing Sun and Oracle complain. But the people that we really should be talking to are the small companies who are being bullied—and they have told me that and I’m sure that there are other reports here that they’ve talk to about it—but are afraid. They can’t report it because they are in a situation were if they do, they could go out of business; and if they don’t, then Microsoft is controlling their businesses. I’m wondering how much you’re talking to small companies and IT folks at big companies who are in a situation where they don’t have a choice but to deal with Microsoft or to buy Microsoft products.

Stephen Margolis

Well, Stan covered some of this. If you’re competing against somebody who is very good, you might feel you have something to complain about. Playing one on one with Michael Jordan is not a great situation; you might come away from it kind of sputtering. But the other thing I think is very important from a social point of view—and I guess here I’m about to get academic once again—is that for an inventor looking for a place to go, it’s not irrelevant that someone else good is already working on the problem. In fact, from a social perspective we would like inventors not to invest their resources or use their creative talents in places where there is a lot already going on.

In economics we have this problem that is referred to as “the waste of the race.” Intellectual property law, patent law in particular, awards the whole invention to the first to patent or the first to develop an idea. It’s argued that the result of that is: once inventors are on to an idea, they speed up the process and increase the cost of inventing in a way that is not socially constructive; that is, the racing itself isn’t useful; it wouldn’t harm society much for the process to slow down just a bit, but it might cost inventors a whole lot less.

A partial solution to this is that inventors look around at what other people are working on and if they see that there’s somebody that’s already got something going or they see that somebody is already advantaged in the development of a particular technology, then they stay away from that. Well, that’s a good thing; you don’t want duplication of effort. So in areas where Microsoft is working or can work, where there is no blinding insight that one small firm has that Microsoft doesn’t have, it is probably a good thing for inventors to look for niches where they aren’t duplicating the efforts of others.


I’m not talking about somebody who’s saying, “Hey, I have a great word processor.” I’m talking about people who are saying, “I have this great speech technology,” or whatever it may be, and Microsoft has mentioned that they would like to get into this area. And you often see that Microsoft doesn’t come out with this technology for years but this effectively shuts everyone out for a period of time. I’m not anti-Microsoft; I think that they are a good company in a lot of ways. But I disagree with this gentleman who says that they have not stifled the competition. I think that has happened, either because of who they are or because of direct means. That was my point.

Janusz Ordover

I’m just looking at a chart in the April 20th edition of Business Week which indicates that the number of software startups has increased significantly between 1994 and ’’97. There was an amazing explosion in the number of technology—and I presume some were related to computing—startups. Whether or not it’s true what you are reporting—I have no reason to doubt your voracity and I have no reason to doubt that some people would rather enjoy options to selling themselves to different players or standing on their own two feet. Nevertheless, at least this simple diagram tends to suggest that the industry is doing very well, and that it is doing better now, between 1993-97, than it was doing in 1993.

So the question is: can one exactly identify what we have missed out had not Microsoft been sitting out there? And even if so, what is it that the government is suppose to do? One could attack vaporware [software which never becomes available as announced] and Judge Sporkin tried to attack vaporware as being an antitrust violation. Can we write clear guidelines for what is vaporware and what is not? It is my understanding that in the software business and in high tech, everybody advertises long before they have the product on the market. For a variety of good reasons; some bad also. Again, I want to be sympathetic, but one also wants to know whether there is anything that the Department of Justice or Congress can do that would vastly increase the output of software and technology from startups without actually making the situation worse. If there is such a solution, I guess it should be debated, but I have not of heard one that would have costs that are small relative to be potential benefits.

Stan Liebowitz

All I can say is that the idea is to supposedly protect competition, not competitors. If the competitors know of something that Microsoft is doing that crosses the line, they should come forward. If Microsoft is putting horse heads in people’s beds when they wake up in the morning, then that’s probably going over the line. But if people are just nervous because they think Microsoft is going to wind up making a better product than they and they are going to wind up not making any money, then that’s not something we need to worry about.

David Theroux

One other aspect of this which relates to a couple of the earlier questions is the issue of regulatory capture. For example, in most regulated industries you find that the regulations where either brought into being by the major players or that the regulatory body ends up being captured by those who are regulated and used for their own purposes. Those that are the most dire defenders of the policies are the regulated, and it is the small guy who is essentially restricted from the market. We create an actual, physical, government-imposed barrier to entry. That’s the alternative side of this. The question is, the way the market is performing, whether it’s also creating these restrictions. I think it is important that we make sure we consider the various other options that are being presented.


I’m trying to get a handle on this fuzzy concepts of predation. I wonder if we could get a quick thumbnail sketch of what in the industry are clear cut, legally actionable examples of predation, which business practices are not clear cut but are ambiguous, and which are clearly not legally actionable but are considered aggressive business practices?

John Lopatka

Let me tell you that in general the legal standard—and this is a direct quote from the Supreme Court not too long ago—is that for antitrust purposes, predatory conduct is conduct that excludes rivals on some basis other than efficiency. It is a very general statement but that’s what the Court gave us. And if you start to dissect that statement—that it has to have an adverse impact, that it has to harm competitors—then effectively it means it also has to injure consumers, because it has to be on some basis other than efficiency. And if it’s an efficient practice even though it injures a competitor, it is not predatory in a legal sense of the word. So it is easy in these cases, in the Intergraph case that I mentioned before, it’s easy to see how Intel’s practices injure Intergraph. And that’s what the district court spent most of its time on.

In specific predation cases—and to those of us who have spent a lot of time looking at the specific practices that where alleged to be predatory in a host of cases in which came up to the Supreme Court—the problem is that very often these practices are ambiguous. You can tell an anti-competitive story—and there are models that would support it. Or you can tell an efficiency-enhancing story. Integrating a browser into Windows certainly has the potential of hurting a competitor in the browser market, but it may well be efficient because it’s going to work better and consumers prefer the convenience of having the thing integrated. In those situations I think it depends a little bit on your outlook, but a court is going to have to make a choice. What do we think it is? Is the anti-competitive story more telling? Is the pro- competitive story more telling? Depending upon your notions of institutional competence and your priors in this, either you’re going to see predation or you’re going to see something that’s not predatory.


And the third category would be: what business practices are aggressive but clearly not legally actionable?

John Lopatka

That is probably the greatest category out there. That is, the idea that you get a monopoly through historic accident. We’ve been talking about that a little bit and it’s consistent with this path dependence idea.

Steven Bryer years ago wrote a little article about this and at the time there were a lot of economists out there who believed that you could get a monopoly through historic accident and you could keep it forever. And I’ve never been convinced by that and I think most economists would say that that’s not realistic. If you are going to keep it, you’re going to have to keep it through efficiency-increasing behavior or you are going to keep through exclusionary behavior. And if you can’t identify these instances of exclusion, what is left is that these practices tend to be efficient. Frank Easterbrook, a judge on the 7th Circuit, has for a long time said “We’re pretty good, at least we have a methodology for identifying exclusionary conduct, predatory conduct. We’re not at all good in explaining how conduct increases efficiency. We just know that somehow it works to the private benefit of the firm that’s using it, and we’re not quite sure why it’s efficient.” And what that leads him to conclude is that we ought to leave this stuff alone. It’s just much easier to articulate the anti-competitive story than the pro-competitive story. But when you hold it up to analysis, the anti-competitive stories don’t work. What is left is the pro-competitive story. That’s the long answer. The short answer is: I can’t give you specific examples of hard-nosed competition that is not exclusionary.

David Theroux

Okay, that will wrap it up. Thank you everyone for joining us and I hope you can join us again.


  • Catalyst
  • Beyond Homeless