Vertical Integration ICANN Meeting Sydney, Australia Monday, 22 June 2009 >>JOE SIMS: Okay, folks, can I have the -- okay, if we can go ahead and get started, so we don't run too far over schedule. Thank all of you for coming. I hope this will turn out to be a useful session. We're going to try our best to make it so. For those of you who don't know me, I'm Joe Sims. I'm a lawyer from the United States who has been working with ICANN for many years. And what we're going to try to do today is spend some time laying out the relevant economic principles for -- that should apply to this debate. And then, hopefully, we'll get some interaction with the audience and talk about the application of those principles to -- to the actual facts before us. This issue of vertical separation, integration between registries and registrars, has been with ICANN since its creation. Obviously, at the beginning, the debate was informed by the fact that there was only one commercial registry operator and one registrar. As competition has been introduced, first in the registrar business and then in the registry business, this issue has stayed alive, continues to be discussed. Now we have hundreds of registrars. We are on the cusp of maybe having hundreds of registries. And so it's timely to focus on this issue again, taking into account those facts, antitrust is about facts on the ground, not historical facts, hopefully the best view we can take of future facts and that's what we're going to try to focus on today. Obviously, vertical integration is a very common phenomenon in business. There are lots of vertically integrated businesses out there. Think of, you know, any number of companies, oil companies, manufacturers of all sorts of goods that also wholesale or retail those goods or make components that they use to manufacture their goods. Vertical integration, like almost any business structure can have good competitive effects or it can have bad competitive effects. It can benefit consumers by leading to lower prices, better quality, better service, facilitate innovation, but it can also harm consumers by making them face higher prices, reduced quality, or too little product variety. So vertical integration, like everything else can be good, can be bad, depends on the facts. So we're going to -- we're going to start by introducing the panel. The participants, the people who are going to do most of the talking, hopefully I'm not going to do very much, are Steve Salop and Josh Wright. Let me talk a little bit about their background Salop is one of the most recognized and respected antitrust economists in the world. He's the author of any number of very significant publications, probably the most prominent of which was the article, the first listed there with Tom Krattenmaker where he developed and articulated really for the first time this notion of raising rivals' costs as a significant potential antitrust concern and that concept is now deeply embedded in antitrust theory and policy. Steve is a professor of law and economics at Georgetown University, he's not a lawyer, but he plays one a lot. And he teaches law and economics to law students. Josh Wright is also an economist and he is a lawyer. He teaches at both the law school at George Mason and the economics department at George Mason University. He's also written widely in the area of vertical restraints, vertical contracts. And so these two gentlemen are actually, unlike many of us who sometimes put ourselves out as experts, these two guys really are experts on the economics of vertical integration. So here's our agenda: Our agenda is we're going to start with a presentation of what the relevant economics and to a little extent, law, is, in this area. Our objective is to try to lay an educational foundation, not to take positions, not to advocate one side or the other, but, to the best of our ability, we are human, so we have our biases, I'm sure. But to the best of our ability we're going to be try to be as objective as we can. And then when we finish that, we're going to move to a more open discussion focused on ICANN facts. The first part of this will focus more on principles, with a few ICANN facts thrown in. The issue here is complicated, as it is in all other instances of the application of economics to business organizations by the fact that vertical integration is not -- can be not just structural connection and a corporation and subsidiary or affiliate but you can also have vertical integration by contract. Agreements of various kinds that have very similar, if not identical, competitive effects. So you have to evaluate the likelihood of harms and the likelihood of benefits of both structural and contractual vertical integration. So we're going to talk about both of those things and talk about what are the circumstances in which you should be concerned and what are the circumstances in which generally there is no competitive concern. So with that, let me turn it over to Josh Wright to start our -- the educational part of our program. >>JOSHUA WRIGHT: Okay. One of the things that Joe mentioned was that one of the complications to do with assessing the economics of vertical relationships generally is that integration and contractual relationships are substitutes. And so many of both -- I'll be talking about both, some pro-competitive virtues about vertical integration that we know exist out there in the world in markets where we've studied vertical integration, but we can also, firms can also achieve many of those pro-competitive virtues by contract and the same applies to the anticompetitive stories that we can tell about vertical integration that happen out there in the world. With all of these we know both the anticompetitive and the pro-competitive effects can happen both structurally through vertical integration and through vertical contracts. So for most of the discussion I might use the term "vertical contracts" generally and it will be meant to apply to both and I'll try to distinguish when something applies to just one or the other. One of the things that economists know, have gone out and studied vertical contracts generally, is that there are theoretical conditions under which vertical integration or contracts can lead to consumer benefits on the one hand, Joe mentioned things like lower prices, increased quality, more innovation, service. There are also conditions under which vertical integration can lead to some competitive problems. The main goal of my half of the presentation is going to be to lay out what I'll call the general economic analysis of vertical contracts. And when I hand it over to Steve, we'll be talking a little bit more about application to ICANN-specific facts. But my goal is primarily to get across the general conditions under which we think vertical integration might lead to competitive benefits versus competitive harm. This turns out to be for economists a relatively tough task to tease out pro-competitive from anticompetitive effects of vertical integration. This is, to borrow another phrase that Joe used during his introduction, an antitrust analysis, the analysis of these vertical -- vertical arrangements is one where facts matter, facts matter a great deal, institutional detail matters a great deal, identifying the conditions under which these pro-competitive or anticompetitive effects are likely to exist is one that depends critically on a sort of roll up your sleeves, hands-on, fact-intensive analysis. Again, this means, on the one hand, that sort of bright-line pronouncements of about when vertical integration is going to be good and bad are very difficult to make. Answers are difficult to come by. But I hope that by sort of exploring both sets of the anticompetitive and pro-competitive stories, we can be helpful in identifying the conditions that we should be focusing on as a factual matter. I'm going to start with the competitive harms and then sort of end with some of the pro-competitive effects of vertical integration. So there are really three general classes of anticompetitive effects that we think about with vertical relationships, vertical integration, vertical contracts. These are exclusion of rivals, so in which case we're talking about the creation or maintenance of market power. Evasion of price caps. And the third being the exchange of sensitive competitive information. That might lead to exclusion on the one hand or collusion, price-fixing, on the other. So I'm going to discuss each of those in turn and then shift over to some of the pro-competitive stories. So I want to start first with I think what is probably the primary anticompetitive effect that focus think about when they talk about vertical integration, vertical contracts leading to exclusion of some sort. Now, there are a number of ways this can happen, so I'm going to talk some about general examples and leave some of the fact-intensive application to Steve. But we can think about in our general setting with an upstream/downstream firm in the setting we're talking about here, registries upstream and registrars downstream. The easiest, most intuitive anticompetitive story to tell is one that economists will refer to as input foreclosure and so an input foreclosure story about maintaining market-mile-an-hour in the registry market would work in the following way. We would think about a registry vertically integrated or with vertical contractual relationships with affiliated registrars refusing, like having the affiliated registrars refuse to deal with competing registries. That's sort of the fundamental exclusion story when we're talking about maintaining market power in the registry market so when the affiliated registrar division again, I'm using the term "affiliated" to encompass both a vertically integrated infrastructure division of a vertically integrated registrar- registry or we can achieve the same relationship via contract. In those instances, when we've got the affiliated registrar refusing to deal with a competing registry, the effect is ultimately that the registry competitor can't achieve, economists say, minimum-efficient scale, they can't achieve a sufficient distribution, they're denied access to this input and barriers to entry are increased and our original registry can raise barriers to entry and keep its market power for a longer period of time resulting in either the same harm to consumers for a longer period of time or giving the registry the ability to raise price. Now, in that story, market power in the hands of the registry was sort of a prerequisite for the condition. And what you'll see on most of these anticompetitive stories, these effects, is that they'll turn on the existence of a market power at one level or another and we'll talk about how some of those result in consumer harm. Equally -- excuse me, if we turn that story on its head, we can tell the story about create rather than maintaining market power in the registry market. We can tell an -- a story about an anticompetitive effect in the registrar market. Here we can imagine the registry division is refusing to allow other registrars to distribute its registry. There we can imagine the effect being driving the registrars out of business or reducing their quality and giving the affiliated registrar, the vertically integrated registrar, a competitive advantage and the ability to monopolize registrations of the other registries. This is a similar story to the maintaining market power story with the register, with the registry turned around. So here the effect is to create market power on the registrar level by refusing access for the unaffiliated registrars to registries. Now, again, with respect to each of these basic stories, whether maintaining market power in the registry level or creating it on the registrar level, there are a number of preconditions that we normally think about as economists as sort of facilitating that sort of exclusion or creation of market power. One is the existence of market power on either level. And we talked about that. And the second is whether the excluded firm is able to realign its supply contract. So what you might think about is whether, if we've got a registry, who's being excluded from access to registrars, one of the questions we might ask, we care about, is whether they have access to sufficient unaffiliated registrars to be able to compete effectively. These are just economic reconditions of the theories. Note that although we've -- we can do these examples all with vertical integration, it's important to note that these aren't specifically, as I mentioned earlier, vertical integration concerns. These are largely concerns about market power that exists being either maintained or used to create additional market power through the use of some vertical contracts. Vertical contracts including vertical integration. In fact, economists generally, when referring to sort of exclusive dealing preferential contracts involving preferential treatment of distributors, generally have a term that they use called a vertical integration by contract in which you can achieve many of the same, in this case, competitive harms by relying on contract rather than integration. The second basic antitrust story we can tell with respect to vertical integration is, is the evasion of price caps. Again, if we have price caps put in place to -- perhaps to respond to what we perceive to be market power in some registry and we have a price cap, then the familiar economic story with vertical integration is that if we allow the capped registry ownership of a registrar or we can do this by granting exclusivity to the same or, again, this contract integration substitution point. Then that integration can allow the registry to evade the price cap and ultimately harm consumers. Notice, as an economic point, that this incentive to discriminate the incentive to either use integration or preferential contracts increases when we've got caps along with market power. This is one of the points that's discussed in the report issued by CRA. Finally, the third anticompetitive story that economists have studied with respect to vertical integration and contracts is the possibility that the vertically integrated unit can harm consumers by misusing information obtained by virtue of the vertical integration. So we can think of a couple of different anticompetitive effects that could arise from that sort of integration. One would be the misuse of information obtained from unaffiliated registrars or registries, we can tell the story in either direction, to favor the affiliated division and competition, right? So you can think of acquisition of competitive -- excuse me, acquisition of competitive information, sensitive competitive information, giving a competitive advantage to a firm at either the registry or registrar level, and resulting in exclusion of those levels, at those levels, of the unaffiliated firms. And you can also think about this exchange of information as what economists will call a facilitating practice, a practice that might facilitate collusion by allowing the sharing of information between firms that would otherwise be competing against each other. So between either on the registrar level is probably the most intuitive way to think about it. So you have competitively sensitive information in the hands of one firm and that information's used to facilitate sharing of sensitive competitive information that can result in a reduction of competition. Now, those are the three basic anticompetitive effects that we think about arising from vertical integration. But economists also talk about and have studied in a number of markets where we observe vertical integration a number of pro-competitive benefits. And so I'm going to talk about, now, the pro-competitive side of the coin and some of the conditions which apply where we think that these pro-competitive effects are likely or not. Now, I'll focus on a couple broad categories, but the same as with the anticompetitive effects, in these broad categories, we can normally achieve these pro-competitive virtues both through integration and through contract. We can think of exclusive or preferential long-term arrangements between registries and registrars as a substitute for integration. Really, the primary economic benefit of vertical integration most commonly observed by economists who have studied vertical integration in various industries is lower prices through what economists refer to as "eliminating double marginalization." So let me give you a couple examples, vertical integration in a number of ways, the principle is it can reduce costs, reduction of costs can lead to lower prices, this is not incredibly controversial point among economists but we can think about the mechanism through which this reduction cost works as having a lot of different applications. So one application you might think of as being sort of relatively intuitive if you think about firms upstream and downstream with patents and each wants to have an upcharge on its royalties, so each has got a patent, you need a couple of products to -- patented products to make the ultimate product that's sold to end users. In the absence of vertical integration, one of the most well- known results in all of economics is that you will get higher prices without integration than with in those instances. What vertical integration allows the firm to do is eliminate that second margin, it reduces price and discounted as a competitive benefit. An area where sort of less abstract, in the patent example, economists have studied in great detail some sort of brick-and-mortar markets that have experienced with vertical separation requirements. One of these industries, perhaps one of the most studied industry in economics with respect to these vertical relationships, is gasoline in part because of the availability of data. So states will have gasoline enforcement requirements, right, separating upstream from downstream firms, essentially imposing at the state level a vertical separation requirement. And what that literature shows is that in states with vertical separation requirements, you end up getting lower output and higher prices than in states without. In large part because what's going on is the vertical integration results in eliminating or minimizing that double marginalization. Now, you can think about sort of closer -- closer to home, as an example of this effect, I am not -- I'm not an ICANN institutional detail expert but I've read about this test case for allowing vertical integration in single owner TLDs and you might think about vertical integration in those cases allowing the owner to reduce the cost of having to go out to a registrar to provide those services. And so resulting in lower prices in that instance. A second competitive benefit of vertical contracts and vertical integration separately is inducing downstream firms and I may start using "registrar" from here on out, to provide what I have labeled on the slide here as specialized services. And that point in this bottom bullet allowing more scope of innovation in the delivery of products and services go together to some extent so we will talk a little bit about that. We might imagine instances in which the upstream firm would like the downstream firm to provide some specialized services and this context we might think about a registry that wants to offer domain names only to registrants that are going to offer highly secure access to information, certain privacy protections or other specialized services you might have -- more -- more -- better examples than I do. But if we imagine this general set of specialized services that without some sort of relationship with the registry, the registrar has little or insufficient incentive to provide. This might be because the market for registrants who are demanding those services is small and so we've got to compensate the registrar for providing them, as one reason. So there are a number of different ways you can think about the registry as compensating the registrar or, more generally, for the upstream firm to compensate the downstream firm for providing those services which they may otherwise have insufficient incentive to do, vertical integration is one method by which that occurs. Other methods through which that occurs can include exclusive contracts and other vertical exclusivities between the registry and the registrar to induce the provision of maybe these -- sort of think about narrowly focused specialized services. A subset of this concern but probably belonging from an economic perspective and this a part of the discussion is the use of vertical integration in contracts to facilitate the provision of what economists will call "promotional services." And so registries benefit from premium shelf space, for example, supplied by the registrar, in many product markets, again, not claiming to be an expert in this particular market, in many product markets, manufacturers will pay, will have to compensate through some sort of vertical contract the downstream firm to provide these services and so an example, which I'm familiar with and write about a good deal in my own work, is the example of slot allowances in supermarkets. Oftentimes in supermarkets what we observe is the upstream firms competing for, in that case, the eye--level shelf space or -- because I now have a 2-year-old child I've learned the second most expensive is the 2-year-old eye--level shelf space, competing for that shelf space because it's valuable to the upstream firm because it induces sales. And so we often see vertical relationships between the upstream firm, say Coca-Cola, competing for this eye--level shelf space and you'll see competition between the upstream firms for this space with different sorts of allowances paid to the downstream retailer which, from an economic perspective, provides some ultimate benefits to consumers, right? So you've got competition for the shelf space and then competition at the downstream level leads these payments to be passed on to consumers in different ways either resulting in higher prices, increased quality, they put in a nice deli, ATM in your supermarket because they've got this additional revenue stream to pass through to consumers. So in other markets we know that it often makes economic sense for these firms -- for the upstream firm to pay for -- to pay for provision of this shelf space and that the downstream firm does not have adequate incentive without these contracts to do so, and so we can get a good, healthy amount of competition for the premium-level shelf space that benefits consumers, and frequently we see these contracts involving exclusivity and other sorts of vertical restraints on the retailer or the manufacturer. Now, for these reasons -- and I've tried to do and set out what essentially is the economic state of play on vertical integration, and because of these pro-competitive benefits that economists and antitrust lawyers, courts, have recognized as well as the anticompetitive harms, in the United States, for example, antitrust law has evolved over the last 50 years towards being fairly permissive with respect to vertical restraint, so resale price maintenance, tying arrangements, exclusive dealing, vertical mergers, if you sort of go back in time 50 years in the antitrust laws in the U.S., in any event, most of these vertical arrangements were treated very harshly by U.S. antitrust law. In the last 50 years of development of economic knowledge about these restraints, the law has changed a great deal in the United States, not towards advocating, per se, legality of any of these restraints, but towards a fact-intensive, in the U.S. called a "rule of reason analysis," to determine the competitive effects, to figure out whether the conditions that support these pro-competitive stories are more likely to hold or whether we've got a story with factual predicates that support the anticompetitive concerns. A similar story can be told about EC antitrust law, which, although slightly less permissive than the U.S. as I understand it, is moving in a similar direction. If we consider, based on that general economic analysis of vertical integration, we can consider a couple of general principles about designing a rule to handle vertical separation sort of from a pure general abstract point. And we'll get more to application in a second. One principle that emerges about considering the effects of a prohibition on vertical separation from a purely economic standpoint is that, while we know vertical integration can cause these anti- competitive harms -- creation maintenance, exercise of market power -- in some sense a ban on vertical integration's under-inclusive, which is to say too permissive, right? Because you can achieve many of those same effects by contract rather than integration. On the other hand, the general economic analysis tells us that a ban -- a clean prohibition on vertical separation is also overly -- overinclusive in the sense that we know in many conditions vertical integration, in particular, where there aren't barriers to entry or no market power at either level is likely to involve one of these pro- competitive stories, one of these pro-competitive effects. So a complete blanket prohibition on vertical integration would come at a cost of denying those beneficial pro-competitive effects where they exist. And that's true even in more concentrated markets. That ends the sort of general economic analysis of vertical integration. And I'm going to let Steve cover things beyond the general economic principles that we hope are none too controversial and get into some things that maybe will be more. Steve? >>JOE SIMS: While --sorry. Forgot my instructions. While Steve is getting ready, just a -- just a caveat here. For those of you who know Steve or Googled him or want to Google him, you'll find that he is, in addition to being the Georgetown professor, also associated with Charles River, which is the same firm that Joe Farrell, the author of the so-called CRA report is associated with. Salop had nothing to do with that, since that's a firm with almost as many economists in it as my law firm has lawyers. And so he was -- he was not involved in that at all. But, since he is part of the same organization, at least on a consulting basis, I wanted to make sure everybody was aware of that. Thanks. >>STEVE SALOP: Okay. I want to begin to apply this to competition policy by ICANN. But, before I do, I just want to put in a little pitch. I'm an academic; so references, citations are everything. If you want to get deeper into the economic analysis of vertical integration, I have an article. It's in the slides. It's in the journal called the Antitrust Law Journal. If you can't get it online, just send me an e-mail. And I'll be happy to send you a reprint of this article that lays out these kind of pros and cons of vertical mergers that Josh laid out from this sort of so-called post-Chicago point of view that I've taken in my work. Now that the ad is done, let me talk about applying these ideas to ICANN. When Josh talks about stories, that's just the way we economists put it. What we mean by a story is a theory, either an anti-competitive theory or a pro-competitive theory. And that's the start. You have to get your framework. You have to get your logic down pat. But the next step is then to apply the theory to the facts, test the theory against the facts in order to see whether there's a likely problem or whether there are likely benefits. And so the point is that the facts really matter here. And the only way to make good rules by ICANN or the Federal Trade Commission or the telecommunications regulator or whatever is to look seriously at the facts. So where vertical integration, exclusionary vertical contracts can lead to benefits, they can lead to harms. They can lead to both and the benefits can be bigger than the harms or the harms can be bigger than the benefits. But the only way you're going to know that is to look seriously at the factual circumstances. A key fact or the key factor that we keep repeating is market power. And market power is about the intensity of competition. And that's both market power, the intensity of competition at the registry level and the intensity of competition at the registrar level. And, of course, the issue of competition at the registry infrastructure provider level will also become important, if that's what we're talking about. So let me talk a minute about market power a little more abstractly. You can have market power on the sell side as a seller or you can have market power on a buy side. A buyer can have market power. So for a seller, market power is the ability to raise price above the competitive level. That is, to be a monopolist. Microsoft has market power in the sale of operating systems. But there can also be market power on the sell side -- I'm sorry, the buy side. A buyer can have market power. So -- and there -- what market power is all about, we economists call that monopsony power. Market power on the buy side is the ability to profitably suppress prices below the competitive level, to pay less than the competitive price or to force subcompetitive terms on the people that you buy from. So in your context, the concern is that registries have market power, like dot com may have market power. And there's a reason why you have a price cap is because you're concerned that dot com is market power. In this vertical separation debate, there's been a concern that registrars have market power on the buy side, that registrars can extract coercive terms from the small registries. So that's a concern that's come through very clearly in the PIR and PIR material and in the conversations I've had with people on the floor. Now, intensity of competition, that's the opposite of market power. Where there's market power, competition will not be intense. Where there's no market power, competition will be intense. And that's going to be a key factor. Okay? Where there's no market power, where competition is intense, then there's unlikely to be a concern from vertical integration or vertical contracts. Now, it's possible under some circumstances that vertical integration, exclusionary vertical contracts, can create market power. But that usually starts from a foundation, from a base where there's at least some market power. And then it gets enhanced. It gets expanded by the vertical integration or the exclusionary vertical contracts. Okay? Where there's no market power, no prospect of market power, then vertical integration is unlikely to be much like a competitive concern. And, where there are no barriers to entry, then there's less likely to be a chance for market power. Because, if there are no barriers to entry, then when a firm begins to try to exercise market power, new entrants will come into the market and trump the exercise of market power. So that raises a really big issue, which is: Is there market power here? Now, let's look at the two sides. On the registry side, it certainly appears that there's a view that -- and I'm trying to be objective. But I guess I share this view -- that dot com has market power. Absent the price cap, there's a risk that dot com would be able to raise its registry prices perhaps by quite a lot. Okay? But -- and so you might say that well, there's a real concern about vertical integration by dot com. And you've got rules, been around for a long time, that says dot com can't vertically integrate. But what about with respect to the other registries? Well, there may be an issue about dot net. VeriSign runs both dot com and dot net. So whatever natural market power dot net might have, well, they might be able to share in the dot com market power. And, in fact, maybe VeriSign can evade the dot com price caps by messing around with dot net. So market power problem there, too. One that might warrant restraints on vertical integration. But these new TLDs, do you think they have market power? That's a question. That's a threshold question. Because, if you think they don't have market power, or a prospect of market power, then there's less -- there's less concern. And I'm going to come back to the trademark issue in a minute. On the buy side, there's a concern in material I've read that certain registrars, certain large registrars have market power, not necessarily on the sell side. I can't tell from reading the material whether there's a concern about the sell side. But there's certainly a concern that they will have -- that they have market power on the buy side. That they can coerce the registries, the small registries into, say, having the registries buy their registry infrastructure services from the powerful registrars. So there's a concern of buy side market power there. But, you know, is it plausible that the registrars really do have market power? That's a question. And I hope when we get the Q&A from the floor -- and I hope when the people talk, we'll talk about whether they believe that the registrars have market power and which registrars have market power. Is it all registrars that have market power? Is it just Go Daddy and NSI and ENOM? What's the question? Because it might be that you only want to restrict vertical integration by the registrars that have market power and by the registries that have market power. And let the other ones integrate in order to get the benefits that Josh was talking about. So this is an area where we need -- where we need information, where you need information in order to make a wise decision. You also have to separate -- and Josh alluded to this, but I want to go into it a little more deeply. You need to distinguish between market power that's going to lead to vertical -- a vertical merger -- vertical integration problem. That is, market power that will be facilitated by vertical integration, and market power that doesn't need vertical integration in order to be exercised. Let me give you two examples of that. And they're, I believe, both at the center of the debate that you're having. In the trademarks -- in this area of trademarks, I've read a lot that argue that trademark holders can be exploited by even small registries because the trademark holders have to engage in defensive registrations. Okay? And there's, of course, a debate over whether there's a big problem or a little problem. But that -- in economic terms it would say that even a small registry would have market power because the large trademark holders are -- have to buy registrations in order to prevent a trademark infringement. Okay. Well, assume that's true. Will a ban on vertical integration have any effect? The answer is no. That's a problem about market power. Those little registries that can abuse the trademark holders, they don't need to be vertically integrated in order to do it. The large trademark holders will seek them out and be willing to pay high prices in order to tie up those trademarks. So for them, if there is a market power problem, it has nothing to do with vertical integration. Vertical integration is not going to make it worse. Because they're going to be small in the end. Second, second issue that I want to raise, the PIR where I see concern that the large powerful registrars will be able to coerce the small registries into demanding that those registries give their infrastructure service business to the powerful registrars. And in the material the PIR laid out, they laid out five ways to gain the system, if you allow vertical integration between registrars into infrastructure services. Well, that's a potential problem. That's real. Because, if the registrars have market power, they can coerce the registries into letting them vertically integrate. And then, when they control registries, they can then raise barriers to entry to other registrars. That's sort of what shows up in my papers. But the ban -- banning them from engaging in vertical integration into registry services won't help. That was Josh's point about the rule being under-inclusive. Suppose I'm a registrar with market power. And I can't exploit my market power by forcing the registry to buy infrastructure services from me. Well, I don't -- so, you know, who cares? I'm just going to go to the registry and say, you know, if you want carriage, you're going to have to pay me something. You're going to have to compensate me. I know the price that you're charging for domain names is $5. You can charge that to everybody else. But you need to give me a lower price, because otherwise I'm not going to carry you. Or, jeez, maybe -- maybe that's not permissible because the registries are forced to give the same deal to everybody. Then they'll say, well, if you want carriage on my site, so happens I offer advertising on my site. And, if you want me to carry you, you better buy a lot of advertising from me. And, by the way, my advertising is really expensive. Well, if there's a market power problem and if you don't regulate the registrars, they can do that. They don't need to go into infrastructure services. So in that situation, maybe vertical -- maybe banning vertical mergers, vertical integration will help but maybe it won't help very much. So you need to confront that market power problem as well. Okay. Next issue price caps. Josh talked about that a lot. Where a registry has market power, there's even greater anti- competitive incentive to vertically integrate if you also have price caps. You know, if dot com can't raise its price of domain names because there's a price cap, well, what they could do is go into the registrar business and say, well, if you want to buy -- you know, if you -- you're going to have to pay a lot for me for hosting. And they can extract the profits into so-called unregulated division. So the price caps are really important. But, again, that's a problem for price caps and market power. If the registries do not have market power and they're not subject to price caps, then that's not going to be a concern. So you need to know market power. You need to know whether there are price caps. And, again, I want to stress this is a problem for dot com, dot net. But not for -- less so, it seems to me, to a -- to a TLD that lacks market power. And next, magnitude of benefits. Josh talked a lot about that. Some vertical integration, some vertical contracts lead to significant economic benefits. Some lead to no economic benefits. In my consulting practice, you know, sometimes firms make claims of efficiency benefits from vertical mergers or from vertical contracts that are just pretextual. They just come up with -- some economist comes up with a story. Well, you have to test their claimed benefits against the facts. See whether it's plausible. See whether it actually holds up to analysis. Some do, some don't. And it varies. So you need to analyze that in order to come up with good rules. And, finally, there's the issue of the restrictiveness of the conduct. There's full vertical integration, you know, 100% cross- ownership. But then there's partial cross-ownership. Partial cross- ownership often leads to less competitive harm than full cross- ownership. In cross-ownership there's having a pure equity interest, you know, just having a financial interest. And there's control, governance over the entity. Where there's control, it raises greater problems and actually leads -- could lead to greater efficiency benefits than where the ownership is just a pure passive equity interest. The more restrictive, the greater the competitive risks. Okay. So on all of these you need to know the facts. Okay? Knowing all the facts is key. And the facts means market power, competitive dynamics in the markets, claimed pro-competitive rationales and a thorough analysis of those issues. What are the implications of that? Well the implications is there are no easy answers. Simplistic rules make mistakes. Okay? And so economists have a general bias against blanket prohibitions. If the facts matter, then a blanket prohibition, a blanket rule regardless of the circumstances is likely to make mistakes, likely to be on balance overly restrictive. Because it brings in -- it has the -- it applies to prohibitions to firms that have market power, that have price caps. But also to the firms that have no chance of getting market power. The rules apply to situations where there are large benefits from vertical integration and it also sweeps in situations where there are no benefits from vertical -- where there are big benefits from vertical integration and little benefits. So doomed to make mistakes a lot of the time. It's better to have standards that depend on the facts. Okay? The same time, complete laissez-faire is going to tend to be too permissive. Laissez-faire is good if there's no chance of market power, but laissez-faire is not good if there is market power. So blanket rules, whether it's what we in antitrust called per se illegality or blanket rules, we used to call per se legality, both are going to tend to make errors. Better to have case-by-case analysis. Okay, you look at the law to apply the sort of law generally, some areas of law have blanket rules, others have case-by-case analysis. So take the law against speeding in your automobile. That tends to be a blanket rule. In the U.S., 55 miles an hour, that's it. It's not case by case. When I get pulled over on the interstate, I can't say to the state trooper, "Look, I've got a BMW. I went to driving school. I've got fabulous brakes. I can go from 80 back to 0 in 110 feet." He's going to say, "Come on, 55." It turns out Montana used to have a rule that said no speed limit. You can't drive unreasonably fast. You see, you can, even for speeding have case-by-case standards. But usually in speeding, it's a blanket rule. Okay? But in other areas, like antitrust, it's not a blanket rule. It's case-by-case analysis. You look at the facts and you decide on balance, given all the facts, is this vertical integration proposal more likely to cause harm or more likely to cause benefits? And those are -- those sorts of standards tend to be what gets used. Because the blanket rules lead to -- too often lead to mistakes. Okay? Finally -- you know, two other points I put in sort of to state the obvious. You should only adopt rules that you intend to enforce. The rules aren't going to have their intended effect unless they're enforced. And, you know, I -- take sort of your concern with vertical integration, your rules say that the registries can't own registrars. But, as I read it, it looks like the registrars can own registries. Well, that's not going to work. That door swings both ways. Okay, so you're going to have to enforce the rules that you -- that you intend to promulgate. The other rule -- it's an obvious point and I've heard it, sort of this is a bullet that I did not have on my deck until I came here and talked to people, which is that changes in the status quo can create long-term benefits. But they -- you know, invariably, are going to cause some short-term transition costs. I worked at the U.S. Civil Aeronautics Board during the period when airlines went from regulation to deregulation. There was a lot of resistance by the airlines to deregulation. Some of the airlines feared competition. They liked it. They were living the soft life. High prices, regulated by the Civil Aeronautics Board. No fear of entry into their markets. The Civil Aeronautics Board kept entry out. They didn't need to compete. If one of them had a desire to compete by giving you a better, you know, a bigger seat or seat that went back further or give you a better sandwich, the Civil Aeronautics Board would step in and prevent that competition. So they liked the cartelization by government regulation. Other airlines, it wasn't so much they wanted to cartelize, it's just deregulation, markets are -- market's a little chaotic, you have to work harder in markets, they're dynamic, there's change, and they feared that, it wasn't they didn't want to do it, they just feared it because it was change. Now, I think pretty much everybody believes that airline deregulation was a good idea. Drove -- took a couple of years but prices have come down a lot. If anything, airline deregulation suffers from success. Too many people are flying, the planes are too crowded, the airlines aren't making enough money, they wish they made more money. But there's been innovation. Consumers benefit from low prices. So I think -- you know, I think the same issues apply to your industry. Your world is one that is involving dramatic growth, innovation and opportunity. Seems to me that it's important to support that innovation by keeping the rules flexible. It seems pretty early in the history of your industry to cement broad prohibitions in place in the business structure. And if you go with greater flexibility, if you permit change, it's going to involve some transition costs, for sure, but I think it's also going to lead to significant consumer benefits so thank you. [ Applause ] >>JOE SIMS: Okay, so much for what we hoped was a useful survey of the economics and a little bit of the law of vertical integration. Now we'd like to participate in a discussion with anyone who wants to participate. We have several microphones scattered around and you should go to the microphone, I think you would be useful for you to introduce yourself and then ask your question and then hopefully one of these two guys will answer it. >>NAOMASA MARUYAMA: Okay, my name is Naomasa Maruyama. I'm not sure, this is -- my question is a good -- this is a good place to ask my question here, but I, in the past I -- one effect, good side effect of this vertical separation of a registry and registrar, that is the -- it's actually preventing the so-called corporate TLD, that is the TLD solely used by one corporation. I used to use a word "proprietary TLD" for this. So I think it's very important to discuss this issue with the discussion of the policy-making about the corporate TLD. Whether the ICANN will permit corporate TLD or not. Actually, one comment for the guidebook by the Bank of America, I think, raised this issue. So if this discussion probably, the two professors discussed about the economic effects of the vertical separation or the integration for the market of -- domain names supplying market, I think, but if we gave up the jurisdiction of the vertical separation for this issue, then probably some corporate will apply for their own TLD and that cause the domain name, this issue, that will cause this issue into one TLD, bringing from the domain name supplying market to another market, that is the corporate advertising market. So my question is: Where this issue will be addressed in this ICANN community? Is this -- this workshop is the place to discuss this issue or not? And if not, where is the place to discuss this issue. >>JOE SIMS: I think the answer to your first question is this probably isn't the right forum, and I'm quite frankly not educated enough on what the ICANN -- where in the ICANN framework and timetable this all applies so hopefully you could ask one of these folks sitting up in the front and they would be able to give you an answer to that question. Paul -- Paul can give you an answer to that question. >>PAUL TWOMEY: Sorry, Joe, but I think this is the place to discuss the subject. Sorry. >>JOE SIMS: Then I must misunderstand the question. >>PAUL TWOMEY: If I may, I think the question being raised is Corporation A applies for a TLD and basically proposes that it be both the registry and the registrar. Now, there is some discussion that -- being discussed in this area that it comes down to what are the rules. If it was only to sell it to its own staff or only use it for its own purpose, how does that fit your set of rules, that may not be fully what you were saying but I think that's one of the propositions that might be considered or should there be a number limitation. Obviously, there's a -- if I could -- put some real consensus on the table, there could be some very large Internet companies that could register in their name, be registry and registrar and then have millions of domains very quickly. That's very different than if it's a -- how do I put it? If it was a consumer goods company and its intent was simply to use the second level to promote the brands of its products in advertising. So you can imagine it could be company X which would have, say, 3 or 5,000 second levels but it could be company-wide that could have several million very quickly. I think that's at the heart of the question. >>JOE SIMS: And just so I understand, is the concern there that if -- if it was the latter, the several million company, that that would be foreclosing other registrars from competing for those registrations? >>PAUL TWOMEY: I think that's one way of putting it, yes. >>JOE SIMS: Okay. >>JOSHUA WRIGHT: Right, so I think this does fit into the general economic framework relatively nicely. The question in these instances for the anticompetitive concerns for these harmful effects that I think you have in mind turn on the same question about the existence of market power on one level or another. Now, from the way I understand the question, that would be the relevant analysis for whether this, I think he used the word proprietary owner, would have market power. It's -- at the registry level it strikes me that unless there's some reason to believe that there's significant market power with that entry then the general economic analysis is that entry is good. >>NAOMASA MARUYAMA: Yeah, it's very interesting to hear the specialists in the economy about this aspect but also I think it's very important to address this issue from the aspect of the root zone scalability. So this is -- problem is very important, I think, that is my comment. Thank you very much. >>JOE SIMS: And that I'm -- that issue I'm sure is beyond the competence of this group. [ Laughter ] >>STEVE SALOP: Yeah, I was just going to say, you know, if you go back to the benefits and harms, a proprietary registry may have -- where they act as their own registrar, that may have significant economic benefits, they get better control, better security, there may be no market power harms and so our analysis, if that were true, would say it's a good thing, not a bad thing. >>NAOMASA MARUYAMA: So the serious problem is that there's no policy discussion about this proprietary TLD, that is probably our problem, thank you very much. >>ROB HALL: Rob Hall with the Momentous Group, we own a lot of registrars I guess would be one way to put it. >>JOE SIMS: You own everything. >>ROB HALL: Oh, yeah. Working on it. My question I think is perhaps one of semantics. We talk a lot about registries and registrars. I think there's a new entity being created as we watch this process go on called -- what I'll call registry operators, and I'm wondering if your analysis -- if you could explain where is the real danger -- and perhaps it's both -- but if we take the example of PIR being the registry, a nonprofit group, and Afilias being a registrar operator for them with no cross-ownership that I know of between the two, is the analysis the contracting party with ICANN should be prohibited from being a registrar, is it the registry operator that may be operating many registries? And I wish we'd start using different terms to distinguish these things because I think it often gets blurred. >>STEVE SALOP: Well, I think that's a problem. You know, I mean, if the registrar owns a industry infrastructure provider and that registry infrastructure provider controls the industry. >>ROB HALL: But my suggestion is if it doesn't, like where are the lines that we're trying to prevent. So if the registry that contracts with ICANN is in no way owned, is it okay to have a registry operator that's owned by a registrar or vice versa? >>STEVE SALOP: I think what you're worried about is the maintenance and exercise of market power. So if there's robust competition among registrars and robust competition among registry infrastructure providers and robust competition among registries. Then having one own the other is not a competitive problem. It's where you have market power, and that market power can be used to create or enhance market power by raising barriers to industry excluding competitors at another level that the problem comes in. >>JOE SIMS: So Rob, if you think about it this way, let's think about a registry called dot Salop that just gets started tomorrow and a registrar run by Josh Wright that just gets started tomorrow, and then Afilias as a registry operator, right? Now, if Josh Wright, the registrar, who just got started tomorrow, buys Afilias, right, Josh Wright doesn't have any market power in the registrar market, he just opened his business, right? Salop, dot Salop doesn't have any market power in the registry business unless you think dot Salop is its own market and he's a monopolist and that's a different discussion. But as a new registry, doesn't have obvious market power. So if Afilias is just a registry operator and that's all they are and they don't do anything else, why would there be a competitive problem if Wright, the brand new registrar, bought them? So now Wright's in both the registrar business and in the registry operations business but there's no competitive problem that's obvious from those facts. Now, if, instead of Wright buying it, it's Go Daddy or, you know, NSI back when they had 100% of -- then you'd have a very different analysis and so that's what these guys are trying to say is it all depends on the facts on the ground of the individual players and it's very hard to come up with rules that apply to all facts because the facts, by definition, are going to differ. >>ROB HALL: And I agree with that analysis, Joe, I just worry that we talk about registries and registrars, and there's actually this third entity that's going to play an important role that we may want to start discussing. >>STEVE SALOP: Let me play around with Joe's hypothetical for a minute, take this a little further. Suppose I'm a registrar, I'm a small -- I'm a small registrar and I come up with an innovation in registry operations. And I go and I go to registries and I say, "You know, I'd like to do your back-office operations, and I'm really good at it." Well, that would create competition for Afilias, against Afilias, that's not any competitive, that's pro-competitive, that's a good thing, okay? And if it turns out I'm not -- that I'm not a small registrar but I'm a big registrar and I come up with a better way of doing back-office and I go and compete with Afilias, again, that's a good thing, that's not a bad thing. It's only a bad thing if it generates market power. So let me give you an example of that. I'm a powerful registrar, Go Daddy, say, and I come into a registry back office and I go to a registry and I say, you know, "use me or -- use me for your back office or I won't carry you," and they cave in. And as a result -- you know, I do that with a lot of registries. And I knock Afilias out of business and I knock the other providers out of business. Then I become monopolist in providing registry back office. And I use that by -- I can use that to create barriers to entry into registrars. I can use it to, you know, prevent other registrars from getting started. That's where the problem comes in. But that's a complicated story, I mean, there has to be potential for monopoly at each of the levels. >>ROB HALL: So perhaps the element therein is how do you write rules for this? >>STEVE SALOP: What do I think? Antitrust, case-by-case. You don't have blanket prohibitions, you look at the facts, and either you or the antitrust authorities say no, this one doesn't go. >>ROB HALL: So you would typically have to build into our contracts, then, at each level, the ability to review this at some point; is that right? >>JOE SIMS: The challenge here, Rob, this is the real nub of the debate that's going on inside ICANN right now, almost every economist would give the answer Steve just gave to you. Fact -- case-by-case, facts. Well, I mean, case-by-case, facts, that's complicated. That's, A, hard, got to get the data, right? I've got to get all the information from all the players and maybe even from all the other players in the market so that I can actually judge whether somebody has market power or not. Then I got to actually analyze that stuff which means I've got to have some skill at analyzing this that's a skill which not every human has and so you've got to buy those resources and then you actually have to enforce those rules, right, which may or may not be all that idea to do. So case by case, that's really complicated, right? So that's the way most antitrust works out in the world, it's case-by-case and it's really complicated and that's why you have specialized agencies that do that work. >>ROB HALL: But you're arguing you're only dealing with the fringes, the large players, if you will, the ones that might have market power not the -- >>JOE SIMS: That's right. >>STEVE SALOP: That's what "case-by-case" means, you only worry about situations in which there is a problem. Dot net is different than dot Salop and, you know, for quite awhile at least that's what's going to be the case. >>JOE SIMS: And so what a lot of people in ICANN would like to do is they'd like to say, well, look, let's come up with some rules that we know are not going to be perfect, they're not going to work exactly right in every situation but they're going to be good enough to protect against the harms that we think might be out there in some specific fact situations and they're not going to be too restrictive so that we don't lose all the benefits of all this potential vertical integration. That's what, you know, a lot of people inside the ICANN community are talking about, trying to find that sort of, you know, middle ground. It's a really hard thing to do. >>ROB HALL: You'd have to allow for exception for review of it, I would think, still. >>KURT PRITZ: Joe, I just want -- Hi, I'm Kurt Pritz -- make an announcement. The session's scheduled to end in about 15 minutes. We can extend it if you fellows can. The facilities are available. >>JOE SIMS: We're happy to go as long as you want. >>KURT PRITZ: But be careful about saying "going as long as you want." [ Laughter ] And so given that there are ten people at the microphone, Joe, you know, you'll use your judgment as far as moderating. >>JOE SIMS: Well, I've got another meeting that I've got to go to by 3:00 so we've got to stop by then. >>JEFF NEUMAN: Hi, my name is Jeff Neuman, I'm with NeuStar, we're the current registry for dot biz and dot us, but we're also the technical registry operator for dot travel and dot tel. And I'm glad, actually, to piggyback on Rob's comment. I'm glad -- this is the first time that the whole subject of a technical back-end registry operator has ever been brought up. All of the economic studies that have been done before has blanket just said registry and registrar and, in fact, there are also other variations of registrar, there's resellers which is a whole other market so, for example, if a registry can't be a registrar in theory second be -- they could be a reseller and still exert that same type of market power. The other point I want to make and I'm glad we're talking about facts because, you know, your discussions was all on the academic level. But today, you know, given that new TLDs are coming in a few months, you know, we're debating this issue but what's going on in the market today is that there is a market for registry operators, and it's not just the existing ones, it's not just Afilias, NeuStar, VeriSign, it's a number of them that you see out there with booths and some of them have pretty nice ones with famous chefs. But there's -- there's a lot of them out there. And the things that -- that's going on in the marketplace today, in actual fact, is that if a new gTLD operator registry, the one that wants to hold the contract with ICANN is going and shopping around and they're getting some of these affiliated back-end operators saying, look, if you use someone else to be your back end operator and don't use me then I'm going to charge you all these fees to have my affiliated registrar carry your TLD. They're saying if you use Afilias, let's say, then we will charge you a couple hundred thousand dollars a month, we want an equity stake in your company and we want a certain fee for every domain name sold regardless of whether it's our registrar that sells it or not. This is going on today, this is not made up, this is going on and this is something that if we're going to have this new round that starts in four-month and we're getting into this academic debate people are spending some real money right now on applying for new gTLDs as a registry, as China competes for a back-end registry operator, and I'm just worried that the fact that this is all going on today and we're talking this academic sense, I mean, you know, investors are -- we're not sure of what to do. >>JOE SIMS: It's a fair comment. It's not the first time that comment has been made at an ICANN meeting about ICANN moving slower than the market. But, I mean, I understand the point. It's a very valid comment. What we've -- what we're trying to do here today is to facilitate a further discussion inside of ICANN so that they can come to an answer because I think everybody inside the staff and among the community, the most important thing for you is let's get to an answer, right? Now, you'd rather have the right answer than the wrong answer, obviously, but you've got to have an answer because you've got a business to run so that's what we're hoping this will lead to. >>RICHARD TINDAL: I'm Richard Tindal from ENOM. We agree, we think it's about market power and we think that's a pretty rare occurrence in this city so let me jump to something that Jeff just said. We're the number two registrar. We're competing to provide back-end registry services to new gTLD applicants. And it's a very a competitive market. There's a group outside that Jeff just mentioned, Mind Over Machines who have no registrar capability and they're competing very effectively with us. So I think the occasions of market power are pretty rare. I just wanted to play back what I thought I heard as the three potential anticompetitive effects in sort of general theory. Evade price caps, exclude rivals, and exchange sensitive information. So in the context of the new draft applicant guidebook, we think all three of those potential problems are adequately covered. First, there are not price caps on the TLDs, so we don't think that's an issue. There's an equal and access -- equal and open access provision to registrars, so you can't be the only registrar, so we think that is covered. And the data that the registry gets is public data. So we're not really talking about a registry getting access to sensitive information. So just a couple of thoughts, quickly, about market power again. So we would like to apply for TLDs ourselves, let's say we apply for dot cool. We think that dot cool's going to compete with dot com and dot net and, you know, we don't think we're going to have market power in that sense and even if you believe that we're a market unto ourself, we think there could be a dot very cool or a dot extra cool or a dot too cool so we think those are going to compete with each other, we don't think we're going to have market power in our TLD. And then finally, market power as a registrar, we are the number two registrar by volume. I've got to tell you, it doesn't feel like we've got much market power at all, it's hypercompetitive, margins are very, very thin, we're always looking at ways to innovate in terms of services and quality. So even as the number two registrar, we're not feeling that powerful. >>JOE SIMS: Just one factual point. Five years ago, were you the number two registrar? >>RICHARD TINDAL: I think Paul was in his garage. >>JOE SIMS: Well, yeah. So it's just one of the indicia of a competitive market and it certainly looks like the registrar market has become a very competitive market as you see changes in the sort of standing of various people. The number one, two, three, four guys don't stay the number one, two, three, four guys, when you see them staying the same, sometimes that's an indication the market isn't very competitive. >>STEVE SALOP: Let me ask you some questions about your lack of market power. Do you charge prices that are higher or lower than your major competitors? >>PAUL STAHURA: Paul Stahura, and the founder of ENOM. I can answer that one. Lower. >>STEVE SALOP: Do you keep customers over a long period of time or what fraction of your customers do you lose every year and what fraction -- what's your won/loss rate for existing customers? >>PAUL STAHURA: It's probably 75% renewal. >>STEVE SALOP: So you lose 25%. And what fraction of your customer - - what fraction -- what's your growth rate? I mean, what's your new customer-to-existing-customer ratio? >>PAUL STAHURA: I don't know the number off the top of my head but I know that our registration base is growing at about 10% a year, this is all public information. >>STEVE SALOP: I'm raising these, because these are -- this is some of -- some of the facts that you'd like to know. >>PAUL STAHURA: I'm with Richard, I wish I had market power. >>STEVE SALOP: Do your customers feel locked into you, that once they buy from you for a couple years they can't move? >>PAUL STAHURA: No, because I, I guess. [ Laughter ] They may feel locked in because we provide, you know, superior service. [ Laughter ] >>STEVE SALOP: No, I mean registrants, registrants, as opposed to -- do the registrants -- >>PAUL STAHURA: No, because it's very liquid. Registrants could, you know, easily move, you know, ICANN has been pretty good with implementing a transfers from one registrar to another one and it's really liquid, registrants move in and out all the time. And one other thing about Jeff, what Jeff says, was, you know, there might be some registrars who say, hey, if you use my registrar back end for your TLD, you know, if you don't use it, I'm going to punish you by not offering your TLD on my platform. A, I never thought of that, even doing that because I don't think I have the market, I don't think I have the power to do that, now apparently some registrars are doing that, I might think about that, because there may be some, what I would guess is, you know, you know, gullible registries out there that might take me up on that deal. What they should say is just say no and call their bluff. So if the proposed, you know, if this potential registry said no, I'm not going to take you up on that deal, I would bet you that that registrar would still offer that TLD. >>JOE SIMS: Thank you. >>STEVE DELBIANCO: Thank you. Steve Delbianco with Net Choice, and also a member of the business constituency here at ICANN. And I'd like to bring the perspective of registrants, those who invest in building and developing a second-level domain name or maybe a whole collection of domain names that they comprise in a business, because I know, Josh, you've mentioned the importance of exclusive or preferential vertical contracts, but from a registrant's perspective, we are also impacted by, well, good old-fashioned spite and rivalry. Even if it isn't power at all. For instance, an example where our registrar might deny shelf space to an entire top-level domain because they don't want to enrich a competing registrar who's going to make money off of rental registrations in that TLD. It's not just a hypothetical example. I'll give you a real example with the dot ME top-level domain. It's a Go Daddy project and I've noticed that at least a couple of large rival registrars, One in One, for instance, and Melbourne IT, don't even show dot ME for sale, it's not on the shelf at the two-foot level or 10-foot level. If I may, a registrant in dot ME, that's suppressing the acquisition of other ME to second-level domains and it's probably also going to affect my traffic and if I'm trying to build a social networking service and I want everybody to get a name dot ME I'm going to have to try to find ways to get those in people's hands because they're not on everyone's shelves. So it's not always about market power, sometimes it's about market performance, particularly if ICANN wants to see these new gTLDs succeed. Remember the network effects, another economic concept I didn't hear at all today network effects would say that efforts to suppress the availability and distribution of a TLD are going to suppress the ability of that TLD to be successful. >>JOSHUA WRIGHT: Just a brief comment. I don't want to suggest from my presentation on the economic benefits of preferential contracts that this means that these abuses aren't real or don't exist. In fact, you're in a much better position to know that they exist than I. But I think, as a matter of the economics, the -- there's a pretty strong link in economics or most economists would say between market power, competitive intensity, as Steve said, and market performance. This doesn't mean that you don't get these sorts of abuses as part of the competitive process. But it does mean, in terms of the general economic analysis, that the existence of those abuses does not follow from the existence of those abuses favoring a blanket prohibition. You might want to think of sort of these facts matter type rules to target out those instances without losing the harms. That's the trickier -- that's the trickier nut to crack, if you will. >>STEVE SALOP: But I want to go a little further on this. And that is that -- take search, Internet search -- Google, Yahoo being -- you know, there's certainly big network effects there. But I don't think that the antitrust authorities would ever consider a rule that required every Web site to have search bars for 150 search engines. I mean, you let -- you tend to let the market work. And not every search engine gets on every Web site. You allow competition there, rather than having a planned economy. >>STEVE DEL BIANCO: And this is not a matter of whether the antitrust authorities are concerned with the market performance of a new TLD. It's how ICANN and our community are concerned about it. And I appreciate that you guys are doing the panel. That's helpful. But I would like to ask ICANN staff at some point to tell us when will we next talk about this important topic? After staff, for instance, has developed some guidelines. I would look forward to more public and open discussions of this before we finalize the new TLD contracts. Thank you. Mike? >>MIKE PALAGE: Thank you. My name is Mike Palage, long time participant within the ICANN process, currently participate within the registry business and the IP constituency. One of the first questions that I'd like to talk about is during the -- I believe it was the CRA report, the presentation, you were talking about the need for get good facts to get good rules. One of the questions I have is over the years, there have been a number -- there's been increased litigation involving certain actions by registrars. Cybersquatting, massive cybersquatting in a number of instances. There also have been a number of reports by ICANN by the security and stability advisory committee regarding frontrunning by registrars. Now, all of these issues are actions harm end users or registrants. And, in reading all the reports, I have not seen any analysis of these potential litigations or other reports where there have been alleged misconducts by registrars that have harmed end registrants. So, getting back to good facts make good rules, have -- has this been considered and do you think these are facts that need to be incorporated into your equation? >>JOE SIMS: I don't see why that has anything to do with this subject. That's a subject that may well -- maybe it deserves attention. But what does it have to do with vertical integration between registrars and registries? >>MIKE PALAGE: I think -- >>JOE SIMS: Pardon me? >>ERIK HUIZER: That's obvious. >>JOE SIMS: It's not obvious to me. Explain it to me. Go ahead. >>ERIK HUIZER: Now you have a registry. Like I have this brilliant idea. The market is created by the registrants, right? They create the product. I have this brilliant idea. I'm now sure that if I tested the domain name that belongs is available, I tested with the registry sure to be neutral place. If there's a vertical integrated registry, registrar, I'll test it. It might all be that all of a sudden this name can be sold to a company and I can only get it through an objection. Front running makes it easier -- it's easier, if there's vertical integration. >>JOE SIMS: Of course that assumes that everybody is vertically integrated right? >>ERIK HUIZER: No. >>JOE SIMS: Don't test it with the guy who is vertically integrated, if you're concerned about that. >>ERIK HUIZER: Where do I test it? >>JOE SIMS: Well, you're assuming everybody is vertically integrated. >>ERIK HUIZER: Does the end user know everything like that? It's impossible. >>JOE SIMS: We can only hear people that are on the microphone. So you'll have to stick with Palage for the moment. >>MICHAEL PALAGE: That can be a bad thing. So, Joe, to address your concern here. One of the issues that was really a systemic problem with the broader gTLD community a number of years ago was tasting. Right? And there were a number of people in the registrar side of the market that were, if you will, profiting off of that. And it was the registries that stepped forward. It was PIR with their initial funnel request to halt it. Then NeuStar and -- then it was NeuStar and Afilias that came forward. And it was that momentum by the registries that helped pave the way for a PDP that resolved really a systemic problem that was really a problem. So in that scenario, I think that segregation, vertical segregation where the registries were actually able to step in and prevent a harm was an important factor. So, again, I'm not going to get into an argument with you. I've learned, growing up in Philadelphia, I'm not going to bring a knife to a gun fight with an economist or antitrust attorney. So I'm going to keep this real simple. That's just my analysis that I think that separation does serve a function in protecting the end registrants. A second point -- and I had raised this with general counsel during the briefing session with the registries. ICANN operates in a global marketplace. Unfortunately, the economists and the -- the data points that we've looked at to date have primarily been U.S.-centric. The lack of sort of other competition authorities, the DG4, you know, looking at the global marketplace in which ICANN operates, I think that that, again, goes to this issue of good facts, good rules. And if we're only looking at U.S. rules as opposed to global rules, I think, again, that's my opinion that that's sort of a noncomplete puzzle. >>JOE SIMS: Well, look, that's a fair point in general. I don't think it applies to this discussion because these economics are not bounded by national boundaries. These economics work everyplace. They don't work just in the United States. >>MICHAEL PALAGE: Okay. >>JOE SIMS: These are economic principles that I think you would find accepted by Polish economists and Chinese economists as well as American economists. >>MICHAEL PALAGE: Okay. Just one final point. As far as, again, going back to good facts, good rules. One of the things I've written extensively on this issue regarding the economic analysis, was the original CRA report there was discussion by ICANN staff that would be available in June of 2008. And then all of a sudden it got delayed. And then we had the final report. With the Charles River we had the preliminary and then the revised. With CRA, we kind of never got the preliminary, never saw it. Then we got a final and nothing else. Is there any problem with making that original draft available? >>JOE SIMS: You're asking the wrong guy. Go talk to John Jeffrey. >>MICHAEL PALAGE: I'll send him an e-mail. Thank you. >>BECKY BURR: Becky Burr, WilmerHale. Two quick points, one on the question of case-by-case analysis. What do you do when the sort of facts on the ground are changing as quickly as they seem to be? The data in the -- in the CRA report was from February of 2008. And with the top four registrars, it was 50% of the market. It's now 58% a year later of top 4. Went from 63 to 74 with the top 8. So the facts on the ground are moving pretty quickly. And that seems to me to be an additional gotcha in the case-by-case analysis. The other -- >>STEVE SALOP: Why don't we take these up one at a time. First of all, it's C-R-A, not CRA. >>BECKY BURR: Sorry. It's the ICANN -- >>STEVE SALOP: Although I didn't write the report, it is the company I'm associated with. When the facts on the ground are changing, you need to do it case by case. You don't blanket prohibitions when the facts on the ground are changing. It goes the opposite way. >>BECKY BURR: I completely agree with you. But, if we did a case- by-case analysis, it won't be until another year that any new TLDs are around. So it's going to -- I mean, I'm completely agreeing with you. It's just much harder. >>STEVE SALOP: I agree, yes, it would be much harder. So you need to be careful that you don't build in flexibilities into the system. >>JOE SIMS: In dynamic industries, which this obviously is one, antitrust is much harder than it is in dealing with, you know, like the steel business, right? Things do change fast. They're changing a lot. So what you're trying to do when you do antitrust analysis and what you're trying do to do, when you do economic analysis, you're trying to figure out what is the future going to look like given these various approaches to things. If I do this, what effect will that have? In a dynamic industry it's a lot harder to tell the answer to that. Your point is a fair point. If you take a snapshot at day 1, day 30 the snapshot may be different. And if you set a rule based on day 1, day 30 that rule may not be any good any more. This is a serious problem for ICANN generally, because ICANN is an organization that doesn't work fast. I mean, that's almost by definition. So I don't know that I have an answer for you. >>BECKY BURR: I don't think it that there is an answer. It's an observation. >>JOE SIMS: But I appreciate that it's a serious problem. The other point I want to make is about the changes in status quo producing long term benefits and also short-term transition costs. I understand that the notion there is that there may be harm to competitors but -- to competitors but not to consumers in that situation. But what do you do in a situation where we have the -- a structural rule, a -- basically, a rule of ICANN has put players in the market in different places? And so they -- there's not a level playing field. And people have invested for 10 years based on a certain kind of rule. The other thing that you add to that is, when you deregulated the airlines, there wasn't the prospect of a thousand new airlines the next day. So, although I read your paper on the way down and I think I understand the difference between, you know, the need for harm to competition and consumers, you have this structural disadvantage that has been built in for the smaller registries that needs to be dealt with at some level. >>JOE SIMS: You mean the smaller existing registries. >>BECKY BURR: The smaller existing registries. >>JOE SIMS: Who are operating under the vertical separation rule. >>STEVE SALOP: Yeah, I think that's right. It's not clear to me as I sit here that they should be -- that they need the, say, the price caps and vertical separation rules. That you'd want to do an analysis of that as well. >>BECKY BURR: Right. I think there's just another -- on the other side of it, for the past 10 years, they haven't been building a registrar base. >>STEVE SALOP: Right. >>BECKY BURR: Because they haven't been permitted to. So I don't know what the answer is. I'm just telling you that that short- term/long-term thing is complicated by the facts on the ground again in this situation. >>STEVE SALOP: Right. But, again, I'm going to -- I'm going to state this as a user, not as someone in the community. I mean, you're working for us users. We're not working for you. The rules that ICANN can make should be the ones that are best for the development of the Internet. And, you know, if it harms some firms and benefits others, well that's the way competition is. That's the way dynamic competition is. >>BECKY BURR: So are there situations where -- that would be -- I guess the question is, are there situations where, in fact, there is a structural rule that is the creation not of the market but of a regulator that changes the next day and the people who have been regulated in one way are differently situated from the next, I mean, it's not the airlines. >>STEVE SALOP: That always happens, yes, it is the airlines. You have United Airlines, American Airlines saddled with higher legacy costs. And they were disadvantaged relative to the new carriers that didn't have union contracts. American Airlines, United Airlines had the wrong sized planes for the deregulated market because with deregulation the airlines started to have these hub-and-spoke systems where the smaller planes made more sense. The older legacy airlines had airplanes that were higher cost, you know, less efficient users of fuel. That's always true in the economy. >>JOE SIMS: You know, that's the right economic answer. That's the right -- not necessarily the right ICANN answer. I mean, this is -- we're not trying to be up here to state ICANN policy. Your point is that there are likely -- if you make this sudden shift, there are likely to be some dislocations, right? And those dislocations are going to impact some people worse than others. >>BECKY BURR: And the impact is going to be the product -- the differentiated impact is the product of an ICANN rule. >>JOE SIMS: Yeah. And that's all, I think, probably accurate and a fair point. And now it's a policy issue for ICANN to determine what, if anything, to do about that. >>BECKY BURR: And my question was just are there examples in the world that we can look to for that kind of stuff? We can talk about that afterwards. >>JOE SIMS: Yes, sir. >>ADAM PALMER: Thank you. Adam Palmer. I am policy counsel at dot org, the public interest registry. Before I ask my question, though, I wanted to thank you for having read the memo that we produced and that I was instrumental in drafting for PIR. And I wanted to emphasize or just bring to your attention that the emphasis of that document was our support for competition, support for a fair playing field for new gTLDs and the protection of the 10 years of progress that have been made in promoting competition within the ICANN community and our concern about a sort of back to the future academic approach that might reverse the progress that's been made in competition. With that said, I did want to make one point, bringing out the CRA. I was about to call it CRA. The CRA report that sort of had the notion that registries might -- or might perhaps play a bottleneck role in the community and roll out of new gTLDs. But, really, I think it's fair to say that the opposite could be true. That, in fact, right now you have a registrar that's free to place or not place any registry that it might desire to do so. You also have a registry or rather a registrar that might be able to be a back end service provider. And those cases we talked about the abuses. And I think those need to be emphasized that, if you're a new gTLD applicant and the registrar is a back end service provider and has the power to force you or compel you that, if you want shelf space, that you have to use them as your distribution channel, that could be a significant harm to small new gTLD applicants or to small service providers. >>JOE SIMS: Adam, can I just ask a question? What's the source of that power? How do they make that happen? >>ADAM PALMER: Through market force by demanding -- compelling them to either use their services or risk losing that critical shelf space -- >>STEVE SALOP: He's asking you how do you know that registrars have market power? What's your evidence that registrars have market power? >>ADAM PALMER: I think it's quite clear from the data and inherent within the system -- >>STEVE SALOP: What is the data? >>ADAM PALMER: The relative, the amount of the distribution channels, the degree to which one might control the distribution within any given new gTLD. Or the shelf space they can control inherently. If you have certain registrars that control significance amounts of shelf space, they would have greater power -- >>STEVE SALOP: But Go Daddy has less than 30% of the shelf space. NSI, from reading your report, has got 7, 8% of the shelf space. Are you suggesting that NSI with 7 or 8% of the shelf space has market power? >>ADAM PALMER: I would still say that there needs to be some transparent analysis of this or examination of this outside of this theoretical assumption that maybe they don't have the market power. >>JOE SIMS: But let me just make one point here. Everybody who makes a product that is sold to a retail customer would love to have that product on Wal-Mart's shelf because Wal-Mart is the biggest retailer in the country. Not everybody has their products on Wal- Mart's shelf. Wal-Mart picks and chooses the products that it puts. There's a lot of people out there still living, still viable, still making money who aren't on Wal-Mart's shelves. Now, it's possible that Go Daddy or ENOM or NSI or somebody has more power vis-a-vis registries than Wal-Mart has vis-a-vis retail products. But on the surface, on the face of things, it doesn't look like it. >>ADAM PALMER: Can I ask you as an affiliated registrar, wouldn't it be natural that they would give priority shelf space to those new tTLD applicants for which they're providing back-end services? And, therefore, you would be foolish not to -- >>STEVE SALOP: Yes, they would have an incentive to do so because they get a share of the profits. And they would actually sell the domain names for the affiliated registry at a lower price. And that's a good thing for registrants. Okay. [Applause] Second thing every retailer has private label products and they sell them for a lower price because they get a bigger piece of the action. And one that leads to lower prices for consumers, which is a good thing. I want applause again. And -- but more importantly, it does not tend to drive the branded products out of business. And what you're concerned with is that it's going to drive the branded products, the unaffiliated products out of business. And I think that requires some more analysis. You're saying that I'm being theoretical. But I think you're being theoretical. You're, one, making an assumption that the registrars have market power. You have no evidence of that. And, secondly, you say you're so worried that the registrars are going to coerce the registries into providing back office services. But, as I talked about before, if the registrars have market power, they can exercise that market power over the registries without the vertical integration. So I'm not sure what the rule, the anti-vertical integration rule does under your analysis other than protect the existing registry information, registry infrastructure providers from competition. And that's the question you need to answer. That's an on- the-ground practical question. >>JOSH WRIGHT: If I can add one point on empirics, I do a lot of work on vertical contracts with shelf space in the retail space. Right? So I understand there are important factual differences. Most of the research in that area, empirical research looking at effects on price and variety and quality and things that those end users really care about, suggests that -- and where you've got all the incentives in play with vertical integration into the retail shelf space from the manufacturer, vertical integration for the retailer with the private label -- you have all of those same incentives at play, obviously, a very different institutional framework is that, where you've got competition at the downstream level, and you've got folks like Wal-Mart who are large in the retail space or supermarket chains that are relatively large and probably have shares that approximate about the numbers I've seen in terms of the registrar space, that research, essentially, shows -- and, you know, caveat for full disclosure that most of it's by me -- that those competition for the shelf space ends up helping the consumers. It lowers prices. It increases product variety. Stores get bigger and nicer. And those sorts of things. And, I mean, that's sort of one of the natural predictions of the economics of competition for distribution. I think that, obviously, the shelf space competition here, there's similarities and differences, but I think not different enough that it's an irrelevant principle. >>KEN STUBBS: Yeah, my name is Ken Stubbs. I have a couple of comments to make. First of all, with reference to a comment you just made, I think there needs to be a significant amount of additional review of some of the data. Because there is a term that we use in the industry called "net creates." And these are the additions to the existing database. If you take a look at the net creates over the last two to three years, you find a significant difference in the patterns of who really has market power. So I think that needs to be looked at much more closely. And it was not reviewed as well in the initial report. Also, basically, eliminating a cross ownership restriction will give registrars access to very sensitive information about their affiliated registries and the registries for which they provide back end services. The CRA report says that if the registry star separation requirements are relaxed, ICANN would need to vigorously enforce the equivalent access and antidiscrimination requirements, provisions of the registry agreements, on one hand, including access to competitive sensitive data. But the report also acknowledged that elimination of cross-ownership limits make it very difficult to enforce these safeguards. We know that in the past this has been -- enforcement has been a very serious problem with ICANN. We also know that when the contracts were originally written, especially some of the accreditation agreements that were originally written, the enforcement clauses in this agreement made it very, very difficult for ICANN to deal with certain issues because all you had was a death penalty. Now, I am very concerned about just taking somebody's word on it that don't worry, we'll be nice. You know, we'll act properly. And, when you have the potential in this new round of abuse on a significantly larger scale than it has been in the past because you've had primarily one segment to deal with, now you're dealing with, as Jeff pointed out different segments and different definitions. You have an enforcement problem. You could use the approach that you said, Joe, and that is we've got to get something out there -- and I'm paraphrasing. And please correct me. We've got to do something. The point I'm saying is do we create a paper thin dike and we spend the next 5 to 10 years running around plugging holes in these dikes because we didn't realize here there was a loophole here and a window of opportunity there? I don't necessarily think that's in the best interests of the consumer, in the best interests of stability and security. I do not think it's in the best interests of business and intellectual property constituency people as well. So that's just the way I feel about it. Thank you. [Applause] >>JOE SIMS: Those are all relevant issues for ICANN to consider, I'm sure. The one that also needs to be put in that equation -- and again, thank God I'm not the policy guy here. I don't have to make this decision. But whoever is making this decision has to take into account all those things. But they also have to take into account the very likely consumer benefits that are going to arise from having a lot more choice in the TLD space. Now, if you don't believe there are any consumer benefits, then the balancing test there is pretty easy. If you do believe there are likely very significant consumer benefits -- and I think, certainly, that's my perception. I think that would be the perception from the sort of general economic perspective -- then you've got to balance those things and make some choices here. >>KEN STUBBS: I understand what you're saying. And you know full well I was not advocating for expansion of the space. I'm advocating for responsible expansion. And my concerns are strictly related to vertical integration and the problems that can result from that. >>JOE SIMS: No, I understand. And I wasn't as precise as I should have been. When I was talking about benefits, I'm not talking about benefits of new gTLDs. I'm talking about benefits that could potentially be derived from vertical integration by new gTLDs, which, I think, are at least theoretically, very real. >>BRUCE TONKIN: Hi. My name is Bruce Tonkin. I work for Melbourne IT, and I've been involved in the policy development process at ICANN for a while. ICANN is an industry self-regulatory body, essentially. Rules get set including registries, registrars, ISPs, business users, noncommercial users, et cetera. So there's two aspects of it. There's a role for setting the rules, and there's a role for enforcing the rules. I've chaired two different policy development processes that have come up with a different answer. One process was transfers to make competition easier. In that case, we set some very detailed rules about how to do transfers. And, if you think about it, we had the expertise to do that. And ICANN also enforces those transfer rules. And, you know, we have, you know, a compliance department that focuses on that. Another area that we set some roles was in approving new registry services. And at the time there's a lot of litigation going on. And the question is did we have rules about assessing whether new registry services would cause a competition problem or not, or do we leave those rules to someone else? The decision in this case was to say that if -- we'd post for public comment a new registry service. The industry can respond to that and say whether they think there's an issue or not. But ICANN itself doesn't rule on whether there's a competition issue. What ICANN has decided to do there is refer it to an external competition body. Now, in that case we're not setting the rules. But the only thing we've done -- we're not enforcing the rules really, either. Because all what we've been doing is referrals. So we're doing neither of those two things. We're not setting the rules, and we're not enforcing them. A position that might be somewhere in between there -- and I see this happening in Australia a bit. And I know in Australia we have a competition body called the Australian Competition and Consumer Commission. They don't directly, I guess, have enforcement powers on their own. What they do is they actually take a private company to court, and then the court resolves or makes a determination based on the law. So I'm just wondering a possible situation for ICANN here is to say well, do we really have the skills to write the rules? I'm not sure that we do other than perhaps being overly simplistic and making them real simple saying there is no vertical integration. And then we're trying to find all the loopholes around that. Or we say actually, we don't really know how to write these rules. Or we use another rule body. Let's say U.S. law, Australian law, and Europe law. But is it enough for us just to refer it to one of those bodies when we think there's a problem? That's a bit I have concerns with. Because in many cases those bodies prioritize their cases. And this might be a big issue for me as a registrar or a big issue for someone who is a registry. But a government body set up to regulate things might say, yeah, that might be an issue, but I've got this huge issue with telecommunications or I've got this huge issue with the airline industry. So maybe ICANN's role there is actually be the body that takes them to court. So it's not just a referral just like hey, come have a look at this. It's ICANN actually would take that complaint before a court and then get the court to resolve. So those are options. But my question for you is are there other examples of industry bodies that you know that are industry self-regulatory that do some degree of rule setting but where they're actually taking an approach of literally taking people to court using the laws of the country versus trying to set their own laws or setting their own rules? Sort of get your perspective on that balance? >>JOE SIMS: I'm not aware of an example of that. Bruce, are you guys? >>STEVE SALOP: I don't know whether under U.S. antitrust law -- I don't know whether ICANN would have standing. >>JOE SIMS: The issue would be in different jurisdictions there would be standing issues. And then the rules might be -- might vary from jurisdiction to jurisdiction. And so -- I'm not sure there's a workable -- I mean, I see what the rationale for it is. But I'm just not sure that's a workable solution. It's an interesting thought. But I think it -- it's complicated. >>STEVE SALOP: You know, you could, if you wanted to, you could set up your own little antitrust jurisprudence based on European jurisprudence or U.S. or Australia or anywhere else. You could set up a Bureau of Competition with lawyers that would prosecute cases. >>JOE SIMS: Just think, more lawyers at ICANN, wouldn't that be a wonderful thing? >>BRUCE TONKIN: I guess I'm focusing on the enforcement piece rather than the rule setting piece. A lot of people are concerned. In the past ICANN has seemed to be fairly weak in enforcement in many areas. >>STEVE SALOP: Certainly one possibility is you would just, if you will, delegate this whole issue to national competition authorities. >>JOE SIMS: Which is what they've done with registry services. So -- >>STEVE SALOP: You're certainly right. You would be a prisoner of their authorities if you did that. >>BRUCE TONKIN: That's the issue and then the speed. So what happens is because we're a prisoner of those, once someone starts breaking the rules, someone says that's good. I'll start breaking the rules, too, and you lose control of it. >>JOE SIMS: Unfortunately, I know there's other people that want to talk. And perhaps Steve and Josh can stay here and talk. But we're going to have to end the meeting now because I have to go someplace else. Jon? >>JON NEVETT: The suggestion is, if we can take written questions, we can provide them to the panel and ask them to reply formally. Would that work for everyone that's still in the queue? >>In a working group or something like that. >>JON NEVETT: It would result in making a public record of the questions and responses to them. And then we can also make sure that the rest of this is appropriately placed so we can -- >>JOE SIMS: Any of the rest of you who want to try to take these guys on the way out, feel free. >>I know you have to go. But couldn't we let someone else moderate and stay for another half hour. There's five or six people here, and these gentlemen are here. We've got people sitting in the room. I think it's only fair to give these guys who have been standing a chance. >>JOE SIMS: I defer to the ICANN staff on that. >>JON NEVETT: We're being encouraged to end this because we've gone significantly over. And there are many other programs going on. Not that we want to cut off this dialogue. This is fantastic. But I think we can move this dialogue to a written form. Nancy, can we keep the room for a little while? (Speaker off microphone). We just got another half an hour. But it's also up to the two guys on the stage if they'll stay. >>JOE SIMS: They'll do it. And they can either moderate themselves or, Jon, you can step up and do it. >>JON NEVETT: We're going to go for 1/2 hour. But, if anybody else joins the queue or we can't get through it, we're going to have to take written questions. >>PAUL FOODY: Paul Foody. Sold. I understand from Kurt Pritz today that, as it currently stands, ICANN does not know how many new gTLDs they're going to be permitting. They don't know how many the system can support. Given that, you talk about ease of entry into a market. I was watching the cricket today. The 20/20 India versus Pakistan and Pakistan won. All around the adverts were single word, Yahoo, or -- it was all one company. There is no dot com any more. And, as soon as organizations, cooperations are given the right to buy their own gTLD, that's going to be it. The top thousand corporations are going to buy up the Internet. That's going to close off the Internet to every other competitor, although they're only resource is to -- is going to be to get a secondary domain. You know, a dot com, a dot net, a dot Google, a dot whatever it might be. How is that good for competition? >>STEVE SALOP: Why if the thousand corporations would each have their own -- their own TLD, why would that kill the other ones? >>PAUL FOODY: That's exactly the comment that people used 20 years ago when I was working in the commercial property business. And they were resisting applications for supermarkets. Today the British food industry is dominated by these -- by the supermarkets because they have -- >>STEVE SALOP: That's because the mom and pop stores couldn't compete. They had high prices -- they had very high prices. Their food tended to spoil. Are you really saying that we shouldn't have permitted supermarket chains? >>PAUL FOODY: What I'm saying is that competition is what you are basing the Dennis Carlton report supported the new gTLDs on the grounds that this encourages competition. If only the top thousand corporations are able to get effective domains that they can advertise effectively around sporting venues, everyone else is just fighting for the crumbs. >>STEVE SALOP: And you don't think the fact that -- would these thousand TLDs also have other people having domain names in them? Or would they all be corporate proprietary? >>PAUL FOODY: If I was the like of, say, Yahoo or Microsoft, I wouldn't let anyone else on my side. Would you? >>STEVE SALOP: That's what Yahoo does now. They have a search engine. They let everybody on. >>PAUL FOODY: Same with Apple. All these organizations that are single product. Coca Cola, Pepsi. >>STEVE SALOP: Apple just opened up. Apple opened up. They used to be closed, but now they're open. In general, when there's network effects, there's a benefit from opening up. That's why VHS beat Betamax. That's why Microsoft beat Apple. And I don't think you'd get a thousand closed TLDs. That's not what I would predict. >>PAUL FOODY: I'm predicting 10,000, if you're allowed 10,000. And they will all be closed. >>STEVE SALOP: And these 160 million other people who want to have domain names, they would just not be able to get domain names? >>PAUL FOODY: Well, they'd be able to get domain names. But they'd be buying them in the secondary locations in exactly the same way that only the big retailers get the prime spots on any High Street in the U.K. Everyone else fights for space on the secondary locations. Thank you. >>DAN WARNER: I might speak briefly before Tim because I might say something he might want to respond to. I'm Dan Warner. I previously was with fabulous.com, which is around the 12th or 13th largest registrar, until about two weeks ago. I'm with interweaving.com now. I would say that regulators and regulation is actually much more closely aligned with communism than it is with capitalism. And the reason for that -- not that I don't think the regulation has a reason to exist. I just think it tends to overcomplicate things. One of the problems is that for free markets to really work, you need flexibility, commerciality, and agility. And regulation tends to do exactly the opposite of that. The more you regulate something, the more you increase the size of a tax system until you have four books that you can't even lift in order to create taxes that people don't understand any more just like regulation in the domain space. You actually gain a lot of inefficiency, and you spend more time administrating and actually competing. Competition is hindered by a lack of simplicity and a lack of flexibility. And I can even complicate that a bit. A lot of what you're looking at is in the sense of is there competition in the registration space. I'm a strategist, not an economist. And I certainly am not a regulator. But, as a strategist and having worked for fabulous where I look at a lot of different things, you have three major groups, of course, which is the registries, registrars, and registrants. Now, in registrants there's professional domain owners. And some of them own hundreds of thousands of domains. And they can be quite powerful themselves. You have the registries. And you have, like, Go Daddy. And Go Daddy actually does have a very dominant share of the market. And you say 35% isn't a lot. But, if you look at the retail market and the consumer mindshare and if you start to bring in, in particular, all the things that are involved in this space, you take it and say this isn't about registration. Actually, most of the money isn't in registration. And that's largely because, through distortion of overregulation, in some ways, it's actually forced the businesses to say all right, well, I'm going to make my money in Web hosting, which tends to be the most profitable sector. So you end up with doing packaging deals and things. You get this domain. Like in Germany, they give you 10 domains, I think for free. But you're paying for the Web hosting. I may be wrong on that. Any Germans in the room, I apologize. >>It actually costs money. >>DAN WARNER: Okay, that's fine. It's actually a complex market and transfer pricing between different products you have. If ICANN thinks it's regulating the price of domains, it's kidding itself. It's just not the way the market works. And what I -- the point that I really want to make and what I want to really ask your opinion on is that, if you simplified all the rules and you took out all of this, how are we going to regulate every single point in the process and every single way that anyone sells everything when you have a hybrid combination of completely unregulated markets, like, hosting and registrations, why can't you actually -- what's wrong with competition? What's wrong -- if someone comes up with dot shop and they only want to sell -- they wouldn't to be the registrar and they want to be the Web hosting company and they want to be the registry, they want to do all that and they want to sell at a thousand dollars a piece or $2,000, $5,000. In an open market when they don't have any dominance of competition, what's wrong with that? Dot com is the 10,000-pound gorilla. Everything else is superfluous. Right? If you look at the retail market, Go Daddy, which I really like as a company, by the way, is a hugely dominant player in the retail market. And, in particular, when you look at where do they stand in the world when it comes to hosting plus registration plus these other services and mindshare with the consumers, they have a hugely dominant share. If you look at ENOM, ENOM is a reseller provider. They are actually distributing -- they don't have that central brand. They're distributing all their mindshare throughout their distribution network. They're an enabler rather than a competitor in that sense. So comparing ENOM against Go Daddy in how they're competing is really not that helpful. So how -- how do we actually eliminate regulation instead of increasing regulation? How do we open the market for flexibility, agility and commercialization? >>STEVE SALOP: You're making the argument for regulation by saying that Go Daddy's got a lot of market power. >>DAN WARNER: Absolutely not, because that's simplistic. I'm saying at the top end of the market, if you take a collection of all these things, hosting, registration, you know, registry services, registries, if you take all those things and you get players that are too big and they have too much market power collectively they need to be regulated. If you leave the rest of the market lone and make that tax system a 10- page book you will end up with a deficient system where the market is the market and commerce actually rules. Do you have a differing opinion to that or it is it -- >>JOSHUA WRIGHT: I'm not sure I heard a question. >>DAN WARNER: Well, do you believe that we should simplify the system to be -- >>JON NEVETT: Sorry, can we wrap up the question? >>DAN WARNER: Okay, can we simplify the system to be nothing more than ensuring that dominance doesn't occur and over -- too much market power for only the top players. >>JOSHUA WRIGHT: Sure, there is -- I think some of that is consistent with the general economic analysis we -- we went through. I mean, you can think of some really simple rules, right? You can have a complete ban and ban the vertical contracts as well. You can have a per se legality rule and let whatever happens happen and maybe you want to carve out some exceptions for dot com and sort of treat that differently. You can have simple rules. The economic point is that simple rules are going to lead to either under-inclusive or overinclusive results. Now, the fact that there are these anticompetitive -- potential anticompetitive effects that would leave from one result -- or that you would throw out the pro-competitive virtues with the complete ban follows pretty straightforwardly from the economic analysis. Now, for the point about simplicity versus administrative costs, right, so on the spectrum you have these two simple rules and you can think of case- by-case analysis in the middle and you might think -- and this is a question for ICANN to determine, not economists -- you might think that the administrative costs of doing that case-by-case analysis are something that you don't want to bear. Or that are too high relative to the gains and instead want to adopt a simple rule or maybe delegate to antitrust authorities. In antitrust, for example, the challenge for antitrust authorities over the last 100 years has been how to struggle with the exact same question. How to embrace the case-by-case analysis because you need the facts to identify the bad instances, but still give businesses some clarity and certainty about what they're allowed to do and not allowed to do and so in antitrust you'll see some combination, usually, of some standards that are very fact-intensive coupled with some very bright-line rules, right, so if you don't have market power in antitrust you can't violate the U.S. antitrust law. >>DAN WARNER: But when you only really are regulating when people trigger that regulation, when you become too big and you're becoming so big and powerful you're anticompetitive. If you left the rest of the market to be the market, why can't you -- what -- what economic reasons are there that you cannot do that? I mean, what's the -- >>JOSHUA WRIGHT: None. >>DAN WARNER: Okay, thanks. Just wondered. >>JON NEVETT: Tim? >>TIM RUIZ: Tim Ruiz with Go Daddy. I'm not going to comment about our market power. I just smile big when I hear things like that. I started with Go Daddy when we were number 181 out of 185 registrars so it's kind of exciting but -- I guess I have more of a concern that I'd like to express than a question, and you can respond to it as you like but, you know, I think the issues that have been raised today, some of them are new, some of them are things that we really haven't spoke out about publicly since this whole issue came up and it's very clear to me, you know, that this is a very complex issue, and I really appreciate the advice that simplistic solutions are going to ensure mistakes. And so my concern is that as we move forward, if we think that somehow we can stick to the existing timeline and really fully address this issue and not make those serious mistakes, I think we're really fooling ourselves. It seems to me that there's really a choice. We either need to rethink this time line, push it out, maybe not even have a time line until we get some of these problems resolved or pull this back, let's go with the status quo for this first round, and then address this issue appropriately, give it the time it needs so that we can really discuss it and vet it and then think about it, rolling something like this out perhaps in the next round. But I really discourage us from, you know, just creating policy on the fly, doing it hurriedly because we want to get the new gTLD process out, we give this the time it really deserves. >>STEVE SALOP: So do you charge higher prices than your competitors. >>TIM RUIZ: Absolutely not. >>STEVE SALOP: Would you contribute data to be analyzed to calculate revenue-based market shares? >>TIM RUIZ: (Speaker off microphone). >>JON NEVETT: Brian? >>BRIAN CUTE: Brian Cute with Afilias. Earlier in the conversation there was a suggestion that there might be different types of treatment for registrars that had market power versus registrars that don't have market power. Both with respect to registrars but also registries, have there been any systematic analyses of market power for registrars or registries that you're aware of or any that ICANN has conducted since the CRA report did not go into that in great detail and that's question one. Question two, and then I'll step away, I assume you've heard some new inputs today that you didn't know before in terms of back-end registry services being one element of this of ccTLDs and reseller channels. What are the indicia that you guys would look to to discern market power given some of the inputs you've gotten today and one last statement: I'm not sure it's been made clear by some that in terms of the status quo, there is a status quo structure with registries and registrars and there's some vertical constraints. It's not clear that if the status quo were to remain that some of the potential abuses wouldn't flow from that anyway because the status quo actually has an absence of certain safeguards, can you comment on all three of those, please? >>JOSHUA WRIGHT: With respect to the first question about -- I mean, I've seen the studies that are publicly available so I've seen, I guess, what you've seen. And out of those studies, with respect to any definitive studies of market power at either level, I mean, I've seen data on shares and the like and I think one of the points that's come out of both Steve's analysis and mine is that when you want to calculate -- do a market power analysis, when you want to engage in a market power analysis, whether you're talking about the registrar or registry level, you want to be focusing on the competitive dynamics. And I think some of -- your second question blends into this, right, so some of these new inputs might be really important to consider the reseller channel being one, for instance, or a -- Ken had the comment earlier about maybe there being a difference in registrar shares, whether we were thinking about current or historic. You might think that the more recent ones are a more important indicia for how you calculate shares which, let me say the following, I mean, shares themselves are an imperfect way to conduct this analysis. There are industries where shares are not a great proxy for market power and there are -- things like Steve has mentioned throughout with respect to pricing behavior and other indicia that would be important. I mean, neither Steve nor I has conducted that analysis. In part my understanding is that there are some real data limitations in doing that. >>BRIAN CUTE: And you've spoken a great deal about shelf space and the research that you've done. Are there industries we could use as an reference for identifying certain indicia that would be useful in this space? >>STEVE SALOP: Yeah, the indicia, the general indicia, you know, we could give you a list, we could refer you to the European Commission merger guidelines or the U.S. merger guidelines. The general indicia are things like -- it's all about -- market power's all about substitution and barriers to entry. So you want to know whether there's a -- whether registrants feel that it's easy to substitute among registrars or whether they feel locked in to particular registrars. You want to know whether some registrars charge higher prices for the same quality service. You'd like to know about price cost margins and profitability. You'd like to know about market shares, not market shares sort of in the whole universe of the sort that's been published in the various reports but you'd also like to get sort of a narrower possible markets like a corporate market versus a consumer market. So, you know, a host of, you know, a whole host of variables that you would look at. And I guess I would say from, you know, the work I've done, that no, there hasn't been a really good study done of market power at the registrar level which I think, you know, ICANN's been focused on market power at the registry level but what's come out in this whole discussion about vertical integration has been a concern about market power at the registrar level and, you know, at the infrastructure provider level. At the registry level, I'm inclined, personally inclined to think, and I'm stating those words carefully because it's no more than an inclination, but I'm inclined to think that dot com has market power but, you know, frankly no one -- no one's done that study in a rigorous way either. I mean, I tend to think that, you know, dot com, as fast as they're growing, if -- probably if they'd raise the price up to $500 per domain registration that would probably be profitable, that people would feel locked in, at least for a while. But jeez at 500 a pop for three or four years, that would -- even if they died after that, it would probably make enough money to make it profitable but that's at the level of the study. That would be the sort of thing you'd do with dot com and there's a price (inaudible) for dot com. But do I think that dot biz could do that, that it would be worth it for dot biz to raise their price to $100 and sacrifice the potential growth? Probably not. >>BRIAN CUTE: There's creative people in the room. Careful what you suggest. >> Jon, we need to close this meeting. It's very important if you don't mind. >>JON NEVETT: You told me I had 30 minutes and six minutes left and two questions. >>NANCY LUPIANO: If you can do it in six minutes, otherwise the doors will open and I'll have to start moving chairs. Thank you. Quickly, please. >>JON NEVETT: So let's ask quick questions and see if we can get them in. >>JORDYN BUCHANAN: One quick question and one quick comment, and I'll do them both together and then you can answer. That's me, yes, not Jon Nevett, oh, yes, other Jon up there. The quick question is: So it's been mentioned that slotting fees often have a good effect for consumers. Are there situations in which slotting fees are harmful to consumers? Just looking around briefly, it looked like there were at least some cases in the U.S. with McCormick that the FTC at least concluded that slotting fees had been used in an uncompetitive way, and I was wondering if you could provide more light as to how that might happen so we might consider that in our discussion and I just did want to briefly mention, though, despite that question, that, you know, there's been a lot of discussion today about, you know, whether or not forcing access by registries or limiting access between registries and registrars is helpful. I worked for a small registry for a while. And certainly in our view was it would have been really helpful to be able to create our own market in order to innovate and create new products that registrars weren't used to sell today, certainly didn't want to force it on anyone else. And so I think in both cases, having regulations that would force registrars to carry our product, it probably wouldn't have been fair, but at the same time being able to just offer up a product on our own as opposed to being forced to go through a channel that wasn't particularly interested in selling our product would have been really helpful too. So I'm going to agree with previous commentators who say that low sets of regulations for new entrants would be really helpful, I think. >>JOSHUA WRIGHT: Let me answer your question very quickly. With respect to the slotting contracts, the McCormick case you cite was about slotting contracts in the spice industry in the U.S. and was a consent decree between McCormick and the FTC rather than a fully litigated decision. There's been no real analysis, there was no real analysis in that decision about the ultimate competitive effects. I don't mean that to suggest that they were not anticompetitive, they may have been, but I would not say that that sort of substitutes for the type of rigorous analysis you would want to inform your policy-making. That said, there are a number of cases in the U.S. involving competition for shelf space where real competitive concerns are raised. I don't think McCormick's a particularly great example, but things you worry about -- and most of those cases in the U.S. involve all things equal, longer-term contracts, you know, years, longer-term contracts although that's a -- not a necessary condition for competitive harm. But in terms of sort of dispositive studies the FTC took out a several year study on slotting contracts and essentially reached no conclusions. That was largely because of data limitations, they asked nicely if companies would like to give them their slotting data and the company said no. I would refer you on the empirical evidence, one of the articles cited up on those bullets on the slide, we'll point you to essentially all the evidence -- will point you on essentially all the evidence there is on slotting if you're interested. >>STEVE SALOP: I'm a little more hawkish than Josh on slotting fees but if you want to have a good example of exclusion from retail distribution I think you can look at Windows tying Internet Explorer, that would be a good place to start. >>JON NEVETT: Last question. >>MARCUS FAURE: Marcus Faure, I'm with CORE Council of Registrars. One quick information on Germany's dot de domain names, they're much cheaper than dot com and there's no 7% price increase. But let me start by saying that CORE is, in the absence of a better model, all for registry-registrar separation, but I concur with Michael Palage that the solutions that we're looking at are many times done with U.S.-centric thinking. Take, for example, as I'm wearing this nice tee shirt of a 2002 Ghana meeting, take, for example, a TLD that targets the market in Ghana, take a dot Accra for the capital of Ghana. There's no ICANN-accredited registrar in Ghana, so there's only two in the whole of Africa. So what you have to do, as the registry, in absence of the possibility to have your own registrar, you have to go to a U.S. registrar on bended knees and ask them to offer the possibility to register your dot Accra, and you have to tell the people in your country to go to the U.S. and register a domain name. Probably they have to pay by credit card, which might not be a favorable payment system in Ghana. So there's no simple solutions here. We at CORE have some solutions, and we're happy to work with staff and to offer those solutions. So my question here is, how do we get in contact, or the offer is for anyone of ICANN staff present contact me or one of my colleagues to discuss this topic. >>JON NEVETT: We'll certainly do that, thank you. So did either one of you want to respond to at all? If not, thank you so much for taking your time, your extra time to come in and answer these questions and thank you for everyone who stayed for the whole program. We'll make sure there's a great transcript of this, we'll help make sure it's all ironed out and any additional comments that anyone has that they'd like to add to the record, please let us know and we'll do that.