SEVENTH ANNUAL INTERNET GOVERNANCE FORUM BAKU, AZERBAIJAN SUSTAINABLE HUMAN, ECONOMIC AND SOCIAL DEVELOPMENT 7 NOVEMBER 2012 14:30 CET WORKSHOP NO. 105 INTERNET REGULATION FOR IMPROVED ACCESS IN EMERGING MARKETS ***** This is being provided in a rough-draft format. Communication Access Realtime Translation (CART) is provided in order to facilitate communication accessibility and may not be a totally verbatim record of the proceedings. **** >> BILL WOODCOCK: Okay. If people want to put on their headsets, I think we will start. So this is going to be a relatively informal session. We have here a number of experts on Internet regulation and the ways in which it can go right and go wrong. Bernadette Lewis it looks like is not going to be with us on the panel today. My name is Bill Woodcock. I will be moderating. I am the executive director of Packet Clearing House. We are an international Internet operations organisation. We support the operation of Internet critical infrastructure, like Exchange Points and the core of Domain Name System in more than 150 countries. We have policy and Governmental work that we do in nearly all of these countries. So I have worked with a lot of regulators on Internet issues over the last 20 years. Bevil Wooding is my colleague also at Packet Clearing House. He is our Caribbean European specialist and he works with Caribbean countries. Sam Paltridge is with the OECD, the Organization for Economic Cooperation and Development and leads their telecom unit. So he does a lot of economic and quantitative analysis of the effects of different regulatory policies. And Jane Coffin is now at the Internet Society doing development strategy but until quite recently was at the NTIA in the United States and before that with U.S. aid. She has also dealt with Internet regulatory issues both within the U.S. Government and in several countries in Eastern Europe. So I think what we would like to do is -- so we are talking today about whether regulation can improve access. And so first let's define access so that we are sort of on the same page. What we mean by access is the combination of affordable price, availability; that is, it is actually something that people can get in their location wherever their location may be and functionality. If they get it, it will have some basic level of working. It is not going to be completely useless. So then the question is whether regulation can help achieve these goals. And I am going to sit down and join the panel and we are just going to talk about it based on our experiences, but we would very much like your questions and involvement. As I said this is going to be a very informal hour and a half. And we really look forward to an interactive session in which we try and address questions and points that you guys may bring up and hear from each of you and your -- maybe not each of you but each of you who are interested and what your experiences have been as well. So I am going to sit down and put on the headset and -so perhaps just start with Jane maybe and maybe you can just chat for a couple of minutes about sort of your experiences from the different points of view that you have had and what you think can make things better and, you know, maybe as I said entertain questions fairly interactively. >> JANE COFFIN: Good afternoon. Can you hear me? Okay. Great. Thank you very much, Bill. I have spent some time as Bill had indicated in two different countries, in Eastern Europe and partially in this region. Several of the factors that I found when working with regulators that was so critical and this is something that I found interesting from a perspective of coming from a very consultative environment back in the United States you need to assess your environment and a lot of this is practical and simple, but it is important to remind ourselves that we are trying to help improve the infrastructure for the country and for the people in the country and as Bill was saying bring down prices and make things more affordable, assess your environment and consult. This helps facilitate stakeholder participation. And I know when we say stakeholders what do we mean. We mean the people who are invested in using the infrastructure or building the infrastructure, regulating that infrastructure or providing policy guidance. I think for a lot of regulators they feel that speaking to companies or constituents in their environment might be a sign of weakness and actually it is a sign of strength if you consult. If you continuously understand your environment, and I found that as a person in policy myself I found myself consulting with Bill Woodcock. It was a very difficult issue that was put in front of me when I was working with Internet protocol when I was working for the U.S. Government and it was important for me to find out more about the Internet infrastructure. I turned to Bill and others to find out what that infrastructure was that -- what I didn't know and what I was missing. So it is very important to assess that environment. Recently the Internet Society has been engaged with many different countries around the world and with companies like Bill's to take a look at providing more Internet infrastructure in the form of Internet Exchange Points. Again you assess your environment and figure out how to bring stakeholders together, understand the technical infrastructure and then from there continue to consult so that you are not making mistakes that harm that environment. You can disadvantage small companies and disadvantage the companies that are already in the market. You also need to talk to your incumbent, of course, to figure out what's going on if you have an incumbent in that market in order to help facilitate that environment. So I will just throw that out. The importance of assessing environment from the physical and technical and human because you are always wanting to increase the technical infrastructure and the human capacity development in your country which is so important and you can do that as a regulator and a policymaker by consulting with technical experts and/or other experts around the world. It is important to just know what you are trying to regulate quite frankly. >> BILL WOODCOCK: Sam, do you want to just take a couple minutes and ->> SAM PALTRIDGE: Okay. Thanks, Bill. So Bill asked me to talk about a question of what regulators can do to help improve access. And what are some of the things maybe that they get wrong from time to time and the discourage or help to access to development. Since I have got three minutes I will use one of them to do a one minute history of telecoms. We started out with monopolies and basically the Government did everything. It was the regulator and operator and policymaker. We tried this for a hundred years and most countries have less than 1% telephone penetration. We are not doing this very well. And one of the problems, of course, was where you get investment from and having the Government as the owner and operator and policymaker meant that any time revenue came in from whatever source it would be really to go back in to telecoms. Quite often go to very worthy causes. Would go to hospitals, libraries, whatever. It often didn't go back in to developing the network. So we made some changes. We split off regulation from Governments. We split off operation from Governments. And we said to the private sector you can have a role here since you have got investment that we need. And part of that was having a regulator to intervene where there was insufficient competition and that quite often happens. We know that because we have limitations. We have limitations, for example, in a spectrum where you can't have free market access, free market exit even because policymakers don't like it when people lose their telephone numbers either. So often they keep inefficient operators in the market. So there is various reasons that you need to have regulators. I could talk forever on why you need to have regulators. But where we have got to in the last especially last 10, 15 years we have managed to transform access in telecoms to where we have over 100% penetration in many countries and if they are not there yet they will be soon. And, of course, the Internet came along and in many ways helped develop that access further because there was more demand for that access. But also it helped in many other ways. It helped drive demand but it also helped regulators. It helped the sharing of information about regulation. And so one of the things that we do at the OECD is share best practice and we talk about what the issues are and how we share best practice. And I was in a session yesterday with the former regulator from Sri Lanka and one of the problems and challenges he had going from United States back to Sri Lanka he can't access the information that he needed and the Internet has helped him to do that. On the other side there is many things that regulators can do to make that better. One of the primary ways is how to make an efficient market and how to make it work efficiently and how to increase competition in that market because that helps drive access. And on the downside we see many bad practices. And we see how that can deter growth in network access. We see in many places the market throwing out very surprising developments. Sometimes I am a little embarrassed to say that many of the leading developments are often not in the developed countries, not in the OECD countries. They are in the developing countries. And I cite international roaming as one example where often you can go across borders in Africa and not pay roaming charges the way we do in developed countries and that's a tremendous outcome of market developments. But regulators have to be careful that they don't intervene in a way that may stop such a development which has tremendous benefits for consumers. And there is some things that I can talk probably at length about what they do and give you many examples but I will come back to that. I hope there is lots of questions anyway. >> BILL WOODCOCK: Bevil, do you want to give us your thoughts for a few minutes? >> BEVIL WOODING: Yes, just a few thoughts. Looking at the question, can regulation help and I think the answer is yes, but it can also hurt access to the Internet and to critical services. So just to repeat something that Jane said, objectives must be clear. I think it is simple but an important reminder for regulators to always keep in mind what are we trying to accomplish as regulators. I am functioning largely in a region in the Caribbean that is coming out of a monopoly mind set and that severely impedes the decision making at the regulatory level. So a big part of my function is simply explaining how things connect. How decisions that might have been acceptable a decade or two decades ago can now inhibit market growth and development. And bringing that kind of information to regulators is an important part of the process of them making the correct decisions. There are a couple of important points that I wanted to just throw out there. One is commercial priorities versus the social development priorities. That continues I think to be a challenge for a number of regulators, particularly in some of the Emerging Markets where you have the income from telecoms in particular playing such a significant part in Government revenues. And that puts Government in a sometimes conflicting position where there is a priority to secure these predictable revenues and what they, providers, would do is to say well, if we were to make the markets more open or encourage more competition, then we can't guarantee you these revenues will always be there for you and that presents challenges relating to infrastructure development and opening up markets to different kinds of competition and, of course, rule out of infrastructure to enable the technology that we all have. So I think developing a holistic understanding of the implications of regulatory decisions is one of the key points that factor in to how decisions are made, that ultimately have a direct bearing on the kind of benefits that are derived from regulatory practice. The other thing is hard data or scientific data is very important to decision making. I have seen cases where decisions are made on the basis of what the providers say are not necessarily on the basis of what is factually true. And this is an important consideration, particularly when you have regulators who are dealing with severe human resource constraints. And that really severely comprises their ability to do the kind of research and due diligence that is necessary to make progressive regulatory decisions. So just to throw those things out there and see what kind of discussion we can generate. >> BILL WOODCOCK: So are there any points that any of you in the audience would like to bring up or any questions yet? Or shall I -- okay. So I think one of the things that I use as a test is whether a regulatory environment allows a teenager, a student in a high school say that can start a business that can compete on an equal footing with the incumbent. In a lot of places that seems like a ridiculous question or people think that shouldn't be a goal but then there are other places where there are very healthy competitive markets like, for instance, New Zealand where the same license is available to everyone at the same price and it is quite affordable and there is no real barrier to somebody starting a new business as a teenager that competes on an equal footing with the incumbent that is not subject to any structural disadvantage. And I think that that's a really important point. I think that regulators -- the main duty of a regulator in my opinion is to preserve opportunities for new market entrance, right? That you can't look at a market and say this market is already as good as it will ever need to be and all we need to do is preserve the status quo and you can't count on current participants in a market to provide all the innovation that will ever be needed for the future. The frustration that your current market participants occasion and their customers should be allowed to drive those customers towards innovation, towards providing competitive alternatives. Before I came to this I ran an Internet service provider. And that was what I wanted. So I had to cobble it together myself. You know, later I saw people who were competing against the incumbent because they didn't think the quality of the service was high enough or they didn't think the price was low enough or they didn't think that the services were innovative enough. And all of those things were true. And anything that prevents those people from turning that frustration in to entrepreneurial energy and making the market better by providing new services is an impediment to the growth of the market and an impediment to the raising of the quality of people's lives. And so honestly I think the main goal of a regulator should be to enable kids to go out and do something better. And there are far too many regulators that think that their job is to have meetings with the incumbent and talk about what the incumbent's goals are and that just doesn't seem very interesting to me. Most regulators I think would be far better to pretty much just ignore incumbents and focus on, you know, new market entrance and the little guys and the folks who are doing something new and interesting that hasn't been done before. Let's see, so -- sorry. >> JANE COFFIN: I wanted to throw out the idea for many of you and I think this is something that if you are a regulator you know inherently, but if you regulate to facilitate stakeholder participation you are only increasing your knowledge. The infrastructure development across your country and you are again training yourselves. It is the human capacity development angle. It is important that your own teams as Bevil said are trained and that's very difficult in small countries or in countries where you have resource issues. I wanted to give you a snapshot of a story I heard in Lesotho. I was in Johannesburg and the regulator from Lesotho was at the meeting. I had to regulate in order to facilitate infrastructure development and I first thought what did he do. In order for the Internet Exchange Point to exist in Lesotho the laws of his country mandated that the infrastructure surrounding the group of stakeholders that wanted to run the IXP, technical experts and other Internet community experts he needed to pass a regulation that allowed for the Internet Exchange Point to exist. So that's what I mean when I say regulate to facilitate Internet infrastructure. And some people think you are trying to regulate the Internet. What we were trying to do is promote positive infrastructure development through regulation where stakeholders were recognized in that country because you do as Bill is saying about a strong incumbent, if that incumbent is not happy with the idea of an Internet Exchange Point because they don't quite trust that model yet, you don't want any legal uncertainty with respect to the viability of that entity, the IXP. So in any event that was one example where it was a positive stakeholder collaboration but also a positive regulation to facilitate that infrastructure being developed. And I think as we all know that the more infrastructure you have the more economic development you are going to have. We have seen that absolute tie with infrastructure, Internet infrastructure development and economic development, and if you are a regulator, your objective should be to help facilitate economic development and not hinder that development and innovation. If a 15-year-old is allowed to go in to the market, are you helping to facilitate that child's innovation? And he could be the next Isha Ida. But we have seen lots of different examples and I know that many of you have different environments. You have different legal systems. You have different regulatory systems. We are not suggesting that all countries are the same. But we are just saying know your market and know your people and know the companies in your space. >> BILL WOODCOCK: I think it is really important as a regulator to use a very light touch. And so there are a lot of regulators who think well, we are going to liberalize. So that means that we are going to issue a second license. Right? And then suddenly the hunt is on for who can pay the most for the second license. And you know that's not going to be a 15-year-old and you know it is not going to be anybody who has any innovative ideas. It is going to be somebody outside the country who has a big pot of money. Alternatively regulators can do class licenses. A class license is one that applies to anyone who chooses to do something. So class license says not that there are going to be three licenses for ISPs issued and you can bid for them. If you want to be an ISP, you can be an ISP, but this set of regulations is going to apply to you and you need to conform to them. And that's a perfectly reasonable thing to do. Because it applies to everyone equally. And it doesn't disadvantage any one competitor relative to any other as long as there is reasonable requirements. You know, that can be done poorly also. So to give an example from the United States, there essentially is a class license on access to telephone poles to put new fiber on to poles that says if you are going to touch any pole you have to put up a bond and you have to rent the pole for a period of 30 years paid for upfront. And so that's all okay if you are a huge company. But if you are just some kid who wanted to rent three poles in order to run some fiber down the block, it is not helpful to have to pay for 30 years upfront and not helpful to put up a multi-million dollar bond. There is other ways to achieving those goals. Yeah, I think the -- the really classic example of a useful class license requirement is instead of saying well, we need an Exchange Point. Therefore we are going to rule in law that one must exist or we need an Exchange Point; therefore we are going to require that everyone peer. These can be really problematic ways of approaching it. Instead if you have an ISP class license that says domestic traffic should not be passed across the national border, a nice simple way of getting to where you want to be, it means that all the ISPs need a way of interconnecting with each other domestically. >> SAM PALTRIDGE: On that I have a question for you. But first a little history on Internet Exchange Points. We realised in the OECD in the middle of the '90s that these are incredibly important and we were learning of something new because we came from the telecom world. Packet Clearing House has been great in helping us with that. We did a report, one of the first reports we did on and we said to all OECD you should encourage the development of Internet Exchange Points. And the first use of that report was actually Kenya because the Kenyan Internet Exchange Point that today everyone says what a tremendous success it had a really rocky first year where the incumbent went to the regulator and complained about and the regulator shot down the Internet Exchange Point and that's because people didn't have the right information that a role of an Exchange Point plays and the benefits that it can bring to developing the Internet in that economy. Now come back to my question, we still have one OECD country that doesn't have an Internet Exchange Point. And I don't know if there is anybody here from Mexico but basically my best understanding of the Mexican Internet market is that there is no Internet Exchange Point and I figure that reflects a lack of sufficient competition in that market. And I wonder how Bill's proposal would either help that market develop or hinder that market because, you know, are our Mexican ISPs benefitting from taking the traffic across the border and getting a better price or is it because they have got no choice in exchanging that traffic domestically? >> BILL WOODCOCK: Yeah, Mexico is kind of a unique case. It is by far the most largest and technologically developed country in the world that does not have a Internet Exchange Point. It has a company that is incredibly strong. That incredibly strong protectionist regime wrapped around it that it perpetuates. That prevents competitors from gaining equal access to the market. So there hasn't been an Exchange Point started in Mexico because they have essentially successfully argued that there would be no point because they would refuse to participate and if they refuse to participate, then everyone still has to get traffic to them in Texas. So -- or New York or Washington D.C. or Los Angeles or Chicago. So the mechanism -- let's just say hypothetically that the Mexican Government issued as a condition of an ISP class license that domestic traffic could not be passed across the national border. In practice what a small competitive ISP would do is they would offer peering to Telmex and Telmex would laugh at them and they would go to the regulator and say we are attempting to comply with this regulatory requirement and we have offered an interconnect to Telmex and Telmex has refused. And we know we are out of compliance with the regulation but we are doing our best to comply. And we also know that Telmex is out of compliance with the regulation and they are doing so willfully with an aim to quash competition. So that is essentially the mechanism by which you use that phrase in a regulation to encourage domestic peering. Yeah. >> SAM PALTRIDGE: So my follow-up question -- for those of you that don't know me I am Australian. The Australian regulator about 12 years ago decided because the strongest incumbent would not peer with the next three largest ISPs decided to mandate interconnection between them and force them to peer basically as best I can tell from a telecoms -traditional telecoms perspective. That's the way telecoms regulators think. But what happened in the next -- at least the next five or six years was because you took the next largest three players out of the market who would have eventually figured out that if they peered in an exchange traffic with smaller ISPs they would have developed more alternative infrastructure. But because the regulator intervened and forced the three largest players to exchange traffic I think Australia delayed the competition that it otherwise would have had which came later. So ->> BILL WOODCOCK: Yeah. Exactly. So I think this is another instance of the sort of simpler case that I showed of a regulator going from a single monopoly license to two licenses. In Australia they went from having one market dominate player and a whole bunch of medium size and small guys to wow, now we are going to have an ogolopoly of four. They created a static market of four participants who were then able to de facto price fix. That's not a useful accusation to make in the absence of specific information. The net effect is that the market stagnated. All of whom were protected against competition from smaller guys and, you know, they were able to find a what the market will bear sort of price that they all charged. And yeah, as Sam said it put quite a crimp in the Australian Internet development. At the same time, you know, the little guys in Australia were not very strategic about what they did and they did not work with the regulator in a sort of open give and take kind of way. It was a very combative thing it seemed. They did a multilateral peering agreement which essentially meant that the big guys were precluded from participating unless they were willing to peer equally with everyone there which is not a requirement that's reasonable and not a requirement that obtains in very many other places. Yeah. So let's go back to the audience. We have got plenty of people here. Marco, can we get a microphone? Can we get a microphone for Marco there? The gentleman with his hand up. Thank you. Please introduce yourself. >> AUDIENCE: I am Marco. I make this comment in my role as a concerned citizen. On your suggestion of having a law that states domestic traffic must not pass the border, a couple of observations. We already noted that at some point you suggested like okay, so in Mexico they asked for peering and then came back with oh, yeah, we tried to comply with regulation. I am not sure whether it is smart to stare at regulation that you already know that people are going to drag. They are not going to be compliant. I mean if you introduce a law and essentially part of regulation is that sort of -- is sort of law, you might want to make sure that it is enforceable. Now with that you also have the option of people saying well, if they deny peering, you must buy their transit from them. That's the difficult risk if you phrase it like that. Maybe as a startup business I might be better off. It might be cheaper for me to touch my package over the border and give it to somebody else. So as far as encouraging small startups, it might even be better to say no. Okay. If taking the detour of 2 milliseconds makes it cheaper maybe it is. Maybe we should encourage people to look around. >> BILL WOODCOCK: So you have made about 15 different points there. No. I will try to address a couple of them and try and get my fellow panelists to address a couple of them. First of all, I would say there is a national interest in not having domestic traffic subjected to other countries' packet inspection regimes, for instance. That was one of the things that drove Sweden in the, you know, 2000 era to overhaul their Internet Exchange Point system to keep Swedish domestic traffic out of the London Internet Exchange. And that's quite a reasonable domestic policy, and it is reasonable for the regulator to regulate in the public interest for purposes like that. Purposes that involve privacy, for instance, of domestic traffic. I think price is not the only consideration here, right? Routing traffic across a national border in order to get price down indicates that there is some really significant domestic problem that should probably be fixed. And saying that you shouldn't pass a regulation that you know that there will be people who will choose not to comply with, I mean what's the purpose of regulation then if you only regulate the things that the regulated entities want to do? You are not again working in the public interest. You are working in the interest of the regulated entity. So I would say that having such a regulation, I mean just not to focus over much on this particular possibility but having that regulation too small ISPs would happily peer with each other and happily demonstrate compliance. You would quickly see which entity was noncompliant. It wouldn't be that you would be passing a law that no one could comply with. You would be passing a law that one entity that didn't want to comply with. >> JANE COFFIN: I am going to take it back to a less developed market because if I am paying long haul traffic costs to another country to exchange my traffic which is what happens if you don't have an IXP in a small country, again Kenya is a great example of where they came up with a better submarine cable model which was competitive because there were several cables coming in. And until they came up with a better model for competition, competition from the backhaul from the cable landing station as well with that IXP being put in to place, you cut down that long haul traffic cost for locally exchanged traffic. Some people just have a hard time believing that if Bill and I are sitting in -- he is in an Internet cafe in one part of Nairobi and I am in another part of -- it is going up to London and up to Berlin and back to me. It is not going to encourage me as a local participant in that market to use the infrastructure which is exactly what you want if you are the companies in that market. And if you are a regulator that may be where we are saying don't let it go over the border, but it is a different type of border. So we have to specify what we mean by border. >> AUDIENCE: I totally agree with you in terms of you must have an Internet exchange before you can regulate peering. And we partly touched on that yesterday as well, like you said you need enough diversity in the infrastructure to actually reach the Exchange Point in order to have an Exchange Point. That becomes a success. Once you have the Exchange Point there you can look in to basically what will always be the incumbent. That has become a law of nature. >> BILL WOODCOCK: I think you are looking at it as sort of a chicken and an egg thing. And a regulator is looking at it I can pass a law and make this happen. You don't have to have an Exchange Point in advance of regulating a requirement that people interconnect, right? If you regulate a requirement that people interconnect, then you let them find a way and maybe that's not an Exchange Point. Maybe it is a bunch of bilateral circuits. The regulator's interest should not be in advancing any one technology. It may be that there are places where an Exchange Point isn't the best solution. We don't know what they are. But ->> AUDIENCE: But my point is I think you should take it from a broader scope and not be too specific. If you pass regulation you have to look at the full package and look at what the issue is. Kenya was lucky enough to have multiple landing stations and multiple sea cables available and not everyone is in that comfortable position and that brings on a whole new game. I agree with you. I wouldn't make it too specific but that's my point. And I leave now to the rest of the audience before we go deep in to this. Thanks. >> SAM PALTRIDGE: So if there is not any questions from the floor I have another one for Bill. Let's take a hypothetical country and we set up the Internet Exchange Point and we say to all the ISPs you must exchange traffic either here or somewhere else but domestically. If the regulator has a different set of incentives that aren't the incentives that they want a more efficient market and that might be pressure from one source or another to raise revenue or impose constraints on how competitors can interact in the market, do you see any dangers from that sort of solution? >> BILL WOODCOCK: This is probably a trap because Sam is much smarter than me. Offhand, I mean I am sure that really smart people with a big economic incentive can always find some sort of loophole or something. I think keeping things really simple tends to make for fewer loopholes. The more specific a regulation gets the easier it is for people to find ways around it. I think there are a lot of regulators that are pressured to preserve business models for Government owned incumbents or incumbents that are still partially owned. Something came up in an earlier session today that there is often a problem in developing countries where they are coming out of a Government owned incumbent situation. And it is very easy for the Government to extract revenue from the incumbent because they own it and take a profit from it. And when they switch to a competitive marketplace where there are other companies that they don't own, then suddenly they don't have an easy way of getting money out of those companies. They have to turn to an income taxation and they may not be good at actually collecting that. But that's not fundamentally the fault of the Internet or the fault of those companies. That's the Government's responsibility to be able to raise their own revenue and they can't send out the revenue collection folks to get the taxes and that's the Government's problem. That confusion, that mistake has a detrimental effect in the Internet on a lot of places because the Government has this pressure to preserve a business model for an incumbent and that means voice minutes revenue and that means that VoiceOver IP gets illegal and are forced through regulation to purchase services they don't want from an incumbent. There are a huge number of problems that fall out of that situation that are all avoidable. >> BEVIL WOODING: I just want to give the real world example of Belize where two years ago the Government took over the incumbent operator. At that time and still some degree today Belize had made VoiceOver IP illegal. So using VoiceOver IP services was actively blocked on the Internet. And even the regulators said the providers are free to interconnect. And the only landing point in the country was within the compound of the incumbent. And then they also were responsible in large part for the regulations that were in effect. The incumbent was the primary architect of the regulations, not the regulator. So you had the situation where the Government took over and everyone thought well, finally Belize is going to be brought in to the 21st century and that didn't happen. And it didn't happen for the simple reason it was one of the few predictable streams for the Government. Every advisor to the Government if you trouble this, if you open up competition, if you do these esoteric things like create Internet Exchange Points, it is going to be a terrible, terrible thing for your revenue collection. You have a scenario now where Belize has the second highest Internet service rates in the hemisphere. Two, recently the incumbent made a statement well, we have Internet access and we are doing our best to make it available to a wider portion of the population as possible. In fact, it is cheaper than the price of a tin or can of Coke, the drink. And what they were talking about in giving that reference was true. They did have a very low entry point for their services but that service was essentially 56K being offered to schools and school children. And then being told now you can do all the things that your foreign counterparts are doing on the Internet which wasn't true. So this is a very real situation that is taking place now to want the Government to have the regulators' hands untied because the pressure is being brought by the Government in terms of overarching policy. When we have these discussions how big a deal is it for citizens in territories for access between the connection between regulatory policy, Internet access and real economic opportunity is not always obvious or not always as obvious as it should be. >> BILL WOODCOCK: I have got a question for folks on the panel which is are there instances in which you think it is appropriate for a regulator to regulate either the price or the quality of Internet service? >> JANE COFFIN: I will give you an example. I won't name the country because I have great respect for it. But the incumbent came in and said they would offer Internet service for to cost. It was going to be free. What it did to the other ISPs in the country it would have gutted their ability to compete. You can regulate and say we want a competitive market but we can't let the incumbent come in and offer service free in a market where there are entities trying to compete. You don't know exactly I would say that you would have to do a cost analysis at some point there and see where the cost may come out. But I would say that if the incumbent comes in and offers service free that is a problem. >> BILL WOODCOCK: Does that prohibit cost subsidy? >> JANE COFFIN: To prohibit cost subsidy. >> SAM PALTRIDGE: (Off microphone). The instance that Jane is talking about is predatory pricing and it is easy to see but hard to prove. Generally it is better for the regulator to very much focus on the inputs. So things like long prices or how other people can enter the market to make it more competitive. Generally it is a bad thing to try and focus on retail pricing. And if I was critical, for example, of the European Union's regulation of international mobile within the EU area, they have controls over retail prices as well as wholesale prices. And I can well see the justification for wholesale regulation because it has just turned out to be a completely inefficient market. Total market failure with international mobile roaming in OECD countries. I can see what has driven them to do it and the political pressure to have an immediate response to the action you are taking. If you look at the retail prices, they sit just below the regulated retail level where nothing ever happens. And nothing ever will happen because you have got no competitive dynamics going on there. So that's about what I can say on pricing. On quality, again you have to be very careful. Ron likes to talk about the budget model in some countries for communication services and he says what is the most important and he used to be the regulator in Sri Lanka, what is the most important thing for many people in these countries is price. And so if you are providing a service that they otherwise wouldn't have, maybe you can't go straight in and apply the same standards that you would in more developed countries. So there is a tradeoff there. >> BILL WOODCOCK: One of the things that I often think about in sort of the price performance tradeoff I analogize to a ladder and I think that if someone is just using the Internet recreationally, that it is nice for them to have good quality. They have some amount of money that they are willing to spend on their web browsing but this is not ultimately what drives the growth of the Internet. Ultimately what drives the growth of the Internet and the reduction in prices and more services and so forth is when people are able to base their business somehow on the Internet as a utility. And they use the Internet to make things more efficient or provide a better or a new service that wasn't possible before. And so again lowering the barrier to entry means having a very inexpensive service. So people say well, we want to have a very inexpensive, really good service but all that really means is that they are poorly calibrated. They don't know really what inexpensive means or really good. A really good service is going to be a really expensive and a poor quality one can't be an inexpensive one. There are people who have very little money to spend, but if there is some low quality service that they can buy that will enable to build a business to make more money that enables them to buy a higher quality service to work up the ladder and that enables a lot of economic. You are also putting a floor on the price. Effectively nobody is going to be able to provide that service at a quality price. And no one who can't afford that price is going to be able to use the Internet either to educate themselves or to start some new business. So I see a big danger in regulating quality and regulating price because you wind up with these floors that are not low enough in my experience. Any other -- yeah. Troy. >> AUDIENCE: Troy from the states. I just have sort of a beginner's question. When you were thinking about various potential policy approaches for increasing the speed by which traffic travels intra one country, are there any assumptions that are being made by -- about the character of content or about content itself? Are you saying well, we want people to be communicating directly or like if -- if for their content someone is in any case going to have to ping a server outside of the country, then is the IXP then providing actually that much of a savings in cost or speed, or are there sort of content assumption driven policy approaches that are being made? It seems like the discussion so far has been -- it hasn't really discussed about the character of content. >> BILL WOODCOCK: One of the important things about content is that people are often very oblivious to that there are certain kinds of content in which users exercise no choice. And there are other kinds of content in which users exercise a huge degree of choice. And so when you see people choosing to have a Hotmail e-mail account or a Google e-mail account, that's an end user choosing to get a service. If this is a Mozambican customer, as long as there is a market open to provide free e-mail they have a choice of using a huge U.S. company that has no local content provision or a little local company that is going to provide service and customer support in Portuguese and can send a guy running down the street to help you configure your web browser or whatever. And so the consumer is making a choice there and one choice benefits the local economy and the other choice benefits the economy of a completely different country and represents an export of capital. So I think that in the absence of an Internet Exchange Point the cheapest place to host content is next to the Exchange Point that people are actually using overseas. And that's really expensive. Right? Because hosting content domestically would mean sending a query out overseas and back and then the response back and back. And the alternative if you have an Exchange Point is to encourage domestic content hosting, production and the formation of a colocation market, right? So hosting services adjacent to the local exchange and at that point you have a whole new sector of your economy going and generates lots of revenues in developed countries, but a lot of developing countries don't have a colocation and hosting market yet because they haven't reached that level of maturity in their Internet market. Yeah. >> SAM PALTRIDGE: Very interesting question and this is going back a bit in history with what used to happen in New Zealand where Telecom New Zealand, the main incumbent was not receptive to appearing with other New Zealand ISPs, one of which was Radio New Zealand which does the broadcasting of rugby games. And this goes back to our discussion about whether you should take traffic internationally. What some of New Zealand ISPs would do is go across to the U.S. and then back to New Zealand. And if you peered with Radio New Zealand you could get the direct coverage straight away but the traffic they sent to the U.S. they kind of could control the speed of that traffic and so ->> (Off microphone). >> SAM PALTRIDGE: But telecom, it penalized their own customers because they were not exchanging that traffic locally. It meant that you were getting the match -- the report of the match, however much delayed than the other people across the road. >> BILL WOODCOCK: And Radio New Zealand because they had to pay to move that traffic to California. In sort of the early days of hard ball peering negotiation in these national markets, there were a lot of people who did some very aggressive tricks in order to make their case for direct peering and that was one of the better ones. Sir. >> AUDIENCE: Okay. Do you hear me? Okay. So my name is Rastis. I am from National Georgian Communication. I represent regulatory authority. I have a question regarding IXP and namely the remedy, what you propose that domestic traffic should stay inside country and it may make operators to set up Internet Exchange Points. In Georgia we don't have Internet Exchange Point. So I don't understand, to be honest, how will it work. Because situation is the following, we have two big operators. One of them is incumbent. They don't want IXP because it will promote small operators to grow up and it will increase competition. So if we say to our market that, you know, all domestic traffic should stay inside country, what will happen? And other small ISPs will have responsibility to interconnect with each other. They don't have infrastructure. They should go to those big operators, incumbent and another one and ask, you know, now we have access to all 30 other operators. And I think that these big operators will be quite happy because these small operators should pay for transit. They should hire lines and this remedy may kill some of them. And I don't understand clearly how can this remedy make big operators and all operators make to set up IXP. >> BILL WOODCOCK: So the little operators then have the choice, right? The little operators have the choice of purchasing transit and -- so I think your question to simplify it as much as I can is if you require everyone to interconnect domestically but you don't specify how, why won't everyone just go buy transit from the two big operators and the two big operators are happy with that and no Exchange Point gets formed? If someone wants to buy transit from a transit provider in order to meet that requirement, then that's fine. It meets the requirement for them and the transit provider and they are both happy. If they don't want to purchase transit, you shouldn't force one of your new market entrants to be a customer of one of their competitors even if they don't want to be. The incumbents are trying to create a situation where everyone is forced to be their customer. So what you want is a situation where one of the little guys can refuse to purchase transit and say instead we are offering peering to the incumbent and the incumbent may say no but then it is clear that the incumbent is willfully violating the term of their contract. The little guy is not violating it on purpose and are trying to comply. It may not work perfectly everywhere but this is one way it can be done. >> AUDIENCE: Hello. Dimitri from Ukraine. I want to add that speaking on changes in Ukraine the big IXP was supported by small operators who love to interconnect between themselves and they have motivated the large operators in the exchange. Another thing that have happened since last year we had a decline of traffic due to technical issue. It was not able to go with the traffic and that led to the competition to another change. So this other change was competing with the original and that at the end first was upgraded and now they are competing again. The condition between exchange is also healthy. And No. 3 is that now, for example, Google is trying to get in to the exchange. So existence of an exchange is one of the motivating factor for the big providers to come to the country because it lowers the cost and they have to interconnect with big operators. Ukraine is not that large. >> BILL WOODCOCK: Bevil, you have an example of an Exchange Point in the Caribbean that were going to require licenses for participation which would have precluded a lot of that. Can you talk a little bit about that? >> BEVIL WOODING: There were a couple of examples. The Government minister responsible for telecom in one case understood how Internet Exchange Points worked and very intrigued about, one, not having domestic traffic. And two, it would result in tangible economic savings for an exchange being kept in country. And he initiated a process through the regulator for the operators to establish an Internet Exchange Point and they resisted quite strongly. And the regulator had to -- I was part of the process to facilitate the discussions. And one of the operators actually told me say we prefer not to deal with this between ourselves as ISPs, but we prefer that the regulator comes in and deals with it. And so they actually wanted regulatory intervention in the process because they felt that would have been the fairest process to establish the Exchange Point. And in that case they -- the regulator had to adjust the Telecom Act to what Bill said earlier mandate that local traffic remains local. That local traffic remains local and once that Act was updated the fight stopped. And they proceeded toward getting things in place. It took a year to get that Act updated though. It took four months to get the Exchange Point off the ground. >> BILL WOODCOCK: Some we were looking at today. >> BEVIL WOODING: That's still in progress. >> BILL WOODCOCK: Sorry. So we are actually just looking at this draft policy document for the Exchange Point in Barbados that was regulatory driven and it is about five pages whereas most Exchange Point policy documents would be like half a page and a lot boilerplate. >> BEVIL WOODING: It was regulatory facilitated and service provider driven. And the incumbent basically approached the exercise in that case as if it were an interconnect agreement in the traditional telecom model. And so the reason it got to so many pages they wanted to cover the base on all the things that you would normally cover on the interconnect for voice services and because the regulator was not aware of the distinction between interconnecting for purpose of traffic exchange and traditional voice and they allowed the process to get to the point if it were to be adopted would ensure the nonsuccess of that Exchange Point, but we still are at a point where it can be debated. But what happened in that case and I have seen that also in a couple of the other countries in the region where the incumbent has a history of defining regulatory policy on behalf of the regulator. And so they have a lot of latitude as it relates to setting things in their interests, setting things in place in their interest. And the big way to counteract that is for the regulator to have their ears open and their minds inclined toward practices and successful better practices in other jurisdictions that allow them to see beyond the advisor is coming from one source and incumbent dominate provider. >> BILL WOODCOCK: The reason Dimitri's comment made me think of that for those of you who don't know the Ukraine is a huge success story in terms of Internet economy. They have been growing at a very fast rate for quite a few years now. And so this path of development that their Exchange Point have been on is a very good, very reasonable one and may look like there are sort of struggle and conflict, but that's a competitive market at work. Now they have got content providers like Google who want to bring content to them from other places and the policy document that Barbados has would prohibit that because Google is not a licensed telecom provider in the country of Barbados and never would be. Shouldn't have to be and we see the same mistake happening in other Exchange Points. It is not a hugely common one, but there are other countries where to participate in the Exchange Point you have to be a licensed provider in the country. Hong Kong was like that for awhile and then they got rid of that and it helped a lot. Another example of where light touch is hugely beneficial. Not having an extra rule just because you can. Yeah. And in the meantime Jane has something to say. >> JANE COFFIN: This point that I am going to make is more of an access point separate and apart but would facilitate Internet Exchange Point robustness I think. There is the concept for many landlocked countries of trying to get the cross-border connectivity that you need. This is a key access issue. So I am going to throw out a term there is or a phrase collective strength and collaboration. If you are two countries that are neighbors and can work with each other to try and facilitate that cross-border connectivity, this is extremely important. We recently had an example where Zimbabwe and South Africa were trying to bring a cable across the river, the bridge that the cable had to go over was a historical bridge. Customs officials and border patrol and regulators and historical society were involved. If you have the ability and you want to facilitate cross-border connectivity, there are some practical, collaborative steps you can take and as companies you can educate the regulator and the ministry about the time to market. And if you are a company trying to do business and build your network, it is expensive to have your fiber sitting on the board. We are talking about a short distance the fiber had to travel and be interconnected. East Africa where there is a collaborative regulatory process going on right now among the regulators, this is where they are coming up creative ideas for connectivity and that's where we can look at good practices from a regulatory perspective to increase the access and bring down the prices and have a redundant network in and out. If you only have one line in and it is a little tricky, complicated, it is also nice to have that second redundant line, but again you are facilitating connectivity and building your national infrastructure and are helping your neighbor if you have that ability to help your neighbor out. >> BILL WOODCOCK: We have been finishing up a paper. They have fiber across the border to Argentina and fiber across the border to Brazil and you think it would help to get a competitive marketplace going and it turns out not. They allowed the two largest providers in Paraguay, they now have an exclusive Duopsony. There are only two customers in the country that can be customers of that fiber. And those two customers in term price fixed the wholesale rates between themselves. And now this is a new bottleneck that didn't exist before. And so, of course, prices have just stayed exactly where they are and things don't get better and so forth. Yeah. >> AUDIENCE: The case in Barbados was interesting for me and I have a couple of questions regarding it. So regulatory authority established an Internet Exchange Point and question ->> BILL WOODCOCK: In process. >> AUDIENCE: In process. Okay. And owner of this Internet Exchange Point will be a regulator. It will be part of regulator or you made association ->> AUDIENCE: That's part of what is being discussed right now. >> AUDIENCE: So this is first my -- my first question. The second question about financing this Internet Exchange Point. Who financed it? Operators or regulatory body or some Internet organisation or who? And regarding who should switch to this Internet Exchange Point just big operators or all operators or what is obligations? >> BEVIL WOODING: First question in terms of who owns it, that's part of what is being debated right now and how does it function. They are looking at association model and MOU between participants. Part of what was wrong with the document that Bill referred to in the way it was written it made it difficult for others to come in to participating in the exchange, but the idea and the intention is that it will be available or accessible to operators and content providers and anyone who has an autonomous system network. That's an idea from an ownership standpoint governed by the participants of the exchange. In terms of the cost, part of the process in guiding them toward an Internet Exchange Point involved giving examples of the different models for building out an exchange. And one of the agreements in that process was that this would be built out on a cost neutral basis. Participants have decided amongst themselves that any equipment brought in to the exchange will be given to the exchange. Government is playing its part. It sees its role as facilitating national development. And so the parties all agreed to the exchange being hosted at a cost neutral Government facility. I think they are looking at an office space that was set up for what was it? Export development. Recent building proper connectivity. Properly outfitted and the agreement between the parties is that they trust Government facilities more than they trust each other's facilities. Trust being a part of the exchange. It was one of the hot topics or points of debate inside of the process. So the understanding is that we are going to keep our cost down and we are going to do that to find a gracious host and that's how they are proceeding with it. The last point was -- one more? >> AUDIENCE: Okay. So obligation to -- for example. >> BILL WOODCOCK: Obligated to participate. >> BEVIL WOODING: In principle in the Barbados case the discussion went along the lines that everyone should be able to participate in the future. The initial participants are all the local Internet service providers who have their own transit. That's the initial participant, but the intention is the exchange point will be a welcoming place for foreign content providers as well as local content providers that invest in their own Networks as well. Documents as written will not get them but still early in the process that can be adjusted. >> BILL WOODCOCK: Question from the gentleman in the back row. >> AUDIENCE: My name is Sandy. I am from Tanzania. I want to go back to the issue of the content. See, most of these developing countries they are creating a network, a properly designed network, but that network at the moment is working fine. But the problem is there is no original local content that is residing in the network. Do you have any suggestions, any mechanism you can advise the regulator and the Government so as to increase local content production in those countries? >> BILL WOODCOCK: So one obvious thing is if you have local television, national television, radio, newspapers, right, all of these things should be available locally. So national television might be broadcasting rugby or it might be broadcasting whatever. Local jobs. It is content that is going to be of most interest to people in your locality. All of those are desirable properties. Other thing is peer to peer file sharing. It is one of the big bandwidth users, and it is interestingly something where user preference doesn't really enter in to the question, right? So if someone starts downloading a file from a peer to peer system, their client will try several different sources and will choose whichever one gives the best performance. If they are able to reach that seed through a local Internet Exchange Point, that will be preferred. Peer to peer file sharing is one of the things that hugely advantages locations that have their own Exchange Points and sufficient bandwidth to them. You might want to be careful about as a regulator to not preclude peer to peer file sharing and local content curation let's say. >> JANE COFFIN: I would actually urge you to hold a public consultation. I know this is one of those -- can be difficult to do. And not give up if you have done it once but try it a couple of times with stakeholders that you wouldn't potentially normally reach out to. Also any of us, I mean the Internet Society, OECD and UNESCO did a study on local content and the purpose of local content for economic development. There is a direct relationship between local content growth to let the content travel over the network and economic development. You do start to have the cycle of once the network -- when the network is developed and your users are more confident in the network there will be more local business started and they will have more faith in the infrastructure. Don't give up. Keep having them. You can start to spur some of the local business community and this is where a regulator can broaden its base constituents. If there are young people who are trying to develop their own company or don't, there is a fellow in Nigeria who is king of Nollywood. Soap operas, excited family drama. It has become extraordinarily popular. This fellow just received a 2 million joint venture grant. This is content that is produced locally and consumed within the continent and it is driving up the desire for the bandwidth, the increased connectivity within the countries themselves and infrastructure quality of service, if you will. But people are using the infrastructure in a way that they wouldn't have before. So you are going to find there are some creative solutions as a regulator. You have whole different models that are coming at you but it could be an interesting public consultation. You might have to do a little refereeing with the constituents but give it a try. >> SAM PALTRIDGE: Just very quickly because Jane said most of what I said. Don't underestimate popular content. But the key other part I think is public sector information. So get Government information whether it be agriculture, health, whatever. Because it is that local information that people really want. >> BEVIL WOODING: That's where I was going with it. There is a natural relationship between the regulator and the Government and Government in most of these countries is the largest holder of local content, already available content and there are some interesting possibilities as it relates to bringing it on to the local network. One example we like to use is digitizing Government archives. Maybe not the most exciting content. It creates some connection between the Internet and content development. And it is very easy to take photo archives, video archives that already exist in Government information service departments and convert those to the Web and make that process of conversion not a public sector exercise but actually a wonderful collaboration between public sector and Civil Society or public sector and educational institutions and that creates some wonderful opportunities for sparking interest in public content beyond that initial repository. So it is a very easy way to get in to immediate content to the local Networks using existing material from Government archives. >> BILL WOODCOCK: Digital content of library of Alexandra in Egypt is the single largest bandwidth source in Africa to 300 megabits. So yeah, there is all kinds of great content and some of it is produced locally. Some of it is just a local cache. Whatever. It doesn't really matter as long as people can get at it. We have got about three minutes left. I don't think that any of us have any prepared closing remarks. So we can take one more question. Anybody have one more thing you would like us to address? All right. Well, in that case thank you all very, very much. I think we all remain available to you to answer any questions you may have later. And hope you have a good rest of your meeting. Thank you. (Session concluded) ***** This is being provided in rough-draft format. Communication Access Realtime Translation (CART) is provided in order to facilitate communication accessibility and may not be a totally verbatim record of the proceedings. ****