The following are the outputs of the real-time captioning taken during the Eleventh Annual Meeting of the Internet Governance Forum (IGF) in Jalisco, Mexico, from 5 to 9 December 2016. Although it is largely accurate, in some cases it may be incomplete or inaccurate due to inaudible passages or transcription errors. It is posted as an aid to understanding the proceedings at the event, but should not be treated as an authoritative record.
***
>> MODERATOR: Good afternoon everyone. Let's start this session. Thank you all for attending the session both in the room and those who have joined remotely. I would also like to thank the organizers, professors Christopher and you inviting me to moderate this session. I'm Alexandre Barbosa from the regional centers for the development of society which is linked to the Brazilian Internet community and the Brazilian information network center it is a great pleasure for me to be here and for those who know me, having myself a background on measurement and working with our demand side ICT statistics, it is an honor to be here to deal with this important session. And I think that we are going to do a continuation of the last, of the previous session that was really very interesting. Looking at the demand side and the challenge that we face. We have today a distinguished group of experts from different regions and different stakeholder segments. We have representatives from government, academia, civil society and inter organizations and I do hope that this roundtable discussion will further contribute to the debate on how to expand Internet access to the unconnected and how to design effective policy based on data to enable access, in particular for those in remote areas, rural areas, and also to bridge the existing gaps and disparities in many countries in the global south. I would like to introduce our speakers this session, starting with our colleague Alison Gillwald. Miss Gillwald is the executive director at research ICT Africa and the university of Cape Town so we have representing civil society and academia. Rajan Mathews, head of COAI, representing business sector. Professor harn in an Galperin from the university of southern California and also the director at this in Latin‑America. [Speaking Spanish] representing civil society. We have on my left Maria Elena Esavillo Flores. The regulator here in Mexico representing government sector. We have also Mr. Moctar Yedaly, head of information society Africa commission representing intergovernmental sector. And miss Helani Galpaya, CEO Asia representing civil society. Thank you.
Well, before I give the floor to our distinguished speakers, let me briefly introduce some relative and important relevant and important background ideas to our discussion.
First of all this debate is relevant in the goals. Also to recall that many organizations and also some initiatives like the broad band commission acknowledges that information and communication technologies are cross‑cutting enablers of development. In particular, universal access to broad band and mobile services. These are crucial for achieving ‑‑ to integrate and accelerate the progress of sustainable foundation development, economic growth, social inclusion and environmental sustainability. It is also important that the framework makes several reference to ICTs. I'm not going to lease all of them, but just to give you an example, go 5 is related to gender equality that we have just raised this issue in the previous panel. So goal 5 is to achieve gender equality and power for women and girls and let me give another example that is related to ICT. Goal number 9 to build resilient infrastructure, promote inclusive and sustainability industrialization and foster investment but those are the not only ones. It's just an example. Of course the demand side iICT indicators can also help monitor achievement in other areas, such as poverty reduction, education, economic growth and inclusive information and knowledge societies. So demand side data is very relevant to monitor their attainment of those goals that we greeted internationally and signature of those goals.
In summary, governments need to have reliable data to make informed decision and to design effective policies fostering the adoption of ICTs and I think that there is no doubt that data collection and measurement plays a central role in the sustainability development agenda as well as in the design of the effective policies. And here goes the questions. Where is the data? And I hope that our speakers will address these question. And also, I thinking about what demand side desegregated data we have available, especially in the global south, to understand this challenge that was posed in the previous session, to really have universal access to broad band. What are the barriers? Affordability is one issue. Sure it is, but not the only one. Digital literacy? Lack of interest? We have in some countries in Latin‑America, where we have ICT household surveys, barriers indicate barriers on use of Internet. And believe it or not, even in countries like Brazil we have significant proportion of population that they don't have interest or they don't have the needed skills to use the Internet.
Are there cultural barriers or gender issues? Well, the answers are likely to be different depending on country, region, income, race, et cetera.
And also most countries, they have administrative data, supply side data. Is this enough for policy design? Certainly not. Because those data, they are unable to shed light on the demographics characteristics of users and also nonusers. So this is really important.
Without further adieu, I would like to give the floor to our speakers. We are going to have 90‑minute session that will be organized like this. Each speaker will have 5‑7 minutes for their initial remarks and after this we will have some time for questions from the audience. And also from remote participants.
And I would like to initiate our debate posing some questions. You are free to address one or more of those questions. And the first question that I would like to pose for our guest speakers are the following.
Are the information needs of governments, the technical community and civil society been appropriately addressed by current research efforts, and by data producers for demand side statistics? What are some good and also possibly bad examples that address these needs from government, from policy makers. What are the key knowledge gaps in the demand side ICT statistics and how should they be addressed? How can further collaboration be fostered between data producers, meaning academics, multilateral organizations, regulators, among others. And also, data users. Are current research methods and standards appropriated, quantitative surveys, qualitative research, are there opportunities for further coordination and knowledge sharing? Are existing capacity building efforts to at least develop the countries? What are the existing funding sources for demand side ICT statistics. This is very key in this context of producing demand side data. Are resources being used appropriately? Are there opportunities to improve resource allocation? And finally, how can current research efforts articulate with explicit targets such as the ones from the United Nations STGs or the commissions. So those are the initial questions that we have thought of to be addressed by our speakers. I would like to give the floor to Alison Gillwald to dress one or more questions. Alison, please? You have the floor.
>> Thank you very much, Alexandre.
I suppose some of the answers are obvious. I'm going to be speaking from a global south and particularly African perspective. We know that there are enormous data gaps. I mean even on the supply side that you were speaking about where you were saying most countries have sufficient administrative data. Across many of our jurisdictions we don't even have the administrative data. So on that side we already have problems. If you look at the importance of this evidence‑based policy, we can just revert to the previous session we were in. Where it is not only about whether the data exists or not, and you know how it's also about how it's been used. And which data is preferenced over other data. So you've got very ‑‑ epistemic communities that are working around models and policy formulation and evidence in data is critically used above other data that might exist. For example the case was given in Myanmar, in a context where extensive evidence has existed around the failure of universal of service funds and the inability of transitional institutions particularly in a ways like Myanmar to manage those. So you have a double negative on the poor. Not only do you have the heavies pushing up the cost of communications, but you have the funds not administrated properly or given to the poor, so it's a double whammy for the poor. If it existed with learn Asia ‑‑ on the negative impacts and failures of universal service funds. To say they're going to be done let's try to do them better. It's actual local evidence was used there, you could have potentially have had I would argue a far better outcome. So I'm going to sort of push away that sort of nice warm multistakeholder approach that in the last if we all work together we can get this right. What we have a shortage of is actually good, rigorous national statistics. And I'm talking here about you know creating public goods in the classical sense. That all people have access to. I think it's important in the context of open data. And about government opening its data too. But also industries opening their data. So the administrative data that you were saying is available at local level. In fact the operators that are providing this data through GSMA and other agencies at these kind of fora are not providing that data at a local level. So we have very poor supply side data. And then we also of course, the supply side data can't answer in many of the critical questions that we need to answer. So it's particularly in a prepaid mobile environment, you simply not get disaggregated data on income disaggregated or other data without which you have not got the actual correct points of policy intervention. So at the moment we're sitting with essentially reinforcing global indices, national indices on imperfect supply side data, highly imperfect supply side data, and limited, sometimes highly flawed demand side data that has really led to a position where in the absence of demand side data we've had very big industry associations, big advocacy groups on the basis of very very limited representative data, making national global claims around which policy agendas and development agendas are being formulated. So I think we need to look at this. If you look at the ‑‑ the ITU of knowledge is their own problems around delays of time in a prepaid mobile environment by the time the data is being received, it's two years old, the markets have already changed. And the data when we speak about this reinforcing data with ITU, with the world bank, this is all the same data with a different overlay. So we really need nationally representative public in the public demand side statistics in order to understand especially in this increasingly complex environment. So the data that we were seeing around gender for example purely the mobile voice environment, looks completely different once you get into an intimate environment where we've discussed income and skills and that sort of thing. I want to make the point here because I think it wasn't made adequately in the previous session, because even if you are doing these on a nationally representative scale, which most of the references to gender, et cetera, or these micro studies that are not claimable at the national level, even if you are doing them at the national level, at the very basic level out of your household survey or your senseless or maybe even a little more in‑depth ICT access survey, if you leave it at the level of descriptionive statistics, it can be misinforming and I think it really raises the question of is bad information or misinformation better than no information at all actually because it's now directly policy arguably indirect interventions. So the demand side and data really requires, especially to deal with this complex environment, levels of analysis and modeling that tend not to be available even in national statistics offices. And so I'm really asking for a greater collaboration between academic communities in many of the countries that are producing very good national statistics on poverty, et cetera, they're collaborating with universities on the actual modeling and analysis of that data. As I said, because the descriptive data can be entirely misleading. So I think what's important in this environment is for developing countries to see what data is out there and it's obviously not preferencing one or the other, but understanding the complementary nature of the data and how we can leverage publishing and private data that's available into a public ‑‑ into a governance framework that makes this data public and open and available for analysis for further use. And I think this really requires plotting some of the emerging, for example, I know ITU and maybe other people doing a lot of work on this already, but trying to see how big data for example can do some of the expensive fieldwork demand side data that it's very difficult to do instantly as it's able to do, but certainly on an annual basis or something like that. To use that demand side data in conjunction with the supply side data that's improved and better delivered to national regulatory authorities, together with the demand side data. Although it's lovely public data that goes a long way to answering a number of questions, as I said it's simply unable to answer critical questions and many of those critical questions relate to digital any quality. That you have to put the money and go into the field so you can model and analyze it properly. Thank you.
>> Thank you very much, Alison. You've brought relevant aspects of making decisions based only on administrative data. And we have in many countries those data over estimate. So this is really very important. And the problem and the issue that I think is funding a very important issue to have a regular data collection from the side.
Let's move to Mr. Rajan Mathews if you could give us the prospective in India.
>> Thank you, chairman. Again, I'm going to principally from the commercial aspects. I represent the mobile operators in India and you're absolutely right we have statistics which says we have a billion plus connected folks but that clearly doesn't speak to penetration and we sometimes confuse the two. If you look at pure households, there are individuals who are probably at about 650 million. But still a sizeable number. A very interesting phenomenon happened in India on November 8th. On November 8th, the government of India demonetized a certain set of currencies. 500rupees and introduced a very interesting dynamic. And the dynamic was that all of a sudden people found that they could not have access to currency with which they needed to conduct their every day business.
What also emerged was the divide in terms of who had access to the money and who didn't. And those who lined up outside of the banks were the folks who did not have access to a digital payment medium. Those folks who had access via a smart phone or even a high featured phone were able to continue to have their lives go on. And so the people who were lining up were at what we would call the bottom of the pyramid. And it showed principally happens when government intervenes and does things and shows people what has to happen when they are not involved in a digitized economy. And it was painful and it's still going on. And I think this is a critical aspect in terms of showing what was the data. One thing the British taught the Indians very will is to keep statistics. And we've learned that to a fine art. So there's never a lapse of statistics. But what we find is a plethora of commercial statistics on industry, very little personal statistics. And a lack of integration between the two. And one of the things that the government has done is introduced what we call a national identity, which was part of the reason why we had difficulty tracking personal data. Because we had no personal identity. And today India has embarked and successfully have a billion‑plus people in what is known as a database. It's a bio metric driven database at in I point in time, any person can buy a thumb print or retina reading, establish their unique identity. And as a result, more and more people are being brought into the digital economy. So from India, this whole aspect of biometric unique identity is allowing for the first time the crossing of the digital divide. So what it does, is because the government is driving a lot of the what we call transfers, be it transfer money for benefits, cooking gas, whatever, the government is forcing people to say listen, unless you have an online bank account, unless you have these, we will only do transfers on your electronic media, either bank account access through a smart phone, so it's extremely, extremely important from a livelihood perspective to show that is not just merely a theoretical exercise of crossing the digital divide. And what we're seeing is the government, because of the government initiatives is forcing people to begin to start getting one digital literacy and two, getting online in terms of actually being able to do digital interventions, be them bank accounts or any other thing that allows them to enjoy government services. So I think this is one of the things that is happening right now. And is also continuing to be the focus in terms of inclusion and in terms of data.
Now one of the things I already talked about is the personalized data. Interestingly, private operators are the ones that have the bulk of the personal data. And it's all there at any point in time we know a phenomenal amount of information about every user that is on our network, especially as we migrate to 4G networks. It's phenomenal. What unfortunately is happening is we don't know what to do with all this data. Operators are least situated in fact to cull the social aspects. We are commercially driven. So we will use this data for up selling and commercial aspects. So we need to ask ourselves in this information that's how do we cull out the socially beneficial aspects of what is sitting in our databases. There is no government agency today that has the amount of information in the granularity that mobile operators have in countries. So I think this is important. And I think it goes to the issue of static versus dynamic data. I think we have dynamic data, not static data. Because we find out on a daily basis what's happening. I think I'll close with this. One of the things that I often hear is about solutions. One of the things we have to deal with is scaleability. So solutions that are one off. Don't help us. Because we don't know how to market it. We don't know how to sell it. And we don't know how to implement it if it's not scaleable. The one‑offs are good for NGOs and others who will go in and do an intervention and do something, but they never tell us how to scale this up to the 1 billion plus. Because we rely on this and then finally, is what I believe is risk taking. In most developing countries, especially in India, the government realizes that they simply don't have the band width to raise the type of finances and to conduct it to scope and scale of business that is required to do the penetration. So they've privatized it. But unfortunately having privatized it, they've starved industry of the needed resources and this is in spades in India. So they do not know how to factor in risk‑taking into the metrics and into the decision‑making of both regulatory and policy making. And so again, from a developing perspective, I would say that governments, in terms of initiating this divide, need to ask if you had given it to the hands of the private industry and if you believe that they are going to be the engines, not to say that we ignore the government sector, but risk‑taking and return on government are important criteria in the equation that determines how digital divide is going to be addressed. And this is one of the issues we've faced in India in this conundrum we have between profit is bad and we want to collect the next 1 billion. Let's get the billions of dollars that are necessary for investment, ignoring what is an appropriate rate of investment to private operators. Thank you.
>> Thank you, Rajan and I think what you have just said reinforced what Alison has said about the need of collaboration between government, academia, other stakeholders to address the challenge that countries like India, Brazil and many countries in Asia and Latin‑America and Africa faces in order to have universal access.
And also it is very interesting what you said about commercial industry data provided by industry. Because usually it lacks method logical rig or and they are biased. So how can policy makers take an informed decision based on industry association data. He some this is very important. In this collaboration definitely helps in‑depth analysis of these data.
Let's move to professor Galperin. You are a deep knowledge of this situation of broad band in Latin‑America and the challenge that we face in the region.
>> Thank you, Alexandre. The challenge we face in Latin‑America clearly is on the demand side. There's about 200 million Latin Americans who are covered by broad band access in some form, but just are not interested in accessing the regular subscribers to fix or mobile broad band. So clearly we have a demand problem in Latin‑America. If you actually track growth and investment and net worth coverage and the growth in online population in Latin‑America, you see that those curves were growing steadily in parallel until about 5 or 6 years ago when there was a decoupling between the continued growth investment and coverage and slowed down in the growth of the online population. So we have a problem of demand in the region. Which makes so much important to understand demand. Why are those 200 million Latin‑Americans are not interested in broad band access. Is it cost? Is it skills? Is it interest? Is it content. All the issues that were raised in the previous panel and also in this panel. So demand side data is key to understand the magnitude of this challenge and to address this challenge in ways that are cost effective. And let me give you one example. I've been working with the Peruvian regulator who conducts an annual demand survey in the country, large‑scale survey. And one of the questions asked is simply why are you not connected and the answer is actually quite surprising when you track over the years less and less people mention cost and more and more people mention interest. But still cost is a little above, it's about 50% where interest is 40%. That's the general population. Now when you control for other factors and you only look at respondents who have children at home, the cost spikes up about 25%. So this is why in the United States now is being called the homeward gap. Parents recognize having broad band at home is essential for their children's education and they just can't afford it. And on that basis, the regulators started thinking about targeted voucher program for parents whose kids are in public schools to address that homework gap. So this is the kind of thinking outside the box. This is the kind of thinking outside the box that having the demand side data would allow governments and everybody involved in this, to start thinking about this.
Where are we in Latin‑America? I think we are, we make a lot of progress in terms of demand side data. Alexandre has been ‑‑ has been involved in. Of this in levels not only from Brazil but in the regional national level, but we still have many challenges and I'll just mention a few. One of them is that we have some surveys that are conducted by national statistics office, very large scale, very serious service. But sometimes they're just one of service. So they're done one year and they're done whenever there's money available. They're not on a regular survey calendar of the national statistics office. We have a lot of inconsistencies still about even basic things like who is considered an Internet user. Is it three months have you used Internet in the last 3, 6, 12, still inconsistency in service about this simple question. Data is not always public and available, which is always an issue. And also often there is poor visibility and poor coordination of this kind of service and the data. And this goes back to Alison's issue of is bad data better than good data. The irony in Latin‑America is sometimes we have the good data and yet the not so good data is the one that gets visibility. And here's an example. GSM study on gender gaps in mobile adoption. I really like GSM. GSMA, by the way. I respect their work enormously in the region and they did a study that you may be familiar about gender gaps and they included Colombia and Mexico in their 11 countries. Each country they sample a thousand people. Imagine trying to estimate the gender gap in Mexico with a sample of a thousand. It's just not possible. The statistics just don't work. And yet you have survey done by the national statistics office in Mexico which surveyed 75,000 households. So this is an example of one study. We have the good data. The data is there, it's just sometimes not as visible. And there's not the coordination that needs to happen between everybody involved in this to take advantage of the data that is being produced.
So in closing I think in Latin‑America we have great examples of very good demand side data. Satiq is a great example because they only estimate demand side data at the level of household. They recognize that you also have to look at SMEs. You also look at schools. You also have to look at hospitals and all the broader digital ecosystem as it's called. So we have good examples in Brazil, we have good examples in Mexico. We have good examples in many countries of the region but yet many challenges as said remains to be done. Thank you.
>> Thank you, Hernan. I was wondering if we have the required level of data for policy making decisions, this is an issue that we still have to address in many surveys. We don't have the required level of disaggregated like regional, like age, sex, et cetera. And also the inconsist tense issue that we have to face. Although we do have internationally agreed frameworks, we still have some countries with different definitions for not only the Internet users but devices, for instance. The definition of computers, et cetera. Activities and so many other concepts. So thank you for addressing that.
To keep balance I will move to my right now. And I will give the floor to Mr. Moctar Yedaly from the African commission to give your perspective on the demand side data production.
>> Thank you very much. Actually I was thinking I would speak last then I am surprised not to speak the last one. Because back home you see when you speak last, you are the lonely one. And I was hoping that I would be the one. However, I think Alison has said everything I wanted to say. However, I want to underline one thing very important. That I am in a business of preparing and making policy happen. I'm advising head of states on how to do policy. And any policy that should be very good, it should be evidence‑based. It has to have data. Otherwise it will be a very bad policy. And fortunately most of the time ‑‑ unfortunately most of the time we do bad policy because of lack of data. So data is very important. But data need to be very reliable first. Data has to come on time. And data has to be appropriate in the sense it has to be tailored to the need and the specifics of that country. And I will cite two examples in my dealing with data and policy in three paradox that are currently existing.
The need of data that are actually worked out is sometimes we assume that the techniques that are being used, the approach in conceiving designing data is universal. Scientific and universal. Which is actually to my view not true. You cannot assume that any teacher in developing country who can afford himself who has a broadband home, you assume that any teaching African can afford that. So for the purpose for costing, if you say that in Africa you have a thousand teachers and therefore those thousand teachers can afford themselves a home, which is actually happening in developing country, then you are wrong. That are not maybe the same thing. You cannot assume that data provide by the world bank and ITU that are actually two years back are the one appropriate for somebody who wanted to take something reliable to move tomorrow. And you cannot assume that collecting data is always something that is, how can I say, reliable. Because when you start to collect data within a country, people who are providing data are seeing that data as analyzing their performance criteria. They will give you data from the government side because you think okay, he's requesting data from me and I don't want to show him that I'm doing my job very bad so I give him the good data and therefore I don't get the other data. But if I to you world bang ask them they are not controlling me, then I can give them data. But in the same time you have a paradox that people who can pay or afford for data to collect them, can access to data and those who cannot pay for them can access to them. And I did two examples. One of them I asked using the new technologies of GIS and surveys and so on. And ask that people to provide my data. I didn't get anything. I went after that if you provided me data on time you get an iPad, I get all datas. That's most of the people do to get data. That's probably not questionable at all. We need to know if data are really reliable, viable, can be used to have a policy‑based thing.
The paradox number 3 when I came to collect data that our supposed to be reliable coming from the U.S. or coming from the world bank, and I come to my member state and I see okay, lately I was on a council from the African union and it was advocating for cyber security matters and I was providing example how cyber security is affecting households. And the majority of my datas are coming from the U.S. I'm told very good your data are coming from the U.S. I need data from Africa. Go and bring them to me. Count your data, there is no policy. So this is where the mix of challenges and paradox in design are happening with regard to data. What we should do now. First of all, we need to employ those local people who are collecting the data. It's not data that are collected by somebody outside of the continent who doesn't know the specifics and interactions in the social context. That probably he's got to have the well‑designed and good.
And number 2 the social majors ‑‑ service technology combined together today can provide some very reliable data.
Number 3 developments need to take as possible to first the fact that data need to be provided on time. So this is what I wanted to contribute. And thank you again for giving me the floor.
>> Thank you, Moctar. You've brought important aspects. One is the timing that is needed to policy makers to have the data. If you take two years to produce a data, it is maybe in the technological field not be very helpful anymore. And also the accessibility in terms of having the data accessible to our stakeholders, especially to have access to the micro databases, to make a more in‑depth analysis or to cross datas in different ways that usually when you have only the aggregated level data, this is not possible. So this is very important.
Let me move now to Helani Galpaya. You have the floor.
>> Asia is not one Asia. And we have such a diverse set of stories on how data is used or not used in evidence‑based policy making. But let me address Rajan, which gives me my first talking point. You said you have so much data and you don't know what to do with it because there's so much personal data. We know this. We have historical anonymized call records in Sri Lanka. We are using it to identify where poor people live, to identify how cities are changing because it's very fast changing from residential to commercial or mixed use areas which national surveys cannot identify because they've done censuses every 10 years. We are identifying where the poor people live, how much and poverty mapping and poverty indices. This is valuable social economic data that is used and usable for policy making. We advice the western megapolis planning project in Sri Lanka to look at where people live and where they travel to. This is useful to see where should we put bus routes in the western region of the country. That's proper use of your data for policy making. It took us two years to negotiate this data as an independent think tank and we had to sign nondisclosure agreements that would put us in jail if we put up a slide without the approval of the operators. This is a huge problem. Yet at the same time operators in the region, without any conditions, give up their data to developed world units and researchers and think tanks who come in and do analysis and publish papers which are absolutely wrong because they don't understand the context. So they will use 25% of the data and say oh, this is what's happening in Bangladesh when in fact if you understand what that region is, you will say that's actually completely wrong. There's a huge problem to access to data by southern researchers and this is a real problem. That's my first point but it can be done.
Second, some of the data that's been most effectively used to change policies have come from really expensive household surveys, which thanks to Canadian government or British government willing to fund it, we were lucky enough to be able to spend that kind of money across the region. 11 countries to get national ICT personal household level use data. They're no longer in this business, so funding is a real problem. When the SRI Lankan government introduced an aggressive tax on mobile phones, on the SIM cards, a flat tax, we had, thanks to household survey data, evidence to show that the mobile phone is not a luxury good. That's really stupid thinking and that this tax would be a 25% cost increase on very poor people because we could look at the base of the pyramid expenditures which the supply side cannot do because they don't know if it's poor people or rich people who are spending this money. We were able to successfully role back this tax. We now no longer can afford these types of data.
When Myanmar for example is doing its planning and some people are talking about the huge gender gap and how south Asia has a gender gap in mobile phone ownership and southeast Asia doesn't, our national household survey showed that despite being in southeast Asia, Myanmar has nearly 30% gender gap in phone ownership. This comes from household data which no supplied side data, even in a country with decent SIM registration rules will give you. Because it doesn't really work you don't know which gender and so on. And we also know that in Myanmar from the surveys, that 66% of the population has a smart phone. And only 3% has a feature old phone. Why is this important? Because Myanmar is different to India. The interventions we plan in Myanmar should be based on fancy phones we can do apps and funky things. It's not the solutions we should be recommending in India. These are the tapes of data that we get if we really study uses. When we talk to women in Myanmar and we look at the representative data we know that women own less phones as I said but they make more calls and talk more. So there is no gender gap in access to phones, but there is a huge gender gap to Internet usage. These are nuanced understandings you get only when you talk to users and in meaningful ways.
The Indian decision to ban differential pricing was an example of a debate from developed countries on net neutrality, coming into our part of the world. The regulator being unable to assess the evidence on the ground, that despite all the screaming and shouting including by the civil society, that this was in the short term not a huge problem and it was in fact helpful. That didn't mean we should have given a free pass to all types of differential pricing but that we needed a more nuanced approach. So we had problems of access to data. We have regulators who are unable to assess a variety of data and make use of it in their decision‑making. We have a problem in funding of good research in our region.
>> Thank you very much, Helani, for highlighting the challenges in having access to private data sources, especially operators. Two weeks ago we had the symposium on ICT and telecommunicatetors this was an issue that was raised how to negotiate with private data sources. We have a lot of challenge like privacy and confidentiality. This is the main barriers preventing access and of course commercial interest and costs et cetera. And I also thank you for highlighting the need of household surveys.
Now we are going to move to our last speaker. Miss Maria Elena Esavillo Flores from IFT in Mexico. Before I give you the floor, I would like to highlight that Professor Galperin has mentioned in the national statistic office here in Mexico is one of the very good practice in the region and in terms of collecting the ICT statistics, Mexico has been providing service in regular basis. On the national household has mentioned about 70 or 80,000 households. A very large sample. So with such a large sample, it is possible to collect segregated data at various level. So I guess that we have a presentation. So Maria you have the floor.
>> Thank you very much. It's been very interesting to hear all of your interventions. And I'm going to ‑‑ no.
Now, about what we are doing in the federal institute of telecommunications. Because when the institute was created, we found out that we had some very poor indicators. So we started to work on that. And since 2013, when we were created, we standardized all our indicators according to ITU definitions and best practices. According to the organization carriers have had to make very big efforts to change you al their reporting to the IFT. Not without resistance, but finally they are working a lot to so that we now have more reliable information. And that we have comparable information among the same carriers and with other countries that we didn't have before. So this is really a huge change for us. We have now also electronic system to receive the data, to possess it. So it is easier done for us and for the operators. These centralized database is also available for public use and this is also news on how we are working. As part of these changes, we now publish data by carrier. To do this we have to have a new interpretation of what private information is and what public information is. Because that was the reason why before we didn't have this division by carrier. Since these are public services, we are interpreting these are public data.
I have some just some of these data that we have been collecting to have an idea of what has been happening in Mexico. This is the change in household expenditure and mobile telecommunications related to income by different life styles and we can see that how the average expenditure has been moving have been declining and mostly for the first ‑‑ we believe that this is coming from important reductions in prices.
This is, we're looking here at different bubbles that represent different cities in Mexico, different states in Mexico and how we have so big differences in access and that are related to educational level to income and this is also related to subscription penetration as you can see. The farthest to the right bubble is Mexico city where we have very high levels of penetration, almost 90%. But then we go to the other side of the graphic and we see states which have very small penetration. So this is part of our challenges to make this bubbles get together.
This is the drop in telecommunications prices. This has been quite dramatic in the country. The deeper greener part is the reduction in telecommunications prices. The other one is average prices.
And here we see the difference in pricing, but among different telecommunication services. Where most often decline has been in mobile, the great part is because we now don't have long distance calls. So it is 100%. But really the most important reduction has been in mobile services that what we also have to look at the other services which have not had the same reductions.
In fixed Internet services, you can see how the subscriptions have been mounting. This is for fixed, but in the next one, mobile Internet services, we have had very ‑‑ a very big increment of subscriptions from 7% in 2011, now we have a 54% of penetration.
And now these are maps of coverage in mobile services for different technologies and here we have a big challenge. This is 2G. Now you can see 3G in the next one which is around the same. The same coverage as 2G. But now we are looking at LTE and the next one and this is where we have to work. We have now this new project, the shared network that has just been auctioned. This is a great project for a country and we believe that this will help to improve this LTE coverage.
In the last one, I wanted to talk a little bit about these efforts that Alexandre brought about. This has been an effort. We are funding these efforts with three institutions. Our background is that we had the ancient model of the nationalal survey of occupation and employment that looked at telecommunications services. This function in a way that the interview was done to the person who opened the door and often this person was not the ICT user. So the data was not very good. It had only national wide statistics. Since 2015, this changed. To have now the ‑‑ [ indiscernible ] and this means that the model is now an independent survey. The biggest change was now the interviews is done to the ICT user. And to all population, covering all population other than 6 years. And the objective is to know the availability and expenditure on computers, TV fixed mobile telephony and Internet connections. The sample is at over 90,000 households and we now have statistics for national level but also for state level and for the biggest 32 cities in the country. This was for 2015. For this year, where the first results will come out in this month in December, we will have an increase to 49 cities, that is data specific for these 49 cities. And next year we will have rural and urban segmentation. So we're working a lot in these efforts to have better data in supply side and demand side. I would also say that in. IFT we are conducting periodic service to consumers, not such as solid or wide as DM ‑‑ but these are also helping us to understand what the demand side is its behavior. Thank you.
>> Thank you. This is a very good example of cooperation between regulator and the national statistics office to come up with important stand alone ICT survey. In Brazil we also have collaboration between regulator NSO and also Satiq which is an independent data producers to produce relevant statistics for policy makers. So now we have a few minutes for questions and comments from the audience. I would like to open the floor. Yes, place. Say your name and organization.
>> Hello. I come from the ministry of science and technology. Thanks to everyone. It was a great workshop. I'm just going to make a comment and I wanted to build on some of the arguments that have been already presented. And I wanted to return to the title of the workshop that is connecting the unconnected. Where is the data. And I believe that the data sometimes is not inside our telecommunications sector. And I believe this because returning to the previous comments, if we wanted to connect the people, first we need to know if the unconnection is due to an affordability barrier or digital literacy barrier or an infrastructure barrier. But if we go specifically to the affordability barrier, maybe we don't have the information. And I wanted to comment that the case of recontact, because in in our national telecommunications development plan we have a program in which we try to address this problem by giving the people a laptop and a subsidy only to the families that live in a poverty situation. And these data is not in the policy maker. It is not in the ‑‑ but we do have it in the country and it's part of the data of the social welfare office. So woo did we do? The ministry that's been in charge of the public policy. We defined a program and then the regulator is the one that takes care of implementing the program. But they don't have the data. So what we did is as a government we coordinated with the welfare office. So they identify the families which are in moverty and they gave the ID number to the service providers. So if I'm a person, a member of a family that lives in poverty, I go to the service provider, I give my ID number and they give me my laptop and my subsidy for paying for the Internet. So the service provider doesn't have to ask me for my personal information or my economic situation, because the welfare office already went to the home of this family and they checked and as the experts they are, they said yes. This person lives in poverty and they do need a subsidy and a laptop. But we couldn't have done that if we went only to the information that is in the telecommunications sector, which is the one generated by the service providers or the one that would have in the ministry from the users. We need specific data on poverty. So in the end as a summary, I believe that it's very important to use all the data that we have as a country. And not always the one that is inside the telecommunications sector or even in the one in the national statistics office. Thank you.
>> Thank you very much. Any other comments or questions? Yes, please.
>> [ Indiscernible ] I was surprised by the comment by Hernan Galperin in relation to the quality of the research done by the GSMA. With all due respect I've been following up the outputs of the DSMA at least in South Africa and the African space and at the when you talk to other Internet service providers, it doesn't seem as objective as it could be and the interest of a given set of providers in the continent are not that much the public service of what science and evidence policy should be based upon. So I would like some clarification on that comment. Thank you.
>> Thank you very much. I will take one more question and then I will pass the floor. Go ahead. You have the floor. Please, go ahead.
>> Sorry. There was comments ‑‑ I'm Clara ‑‑ comments made on our report and I just sort of wanted to respond. I think our report, we were very also surprised by how much our particular the report you mentioned was red and I think it highlights that this huge lack of data on this issue. And that's why I think it has been absorbed by so many people. And I think it's also, I think there's a need for the this to be collected by the national statistics offices to do this. As Helani mentioned there's an issue around cost so we were limited to the number of surveys we could do because of budget issues. This data needs to come from those who are doing national surveys, so that others like us can focus our research budgets on answering some of the questions that are not answered. And then as another small point in that, the modeling we did on the gender gap included also more than just the thousand surveys. But I think my point I wanted to make was just sort of the need to have this national data funded by collected by the national statistics office so the rest of us can focus on where those gaps are and there's a real gap on gender data at national level.
>> Thank you very much. Hernan would you like to make any comment?
>> Yes, thank you. I think the GSM representative responded well and my comment was simply when you have two studies to address the same question and they have such different methodologies and sample sizes, inevitably you're going to try to compare them and one has ‑‑ I mean the estimates that are coming from one won't necessarily be better than the other. Again I said this with my deepest respect for much of the work that GSMA does in addressing the question. I think GSMA does a lot of its best work is on other issues, on many of the other issues that they work and in this case my sense is trying to fill gaps and trying to participate in a very interesting conversation about gender gaps and mobile adoption. So my point was highlighting that sometimes we have missed opportunities. Every situation is different because in some countries, the GSMA and other studies may be addressing information needs simply where no other data exists. So at least there is some reference point for the conversation. But in many other countries, and this includes most of Latin‑America, we've come a long way in terms of having solid statistics that can help us address and estimate what are the real gender gaps and many other gaps with very solid data.
>> Thank you. Moctar?
>> I would just like to highlight something and it probably is not the purpose of our workshop. Yes, we have spoken about data for the purpose of making our cost and planning and so on. And there is a lot of challenges, specifically in Africa specifically in collecting data. But once we pass all that and we collect data, most of them do have some specific personal strategic data. And we need to think about how to protect those datas. There is a lot of data that are being provided by citizens voluntarily in exchange of free services. And while those people think that there are clients every customers, they are becoming a product. And those data may be used, abused and misused. Not for the purpose of ‑‑ I'm not talking about privacy. I'm talking about data being used for specific strategic cyber security matters that are not really in line of business imperatives. And that's something we probably need to think of and I think the way we design data, the way we conceive them, the way we collect them, and specifically the way we store them. Thank you very much.
>> Thank you. Any other comments? Alison? Please.
>> Thank you. I just wanted to return ‑‑ I think the point was well‑taken about the fact that because ICT's increasingly cross‑cutting that it's no longer just a telecom sector thing I think that the example Helani was giving for big data being used is absolutely critical and I think it has to be as we try to deal with greater complexity we of course are talking about the need for this mobile data informing all sorts of other planning and stuff. But I did want to return to the impact of the available data that is there and how it is positively or negatively impacting on policy and development outcomes. Because I think in Latin‑America it's unusual that there's lots of good data. I think in many of our environments, if the IBRC funded surveys that we've been conducting in Asia and Africa and Latin‑America, but less so in Latin‑America weren't done, when they were cut off, they stopped being done. I'm very happy to say that IDRC will be funding in 2017 limited household surveys again I think the issues that have been raised around you know, the wealthiest industry association industry in the world receiving fundings from the big donors to go out and do non‑representative studies and to determine development agendas is really problematic. I think if we're going to be talking about collaboration and multistakeholderism, we're going to say who's paying for the data, what are they paying for it for. I think it's absolutely fine for GSMA to say we're doing studies overtly. Because the markets are saturated. The markets can grow and I think that's made quite openly. Nobody is pretending that's not the agenda. But then to say that is also the development agenda because there's no other data saying well actually the policy interventions are not around the descriptionive statistics that are showing these inequities but are around the fact that more women are concentrated in the bottom of the pyramid without education and income so that's determining so there's no ICT. It's actually a social inequality that requires a much bigger intervention from government in terms of education strategies, keeping girls in school and all sorts of things. Not giving them pink cell phones or discounted packages or things like that. So I think these are really important things to put on the table. And actually say a lot of the advocacy groups, there's a lot of money. The big platforms do not give money to public research, institutions. Certainly not in the developing country. The money that goes from these big platforms goes to the big advocacy groups. So there's lots of money for advocacy, but there's not very much money for public interest research, public universities or NSOs so I think we really need to say who is paying for it, why is it being paid for and what are the outcomes in terms of our policy agenda.
>> Thank you, Alison.
Well, we are running out of time. I just would like to add to what Alison has said that transparency technology and the method is crucial to understand the quality of the data. I usually say that the methodology is the birth certificate of the data. So that you can understand the identity of the data. If you are transparent with the methodology, that's okay that we have diversions in numbers and estimations.
Well I just would like to ask 30 seconds for each speaker, because Carla from Costa Rica has mentioned something interesting. The data is not on the operators. So where is the data? So if you could please give a very brief and short answer. Where is the data? Let's start with Helani and then Moctar.
>> I think some of the most valuable data is with the operators, so I disagree. We just aren't making maximum use of it. And the technology, the algorithms and the capacities we have to analyze that data and really make use of it for development and socioeconomic purposes will only increase.
>> Thank you. Moctar?
>> I will answer the other way. For sure data is not within the government. That's for sure. Now where is the data? My belief is everywhere but the problem is there is no democratic access to data. That's the issue I see. They are everywhere and unfortunately those who can't pay for them to collect them or to store them are the one who has data. So as Alison has said, we need to know exactly how the public and those policy makers will need the data for evidence‑based policy, need to know ‑‑ we need to set up how those can access to those data.
>> Thank you. Maria Elena? Where's the data?
>> Yes. I believe that we have data everywhere. Our challenge now is to have the capacities to process the data, to know where the data is. And also in parallel to assure privacy of personal data. And awareness by the public who is sharing their data.
>> Thank you. Hernan?
>> It depends on your question. If you want to know something about those that connected, the private operators know a lot. If you want to know something about the unconnected, they know nothing because they're not connected. So it really depends on what's your question.
>> Thank you. Rajan?
>> I think very emphatically if you are talking about personal information and data, it's with the mobile operators and it has to do with country specific. Your country may be very different. Please understand in India we only have one network. Most of the other countries are at least 5. You have land line, satellite, you have cable, you have private networks, government networks. In India, one. So if you want it, you have to come to the mobile operators because that's where most of the information flows.
>> Thank you. Alison?
>> So I do think there is's critical data and we want it for mobile operators. But I think in a prepaid mobile environment particularly there is a lot of data you cannot get simply from the big data or even the supply side data from operators. We have to, we've got to go into the field and collect it. And I just want to make a final point because I think a question of that is really in our country, in our countries in Africa, there's very little data like the others. How do we find sustainable ways of getting those public data so that we're not biting our tongues at done or meetings because we want to make sure we get the kind of funding that matches their agenda or aligning whatever they're doing with some other advocacy group just to get the data. How do we create good public good statistics. And I really want us to think seeing that we are at at Internet Governance Forum to think about emulating or get some kind of commitment to the model we have at Satiq at Brazil which is using the demand name sources, but you know hopefully in time we well get Africa to Africa and perhaps the smaller countries that will never be able to afford the fieldwork, we will actually get a fraction of that public domain for statistics.
>> Thank you very much. We may disagree many aspects on answering where is the data. But truly believe that we have a convergence idea that relevant demand side data is crucial for policy making decisions. And despite the growing debate on measurement and in statistics production, on the ICT access and use, there is a lack unfortunately of systematic and liable statistics, production. By the lack of data in nations that can be used in development of information society and to the universal access to broad band. So I think that we can leave this room with this idea that producing demand side data is crucial for policy making despite our different views on where is the data. So with that I would like to thank all of you that has attended this workshop. And give a round of applause for our speakers. Thank you very much.
[ Applause ]
(Session concluded at 13:30)