Welcome to the United Nations | Department of Economic and Social Affairs

FINISHED COPY

NINTH ANNUAL MEETING OF THE
INTERNET GOVERNANCE FORUM 2014
ISTANBUL, TURKEY
"CONNECTING CONTINENTS FOR ENHANCED
MULTI-STAKEHOLDER INTERNET GOVERNANCE"

5 SEPTEMBER 2014
11:00
WS 194
NEW ECONOMICS FOR THE NEW NETWORKED WORLD

 

 


 
***
This is the output of the real‑time captioning taken during the IGF 2014 Istanbul, Turkey, meetings.  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 session, but should not be treated as an authoritative record.   
***

>> MODERATOR: Hello.  Can you hear me?  So there are not so many people in here.  You're welcome to come and sit in this big, nice round square that we have here.  Much more fun.  Unless you want to sit and do something else, you're welcome to do that, too, but if you want to follow discussion, welcome to sit around this table.  We will try to make this pretty interactive, too.  
So I'll give you a short presentation of who's here, what we're going to talk about and also maybe a bit of the house rules for this session.  So we're going to talk about what economics is going to look like in the future.  So the problem that we have today is that we have a whole lot of economical models and matrix that we, kind of anyway, thought worked in the world as it was.  But, you know, things are happening so fast now that we don't really have the necessary metrics or economical models to discuss what happens when somebody does something like Wikipedia or what happens when somebody introduces a smartphone like this?  And it is of some importance because OECT and other good economists are advising governments, making decision to what to do with the substantial part of the world economy based on models that maybe don't apply as much anymore as they used to.
And so the question is, basically:  What kind of metrics and what kind of models and maybe even what kind of governance will we need in the future to be able to run our economy in the best way?  All right?  
So with us here today we have Svetlana Maltseva, she is the Acting Dean from the Faculty of Business, Business Informatics at the Higher School of Economics in Moscow, Russia, where she also is head of the Department for Innovation and Business and Information Technology and Deputy Director of the Institute of Information Technology.
And with her, her colleague Mikhail Komarov.  He is the Deputy Dean for International Relations of Faculty of Business Infomatics at the Higher School of Economics in Russia and also Assistant Professor at the Department of Innovation and Business and Information Technology.  
And to my right here we have Michael Kende who is the chief economist of the Internet Society in Geneva.  
And here we have Rudolph van der Berg.  He is policy analyst with the OECT Division for Digital Economy Policy and the leading expert of machine‑to‑machine and Internet of things and all these things that are very hot now.  
Yes, next is Helani Galpayawho is I hope I pronounced this right, LIRNEAsia CEO.  And she leads multiple research projects and telecom, agriculture, government service delivery and so on.
And then we have, yes, Roslyn Layton.  She is PhD fellow Aalborg University at the Centre for Communication Media and Information Studies.  And she is into political economy.  
And at the far end there we've ‑‑ do you hear me by the way when I speak like this?  It's okay?  Good.  At the far end there is Mike Nelson, who is Adjunct Professor at Georgetown University and also the leading policy expert who, among other things, wrote a good deal of Gore's legislation back then.  So veteran here.
So the question is today what is it that economics does not measure today that is important?  And what does it need to measure?  And maybe perhaps even how can it measure it?
So I'd like to give th­­e word first to our Russian friends who have to take off.  It's about half past 11 to get an airplane.  And we're going to run five minutes per talk.  And it's going to be hard stop.
So, okay, please.  Svetlana?
>> SVETLANA MALTSEVA:  Thank you very much.  Many thanks for the invitation to participate in this very interesting discussion.  I must say that we work in maybe understanding the Internet and the Internet economics very much because our business informatics is the discipline in which of course we must think about, first of all about business, about enterprise and IT, the role of IT in enterprise activity.  But of course in reality, Internet information is a part of decisionmaking now.  And big data project, open data project, all those projects change models of making decisions in the enterprise activity.  
And I want to tell about one of our research that was conducted together with W3C Consortium.  And I must say higher School of Economics is the host of this consortium in Russia.  And it was very large, very interesting research for different Russian enterprises.  And the use of open data, data from Internet sources for decisionmaking for enterprises in different areas, such maybe critical areas for Russia is oil area, oil/gas area, transportation, agriculture and so on.
And maybe I have only, I know that I have only five minutes.  And aircraft is waiting for me, so I will maybe only brief results, some brief example from this research.  And I want to give for the example of using an approach which originally was developed for Internet data to meet the challenges of enterprise systems.
I must say that we investigate big data using and find that, for example, Russia first of all, big data technologies different level, on different level of maturity is used, first of all, in marketing strategies and in finance and accounting.
And now I want to tell about how the link to data project can be used to meet the challenges of enterprise systems.  Link data is about using the Web to connect related data that wasn't previously linked or using the Web to lower the barriers to linking data currently linked using other methods.
Data integration in enterprise is costly long lasting and challenging problem.  Legacy data and business‑critical information is often given and integrated in such systems like ‑‑ supply chain management, one minute?  Oh, okay.  Then to the end.
Mainly information integration challenges.  Enterprise taxonomy, portal intranet database integration, data fusion and linking for knowledge base.  And linked data, linked open data provides the following benefits:  Open standards based on ‑‑ and vocabularies.  Search and semantic content extraction.  Lightweight data integration and several research projects.
I must say that in my finding of this research was the new chain model that can be used for model of benefits of serum system.  And if you want to know more about these results, you can write to me or to my colleagues and maybe to the Department of Business informatics.  Thank you very much.
>> MODERATOR: Thank you very much.  Five minutes.  And we got to know that big data, open data paradigm, linked data and the Web, these will be sources for metrics and building models.
Okay.  Mikhail, your go.  
>> MIKHAIL KOMAROV:  Thank you very much.  First of all I'd like to tell you about statistics of Russian Internet‑based business.  So we've got 4 percent of Russian GDP going to Internet‑based business.  In fact, approximately 4 million people in Russia involved in the Internet‑based companies.  And in terms of market share, in terms of numbers, we've got 120 billion dollars for Internet businesses.  And it's really, you know, good numbers in terms of the country.  But what we are talking about now?  When we are talking about market shares and different types of Internet‑oriented business, leading, let's say Internet arranged business is connected to transactions.  Internet transactions, payments.  Because it's kind of ‑‑ for Internet economy.
On other sides in terms of Russian market, we are talking about fast and quick development of game market.  So games for the Internet and so on.  Online traveling.  Because they are leading also now in Russia.
From the other side I would like to say that Internet‑based business processes when widely in different areas.  Smart phones and development of smart phones mentioned the description of the workshop created new mobile applications markets.  So it's already changed economy and it's already changed Internet economy.
From another side, when we are talking about Internet and let's say connection between the countries, so Internet is transnational, right?  But from the perspective of economy, we are talking about tight connection between the countries because we are talking about data exchange.  And that, I think, in my opinion, metrics for the new economy.  We're talking about data processing time.  From the same perspective as payment processing time.  At least from Russian perspective.  So we're talking about transactions, we're talking about timing.  And you, let's say, new resource, huge resources was mentioned we are talking about data, we are talking about big data, structure date and processing time is a key and crucial point there.
In terms of models, a description of the workshop David mentioned share economy, cloud‑based economy.  And in fact we are talking about policy changing in terms of this new ideas, right?  Because we still don't have really, let's say, clear models for profit sharing, for cloud‑based projects.  Because somewhere these projects connected through the Internet, they are building some things which could be repeatedly produced, right?  But in terms of profit sharing, in terms of transnational, let's say, characteristics of the Internet, it's not probably the same as building some real devices.  So we are talking about services.  We are talking about services which could be used by millions of people.  And that's when why we're talking about profit shares, there should be some new policies developed.  
And also I would like to focus on open government data because in terms of open data, or big data, open government data has key points there.  And the key are on, let's say, benefit.  Open government data is reliable data.  And we are talking about new businesses based on this, on open government data.  Quickly developing now.  Because it rely, it's reliable and services provided by companies use the data as a source of information also providing reliable service.  So these are my thoughts about what's going on with some examples from Russia.  Thank you.
>> MODERATOR: Okay.  Thank you very much.  Since these guys are going to the airport now, I don't know if anybody has some immediate questions for them?  Mike?
>> MICHAEL KENDE:  Just real quickly.  I'd be real curious what your sales pitch was?  When I was at IBM lots of academics said give us funding, we can do this magical thing for you.  So when you talk to companies, whether American companies or Russian companies, what examples do you give them to show how your analysis is helping them run their companies better or develop products?  Are there specific economic data or models that have been helpful for the corporate sponsors?
>> SVETLANA MALTSEVA:  I must say that our faculty works with different companies from not only from Russia.  First of all, it's companies in IT sector, Microsoft, IBM, EMC now we have good connection with ‑‑ those companies, ECP, of course, because those companies lure us to maybe use the most interesting technology for data analysis, for enterprise management and so on.
But when I told about of course due to this cooperation, we always use the results of different research of benefits from IT for business for different companies across the world.  But the research that I told about is the research, first of all, Russian companies because it was interesting for us maybe to implement new decisionmaking models to Russian sector of economics.  And first of all we work with big business because we investigate the ability of big data technologist as innovation and find that only big companies now can implement these technologies.  And so we look to big business, of course, to oil, gas, transportation, maybe education because education is now big, large organizations, and other industries.
Of course we use this in research some results from research of leading companies from around the world.
>> MODERATOR: Thank you.  Okay.  I think you said you need to go in two minutes?  So I hear you saying that there are coming a lot of new economic tools for companies for using big data, linked data on the Web and so on.  And well, I have several questions but you only have two minutes.  One would be the kind of split between macro- and micro-economy.  It's going to take a bit longer for government to start using these things, right?  What do you think?  Is that going to be as fast as the companies?
>> Or just to add to that question.  What kind of economists are you training that IT companies want to hire?  What are the special skills that the new generation of economists need?
>> MIKHAIL KOMAROV:  We said some key meeting points, we are talking about the future decisionmaking processes and improving decisionmaking processes.  So because for the decisionmaking, we need to analyze data; right?  And so we need to, let's say, make some conclusions.  And it's based on some mathematical algorithms, as well, of course, based on the reliability of the data we have.  And answering the first question, our government would take it faster than companies?  I would say business is taking some new technologies, new projects faster than government.  After the government introduce some new policies, right?  Answer.
Answering the second question in terms of kind of IT economists, I would say that in terms of faculty or business informatics, we are oriented on business processes.  So we're trying to teach students to understand business processes and transformation of business processes according to the transformation of technologies.  And it's a key point.  From one perspective.  And from other perspectives we are talking about some, let's say, basic standards, some traditional standards for financial management, for company management, for business management, in fact.  And this combination actually makes our -- well known in the Russian market and outside Russia, as well.  Let's say they kind of approach for new transformation business models around the world.
>> MODERATOR: Thank you so much.  So, Michael, Michael?
>> MICHAEL KENDE:  Okay.  Good morning.  So I wanted to talk a little bit about the app economy, the economy for the apps on our phones and tablets as a symbol of the changes that are being brought about by the Internet.
So some high level numbers that are known for 2013, 102 billion applications were downloaded, or there were 102 billion downloads of applications with 26 billion dollars in revenue.  I'm not sure if that includes what went to the app stores, the Google and the Apple app stores.  
But if you hear those numbers for individuals and companies or developers, it's a very attractive sounding, right?  Low entry barriers.  Anyone can download the software to make an app, growing platform.  It's global, low distribution costs.  Governments are obviously attracted to this.  It's good jobs.  They're flexible.  It's kind of a cottage industry of people making these apps, and some of them have some real social benefits.
So I bring this up -- in my old job as a consultant, a trade group hired us to write a paper about the benefits that we would present to a government to show the benefits of promoting this app economy.  We've quickly found not only were there no official statistics we could use, but even the companies involved wouldn't give us even under NDA.  
So it turns out one of the ways to do it is count wanted ads that looked like anything that looked like an app developer and then make some an sums between vacancies and jobs.  And then you round things up and you think about how many people are professionally developing apps for companies like BMW and the others that have apps.  But you use a lot of averages and a lot of assumptions.  This report I left before it was released so I don't know what happened with it and joined the Internet Society.  But what this really obviously missed was the winner take all in these types of services.  And the winner take all idea is that at a university, the second best professor will always fill the classroom because the best professor's classroom fills and then people go to the second best.  But if they start teaching online where there's no capacity constraints, the first best professor could have all the students, right?  Because there's no constraint on the numbers.
And so this came out in a report that made me think about a week ago by Vision Mobile saying there's 2.9 million mobile developers making 2 million apps.  47 percent of them make between zero and $100 a month.  Actually about 25 percent make nothing and the others make less than $100 a month.  And 1.6 percent of them make more than $500,000 a month.  So the 1 percent thing extreme with the top 1.6 percent making way more than everyone else combined.
So as another way of illustrating that, 26 billion dollars, some of it is for ads some of it is for in‑app purchases for free apps.  But about 27‑1/2 billion are people paying for the Apps.  Of that, there's one company Gung Ho that makes this game called puzzles and dragons.  It's probably on my phone that my kids put there but I never heard of it.  1.45 billion dollars last year they made on apps with 333 employees, not all of them working on that app.  They have a few others.  So basically that's 1.45 out of 21‑1/2 just on one app.  And those lucky employees were working for a company that brought in over 4 million per employee.
Compare that to another company, in this case I chose Dannon Yogurt, a French yogurt company, making also 126 billion in revenue.  104,000 employees.  The CEO is making 3.7 million and none of them because it's in France are making less than $100 a month.  So the point of this is really to summarize quickly the data availability is terrible in these new areas.  It's also hard to interpret.  You can't really use averages.  You can't make assumptions that everyone who is doing this is doing fine.
And I think you really have to think about that.  Is the app economy a sustainable business or is it a lottery where if you invest 47 percent don't get a payback and the 1 percent make all the winnings?  So I think that's just something that we have to start giving better data for so people can interpret it better.  12 seconds.
>> MODERATOR: There you go.  So some interesting points here.  I think the one is that it's so large ‑‑ oops.  One is that it's so large.  It's growing so quick.  We don't have really good metrics to measure it.  I don't think you really ‑‑ the model seems also complicated, right?  And that it is actually problematic this with the winner takes all.  And that fluctuates so much.  I think we've seen this in the networked economy that WhatsApp got, what was it, 19 billion dollars for a company with 300 employees that kind of disrupted the whole segment of the telecom industry.
Is there kind of like any ‑‑ can you foresee any measurements of metrics or any models that you would kind of recommend people here to maybe have an extra look at that might kind of give us, enable us to come to grips with this stuff?
>> MICHAEL KENDE:  I think that's a good question.  I don't have the answers to that.  But it seems something that shows the SKUs so you're not just looking at averages.  But some way of very briefly summarizing the winner take all aspect.  Like that number of 1.6 making more than everyone else combined.  Some kind of a metric that really conveys the winner take all aspect of it that we really haven't used in manufacturing and other areas where there's some equality ‑‑ inequality within the company but you know that the bottom is at a nice enough level, or at least it used to be.  Here you don't know where the bottom is and if anyone is making anything.
>> MODERATOR: And, finally, when it comes to the metrics side, do you see any kind of point of attack or whatever, where one would be able to find good data round this in the future?  That would be the low hanging fruit or whatever?
>> MICHAEL KENDE:  I mean, as the partnership for ICT statistics that's trying to keep up and trying to promote to governments what they should be measuring, but it's very hard to keep up and then come to agreement and get everyone to start gathering that.  I don't think you're going to suddenly have an app developer column the governments love.  But at least something that's a little more granular than I think what exists now.
>> MODERATOR: Right.  So it's a good point there that one has to agree on something.  But I'm just thinking hands-on.  You looked here at job ads.  It seems a bit shaky to me.  Maybe it's a terribly good indicator, I don't know.  Is there anything more than job ads that you think will be a fruitful approach?
>> MICHAEL KENDE:  I think, yeah.  Nobody knows -- that was the beauty of it.  Nobody knew if the job ad approach was right because nobody knew the right answer.  That's the consulting approach.  But I think that the survey really is what demonstrated it, right?  So if you take some surveys and if you're pretty confident, that's what demonstrated in this market.  And I think that may be an interim step that people take to start ‑‑ even the government starts randomly and anonymously serving to build up their statistics in the interim before they can really impose any more formal metric.
>> MODERATOR: Mikhail talked about an economy where one has data transaction, money transaction.  Maybe this can be like taxes that government tells these companies you have to tell us this data, or something, I don't know.  Who knows?
All right.  So, Rudolph?
>> RUDOLPH VAN DER BERG:  Yes, Rudolph van der Berg.  I work for the OECD in Paris which, according to our critics, is a neoconservative, socialist, crypto-imperialist, a more bilderberg type Communist international organisation.  At least if I read some of the comments.  So we are somewhere in between.  And what we try to do is with data, do work on better policies for better lives.
One of the directors in the organisation who does the -- project.  Without data you're just another person with an opinion.  And that also characterizes our work with regards to Internet.  We try to get the data, but it's not always easy.  So quite often we're still left with opinions.
That the Internet is important and changing businesses was quite clear this week here when together with [Inaudible] sitting over there, we were in the city centre and we talked to these carpet salesmen.  And one of these carpet salesmen said "oh, you work on the Internet, oh, that's great, do you know anybody who can stop eBay or just get it off the Internet because it's hurting our business?"  For some reason honest carpet salesmen just cannot have a reasonable business anymore in this new economy because, you know, their honest way of selling carpets are disintermediated by the Internet.  
And that does give a good view that the Internet is pervasive even in those kinds of areas.  But measuring that makes it a lot harder because finding where the Internet really hits is often a result of where the information is.  And information is qualitative.  And it doesn't really show well and national account.  
Now, in the past, we've had similar situations.  In 1978 the OECD started a great project which I think is seminal but underappreciated under around the world.  It looked at the value of telephone lines before 1970-something.  It was thought that the telephone was a luxury.  And there was even a curve which said that if countries were above the curve, they had overinvested in telephone lines.  It measured teledensity as GDP per capita.  And if you were above the line you invested too much, if you were below the line you invested too little.  This was done by Siemens engineer who also worked on the enigma machine pages ago.  Anyway, Siemens mainly used it to prove to countries that they needed to invest.
But then this project showed that actually it wasn't about whether you had overinvested in telecommunications; it was whether you could afford to invest less in telecommunications because even for the United States, which had the highest teledensity in the world at that time, every extra phone line per 100 in the country added so much to GDP growth that it was worth it.  It was really worth it connecting everybody.
These days we still have similar problems.  we have countries ‑‑ recently we did a study on international call termination and that clearly showed that countries that raised the prices for incoming international calls into their country, which is a common thing in southeast Asia, Africa and the Caribbean, these countries were losing out.  They actually collected less per line from the United States than countries who didn't.  So charging for incoming traffic is not a good idea.
We then tried to measure the full impact of the Internet economy on GDP growth.  Then it becomes harder.  It's like paradox in the 80s.  We see computers everywhere except in productivity statistics.  Well, some people worked hard on solving that problem but now we have the Internet paradox.  We see the Internet everywhere just not in GDP statistics.  I've got 30 seconds.
That being said, we can get some nice statistics from the networks.  So, for example, a colleague was able to show that the number of autonomous systems in a country is an indicator of competition and you can see about 0.01 percent extra for each extra AS per 100,000 in the country.  Of course there are problems.  It may cost jobs but I'm out of time so I can't go into that.
>> MODERATOR: All right.  So to follow up with a question.  May it cost jobs?
>> RUDOLPH VAN DER BERG: That's a good point.  Inequality is a very negative thing for an economy.  I know Secretary‑General has been discussing that at several events lately.  You need some inequality but not too much inequality.  And we are now in a period where it may lead to strong structural changes.  We, for example, made a gamble that about 1.3 million people according to official statistics work in warehousing.  If those warehouses become roboticised, which is quite likely given the developments there, then Europe may lose a million people working in that industry.
 Now, they can find other jobs, but that may take a lot of time.  And when we look at the miners, most of them didn't find new jobs.  It was their children who found new jobs.  But that's a 20, 30‑year lag.  That's a lot of social unrest that you have to deal with as a country.  It's not like you can stop but it but you have to look at it.
>> MODERATOR: Last question real quick.  What kind of numbers would you like to have at the OECD and what would you do with them?
>> RUDOLPH VAN DER BERG: Well, we'd like any numbers that show the link between policies and outcomes.  So we recently collected data on where websites were hosted and were able to link that to the quality in the country and the ease of doing business.  That's the kind of data that allows you to link policies and better policies with better outcomes.  That's the kind of thing that we're looking for.
>> MODERATOR: Okay, thank you.
All right.  Let's see.  
>> HELANI GALPAYA:  Taking off from here, I'm from LIRNEAsia.  We're a think tank.  We do research in South Asia, Southeast Asia and the Pacific Islands and we worry about people at the base of the pyramid.  And for social scientists and development professionals who come from other fields, big data, specifically transaction‑generated data is fundamentally changing the way we work.  
Our questions have always been what people do, why they do it and what are the impacts of development interventions?  At least two of those questions, the way we answer them is changing because of big data.  We are working with call detail records, the tracers that are left in the telecom operator bay stations when you use mobile phones and most of the people do.  We working with perhaps the only digital trace that millions of poor people in Asia are leaving and it is showing interesting things.  
We can look at what happens when you build a new highway or a road.  What happens to people's mobility?  We can match that with satellite images of nighttime lights that Google provides which have been shown to be proxies of economic activities in certain areas.  We can see how that changes as a result of a particular development intervention like a road.
We have looked at who people communicate to.  And traditionally the way of gathering, let's say, where people travel, who do you talk, having to go and do large representative sample surveys, which are enormously expensive.  And if you get the sampling wrong, the whole conclusions are messed up.
Increasingly what we are doing is less surveys.  We still need surveys in the short‑term, but perhaps less frequently.  And we are looking at the digital trace that we're looking at the call detail records which show where people go.  And we've looked at who people talk to.  Do these communities match with the maps in countries that politicians have drawn out?  And in Sri Lanka with the CDR analysis, it shows actually not really.  Communities exist where we didn't think communities exist.  
This analysis, I think, will change fundamentally the way election campaigns are run, will change the way advertising is done by private sector companies because people will talking to people that we didn't know they were talking to before.  This is what this analysis shows.
What it also allows us to do is to test hypothesis.  Because of the ability now to work with large sets of data and to match up different sets of data, there is work in Pakistan, not by us but by other people, that is even commercially valuable to identify when people might leave a mobile network operator.  And can you play around with different incentives by giving them different discounts and promotions and run randomized control trials?  
But in our world, more importantly, it's helping bring a large group of people who were previously not in the financial economy.  Because it turns out you can look at people's call re‑load patterns and all the data that generates because most of the developing world is on prepaid mobile, and reasonably predict their creditworthiness.  This suddenly opens up a credit scoring system for poor people.  So the opportunities opens up is huge.  
This brings up issues of property rights, who owns the data?  Who can sell the data that is cogenerated at best by users and network operators.  It is fundamentally changing the factors of production in addition to land, labour and capital, days the data data and how do we quantify that and a value that are going to be challenges.
And at a macro level the way we measure economic activity is changing not just at the microlevel.  So activities like the billion prices project which is trying to find another proxy for retail price inflation calculation, which the governments do with a small price basket, but now you can have a whole lot of other data, a lot of people reporting this and more or less accurate predictions of RPI.  So a whole lot of other things as a result of data.
>> MODERATOR: Great.  So I hear you saying, actually, that the cell phone service providers probably sit on a lot of data.  And are they giving that to you?  Or do you think they will do business with that?
>> HELANI GALPAYA:  Sure.  Lots of them are doing business with that.  We are a not-for-profit organisation.  What we have is anonymized historical data from multiple operators.  Looking at our data, you can't tell who gave it to us.  You can tell it's just about people from Sri Lanka.  It's a huge barrier.  We sign nondisclosure agreements.  We have various barriers to cross.  It is not the same process.  But we've done it.
>> MODERATOR:  And do you notice any difference in the response to when to use this data between the business and government?  Do you think that that gap is going to grow or is it going to shrink?
>> HELANI GALPAYA:  Sure.  I think the operators are wanting to use this data for their own purposes.  I think the business class is clear.  It's increasingly clear for those for whom it's not clear.
I think for governments, the case is being made.  And we are making the case.  And we are basically saying:  Let's use the CDRs, the telecom sector data, to solve other problems like transport, how we design government services.  So it's nonthreatening to the telecom companies because it's a completely different sector.
The biggest challenge we face there is you need to match up the telecom centre data with other data set, i.e., the stuff that is in government's walls in files.  And that's a big challenge than sometimes get data out of the private sector.
>> MODERATOR: Thank you.  Let's go on to Roslyn here.  And Helani spoke about that the Telcos have a lot of information about people now.  And I guess also the guys ‑‑ well, we're doing so much with our smart phones that the smart phones know so much about us that actually these companies are probably going to know almost everything about everybody in the world.  And as Helani, one can use it for a lot of interesting data analysis, for marking, government use, and so on.
What is your perspective on the politics around this now that will come out of that?  Because obviously it's a huge resource and people are going to compete for control over it, right?
>> ROSLYN LAYTON:  Thank you for the question, David.  You know, I feel for the fat to be one of the later speakers because we've talked a lot about these possibilities with network economics and data.  In my work, I like to ask:  Well, what will be the role of regulation or lack thereof to support the development of the digital economy or perhaps to protect users, if you will?
And I think what this whole week has been here at the Internet Governance Forum, what I would hope to do is maybe give you a review of the three approaches that are taken primarily in the U.S. and EU about how we'd go about this.  There's definitely questions around how would we manage some of these issues about big data?  
But something to keep in mind: In the last 25 years, we have gone from 14 telecom regulators in the world to over 155 today.  So the area of telecom regulation has exploded largely because the state‑owned telephone monopoly has been deregulated.  So of course to go through that process, each country set up its own authority to do that.  
But now we're in a situation that in many cases the, how would you say it, old‑fashioned telephone services have been liberalized, for example,.  So now what is the role of a telecom regulator?  And we're in the midst of a convergence, connectivity and computing all coming together.  And we're moving from the single service being delivered across a single network to multiple services moving across a network.
So we still live in a world where all of our ‑‑ the laws that were defined were really conceptualized for an old paradigm that we don't have.
So, you know, there are challenges here for regulators when they look at old fashioned telecommunications which had to do with licenses and charters.  And all these new network economy companies that don't necessarily ‑‑ they don't think about that when they get started.  It's just an app we heard about.  They don't think about do I need a license from the government to do this.  
Different things where we talk about the geography and the jurisdictions.  So if you want to operate within a certain country, if you're a Telco operator.  But if you're over the top or service, you look at the whole world is where you'd operate.
Contracts versus terms of service.  The regulator is looking to enforce the contracts versus a lot of network service providers are offering terms of service that are not ‑‑ don't have the same contractual obligations.  So there's so many areas like this.  Some have requirements around taxation, others don't.  And then all these issues about data protection, data portability and you have kind of a marketplace where all the players have different conditions.  So this is quite hard job for a telecom operator to sort out.
So the kind of the three approaches that have emerged, you know, one is we have a classic telecom regulatory model, which is very much like, we will regulate the network pan how data moves across it.  Whether it's common carriage or netter, net neutrality, these things we talked about this week.  
Another idea that's emerging from the platform regulation where we need to understand if the Internet is a value chain, it doesn't make sense just to look at the network itself; we have to look at the services on top of it, the platform providers, the sort of Google, Facebooks of the world, the devices, for example, the interconnection between the networks.  
And this is something that has come up just this past year with the French government which has sponsored a think tank to look at what could platform regulation look like?  So they have defined a three-pronged test to determine whether a platform is abusing its market power.  So, for example, acquisition. Is Google in such a powerful position that any new company that would start up is in a position to challenge it so it will acquire that company?
It would also look at issues like diversification.  When you are a large provider, you can offer many kinds of products at a very low marginal cost.  So you're a Google search engine, you can easily offer an email product, a mapping product and so on.  And also another challenge is exclusion.  Do you exclude some in the marketplace because your platform is so dominant.  
The final area is called the developmental approach.  It comes out of a Chalmers Johnson.  He wrote about the Japanese economic miracle.  And it sort of said that instead of creating new regulation, we need to transition to a new kind of system for thinking about all these things.  And you could look at the country of Estonia which had no hit telecom communications to begin with when it was liberated from Soviet Union.  So they started at a position of zero.  All IP.  So they never conceived the world about telephone, computers, it's all IP.  So they have one law that applies to everything.  
Interestingly in Denmark, the telecom regulator was disbanded because they said if we look at Internet and broadband, it's in everything.  We don't need to regulate it anymore.  We will put all of our former telecom people across the various agencies.  And then they'll work together with the other departments.  
(Comments off mic.)
>> MODERATOR: We'll come back to that in the discussion because regulation, of course, comes in very much here.
Mike, Mike Nelson.
>> MIKE NELSON:  Thank you, David and thanks for organising this panel and involving such a diverse group.  I just have five minutes so I'll do a quick bumper sticker view of my perspectives.  I'm a Professor of Internet Studies in a unique programme in Georgetown, communication, culture and ecology.  We have economists, but we also have anthropologists, sociologists, policy experts, we even had a few art history majors.  And the idea is to look at how society is shaping technology and how technology is shaping society.
Unfortunately in a lot of the policy world, economics has been the way we analyze a lot of these problems.  Our programme tries to take a truly multidisciplinary approach and tries to do more than just look at where the money is flowing.  So I think this panel has illustrated just how much broader economics has to be in this New World.  As a professor, I try to give a framework that my students can put things into.  I tell a lot of stories.  I also give reading assignments.  And today I'll be giving you a few reading assignments as we go along.
Speaking of stories, I was also at the market yesterday, Rudolf.  And I learned something, which is that you don't wear a tie with cuff links in the market because then you get to pay the American price, which is a special 50 percent surcharge.
I've also got some other stories, though.  And the one that comes most clear because I live in Washington is net neutrality.  Over the last five years, there has been an amazing boom in the business of creating economic analysis of what the right policy for net neutrality should be.  Absolutely extraordinary how many different studies and how many different results have been paid for by the different people on different sides of this issue.
And this is part of our problem.  I mean, economics is almost discrediting itself, because we're seeing the same problem addressed by different economists coming up with very different results.  And the reason you can pick so many different models in this digital economy is because there are so many different things that aren't really right about the classic economic models.
First off, a lot of classic economic models assume you're at equilibrium.  Yet we're talking about disruptive technologies that in many cases are cutting costs by 50 percent a year.
They also often talk about -- they start by considering the individual who's motivated by making money or saving money, when in actuality with social media we're finding that those kind of economics don't explain what's going on.  People are spending a huge amount of their time just because they get psychic rewards or because they're part of a team on a game or because they like to be part of a community.  So we need to bring the psychologists into these discussions, as well, if we're really going to understand some of the behaviour.
In this weird world where there's so many different models, policymakers are kind of at a loss, particularly when it comes to macroeconomics.  And we've mentioned some of the problems there where we don't even measure the right things.  Example that Michael Kennedy gave about the app economy is probably one of the best where we don't even know what's going on.  So how do you make good policy?
And I think the solution is one we've heard already from Svetlana and from Helani, which is that we've got to get the data.  Rather than constructing elaborate models and making decisions based on them, we're getting data in realtime.  So we're starting to see the impact of policies around the world in realtime to help us figure out what the right choices are when it comes to Internet Governance, whether it's allocating spectrum, licensed/unlicensed, whether it's net neutrality, whether it's the domain name market.  Now we have a chance to actually get our hands on more data.
Getting that data is going to be hard.  I think governments have to find a way to push policies that get more openness.
I would say we've done this pretty well in the microeconomics area.  My criticism is month mostly about the macroeconomics.  Some microeconomics have made a lot of money in stock options, particularly Google stock options because they've used microeconomic models to better understand markets and better design online services.
I would tell you first if you don't hear anything else from my talk, if you don't write anything else down, write down an economist walks into a bar.  And the name, Bob Litan.  He gave a Ted Talk last month at Tedxkc Kansas City.  And he analyzes the amazing results that microeconomics have gotten by collecting data on social media sites, on dating sites.  But he also talks about this big gap we have in the macroeconomics.  To learn about that read a book called Economyths, 10 days that economics gets it wrong.
So I'm an optimist and a pessimist and I look forward to the discussion.
>> MODERATOR: Thank you.  Okay.  So it seems that big data is a big thing here, the analytics, it's changing marketing already.  It might be changing how we do governance and budgeting and so on.  And that getting access to this data is one issue.  There are a lot of private owners of it that consider very much on it like the tele-companies and there will be legislation around it.
So let's open up the floor here.  Do you have any comments?  Or do you have any questions about this?  I mean, a lot of qualified people here, I see.  Please.
>> My question is to Michael Kende.  What you said about the winner takes all economy with regard to apps could be said about almost all information markets.  Rock music.  There are lots of musicians, but most of the resources, most of the benefits go to the top stars, film actors, et cetera.  
We could look at sort of the pure application of economies of scale, in information industries, for example, with regard to the cost of replication of newspapers and we could then conclude that there should only be one ‑‑ there are likely to only be one newspaper in a town.  But in actual fact this is not so.
So my first question is:  Are you reading too much into what you're seeing?  And is it also a question of time?  That as the app economy develops, will this change, this particular phenomenon that you're seeing?  And in any case you yourself admit that you're working with poor data so I don't even know whether your percentages can be given credence because, for example, you're studying job ads for app economy developers, but people are ‑‑
>> MODERATOR: Excuse me.  There are many people want to ask.  I understand you're asking relating to the measurements that he needs to make.  How about you start answering and then you can continue with the question?
>> MICHAEL KENDE: I think from this point of view, I think the real challenge is in the policy framework.  It's not random.  And of course people have tastes, so not everyone is going to red one newspaper.  There's going to be some variety with the bands, with the apps.
But, you know, if you're making decisions and you're thinking of training people, the decisions, I think, have gotten different than previous generations where if you train people in manufacturing, the jobs are all going to pay a certain amount and you can set a minimum wage.
If you train the developers, are they all going to try to make the next puzzles and dragons?  And even the companies I don't think have fully ‑‑ they started making plans as if angry birds is going to be a franchise forever, it drops off and the next one comes along.
So I think that it just needs to be factored in.  And I think that the data will help.  And I think people understanding the implications of the data will help.  And then of course we're all human and we're still going to try to make the next big gain.  But we should at least understand the implications.
>> MODERATOR: Do you want to follow up on that answer?  Okay, thank you.
Andrew?
>> ANDREW:  Hi.  A lot of you talked about data.  And of course this is a preoccupation where I work.  And Mike Nelson tells an interesting story of Bob Litan.  I know where he got his material.  There is a National Academy of Science study is that came out that Bob and I co-chaired that looked at the next generation of data from the National Science Foundation, where it's going to come from.  One of the conclusions is you shouldn't look at the NSOs anymore, National Statistical Organizations.  They're losing their budget.  Look at StatsCan, one of the premier statistical agencies in the world, no longer do their ICT survey even though they were pioneer.  They are mandatory service have huge nonresponse items now that lead them to be statistically not very useful.  And their voluntary ones people don't fill out.  And their employees are getting poached by the likes of Google and other places.
So, what role ‑‑ I mean, one, my first question is what's the new role for NSOs, then, in this New World?  And, two, given that they're not likely going to get the data, where are we going to get it from for the analyses that you want to conduct?
>> MODERATOR: Are you directing it to anyone specific?
>> ANDREW:  Many people talked about data, certainly Mike, Rudolf, the other Mike, Helani talked about the billion price product that I think is fascinating and then alternative, kind of a challenge actually when you ask Bernanke what data he's looking at I know he's looking at Bernie Olson's MIT data.
>> MODERATOR: All right.  Who wants to pick up the question?  Mike?
>> MICHAEL KENDE:  This is something I spent time at working on when I was at the White House back in the 1990s.  We were just gelling started in trying to figure out what the Internet economy was.  And we got the census bureau and the Department of Commerce to start spending some time on it.  
But even back then it was clear that the better, faster numbers were ones that you could pull off the net.  If you could get information on how many people were going to what sites, if you could get the companies to provide the data themselves you could save a lot of money and a lot of time and in the end the policy would be better, which would be a benefit to the companies providing the data.  
But it's been hard to make that linkage in a lot of industries.  And I think in the end it's probably going to be necessary for governments to start requiring either as a licensing requirement or just can't do business unless you provide certain types of data.
Tim O'Reilly calls this algorithmic regulation.  It's a terrible term.  But we've got to find a way to describe in very positive terms this idea that rather than telling companies what to do, we are going to ask them for a whole bunch of data.  And when we see that there's something out of whack in that data, government can take action and start pushing for more responsible behaviour by the companies.  
And this could be in the area of financial advising where companies are asked, okay, give us aggregate numbers.  What are you telling your clients to invest in?  And if it turned out that one company was sending all their clients to one or two penny stocks, that would be a red flag.  You'd probably not be able to regulate that kind of behaviour but you could certainly spot that behavior.  
Same thing in environmental pollution.  There's a lot of ways you could see this happening where if companies were in the business of sharing some of the data that they were collecting anyway, we could have a better, fine grained understanding of what's going on in the economy, leading to better policy and, interestingly, leading to less regulation.
>> MODERATOR: Okay.  So I think, by the way, I'll set the maximum time on the answers now to two minutes to get a bit a snappy interaction here.
I was just thinking of Roslyn, do you have any comment on what Mike just said?  I mean that governments are going to kind of demand these companies to give the governments access to all the big data they have?
>> ROSLYN LAYTON:  Thanks.  Well one of the challenges already is that even traditional businesses have already internalized so many digital tools.  So that when we measure today, let's say the Fortune 500, we're looking at their revenues and costs and so on.  But they already did the benefits of a lot of data already.
So, really, to tease this out is extremely difficult.  I've been the past few months been trying to look at American Association of economics and bureau of standards.
One attempt has been made by the United States International Trade Commission, which was looking at digital exports where they specifically were trying to see what was the amount of exports from the United States that were only goods and services.  That was one of the most specific studies I'd seen.
But I would possible only follow up Mike's question which I think it is a very interesting, intelligent approach.  What are the costs of that?  Because many cases if you don't know necessarily what to look at, you may spend a lot of time measuring and you might miss what might be important because how do we know what could be the red flag?  I mean we could define the red flags up front but there could be ‑‑ it's the black swan assisted.  There's also costs to measure that need to be taken into account.
 >> Part of the answer to that is a lot of the data has already been collected.  Just part of the doing business.  In order to know how much to charge your customers, you have to have the cell phone data.  In order to know what to charge your clients, you have to understand what you're selling them.  I agree.  In some cases there may be some data sets that you need that the company isn't collecting and that would be costly.  But I'm optimistic enough to think that we could do a lot just with the data that companies already have.  And they would actually in the end they would be grateful for the opportunity to avoid regulation by doing this.
The other thing to understand is that if these companies aren't willing to share some of this data, it's going to be collected anyway, in some cases by crowd sourcing techniques.  Great example in the area of broadband where there was the broadband survey in the U.S.  They had hundreds of thousands of people around the country reporting what kind of broadband service was available in their particular neighborhood.  And then actually measuring how much broadband speed they got up and down over the course of the day using different apps.
And so if the users are out there collecting the data, it's in the interest of the companies to, I think, provide the full data set so that the real pick customer is developed have.
Again, I have two answers to all problems:  Transparency and broadband.
And this is a great example.
>> MODERATOR: So was somebody ‑‑ you were waiting there.
>> Hi, my name is Salamia Mutin from Lebanon.  Any comments on the World Bank report and claim that every 10 percent in broadband penetration adds 1 percentage point to the GDP?
And the second question is about the publicly trading companies of let's say Twitter and Candy Crush who don't have, let's say, a business model, does not even have ‑‑ they're not selling anything directly, should we worry about this?  Or is it like totally different from ‑‑ is it going to change anything to the stock market to and to the protection of the people who are buying stocks in these companies?
>> MODERATOR: Roslyn?
>> ROSLYN LAYTON:  I'd like to take question one, but maybe I would also say that Rudolf, I'm going to reference some OECD data which is only 30 -- a limited number of developed countries.  And the whole broadband penetration, I think there may be a dozen different broadband penetration studies that have slightly different results.  The one for developed countries shows the mild correlation between broadband penetration and economic growth.  So, for example, what's very strange is South Korea, which has extremely high broadband penetration has ‑‑ they still make the majority of their revenue from traditional industries.  Now they have made their traditional industries more productive, to be sure.  But this is a discussion of the bro broadband in general to GDP.  
Interestingly the two highest countries in the OECD, Norway and Switzerland, the question is the broadband so great because they were rich already or did the investment in broadband make them rich?  Norway gets oil revenue and Switzerland has watches, chocolate and finance.  So that's a challenge.
Now, the other question looking in developing countries where a simple feature phone on a 2‑G network has a huge ‑‑ has like a hockey stick effect in productivity or GDP for a person, the improvement of life if you're a fisherman and now you have access to a mobile network.  
There are two very different -- policymakers alike these things because they want a silver bullet but in practice it's very difficult.
>> Can I just make a quick comment on Angry Birds?  I don't know why buyers of stocks need protection and we need to worry about them.  You should worry about them if you think their information is -- in stock market.  They think they are buying something else.  I think people who are buying into angry birds know they are buying into a game that might be a hit tomorrow and or tomorrow.  If there is information -- we worry about it but otherwise I don't think they need protection.
>> If I can just add to the broadband study question, in some of these, I look at the analysis.  And they're not looking at some of the ripple effects.
I think this is actually the second most important panel here at the IGF, because economics does illuminate so many different decisions that we're talking about.  And the reason I say second is because Andy Wyckoff and Bill Woodcock and Lorenzo were on what Andy claimed was the most important panel and it was.  It was about jobs and the open Internet.  At the end of that they explored what data and what issues we really needed more funding and resources to explore.  And Joe Aladef said the most important thing which was we've got to understand the economics of the ripple effects of the Internet.  
So it's one thing to say okay, we've got more broadband.  That means more jobs in the IT sector.  But there's a lot of stuff that comes from that.  A lot of improved efficiencies in non‑IT sectors that we're not measuring.  And so I think, you know, that these ‑‑ in the long term, these numbers are probably underrepresenting in the impacts of broadband.
>> It's not because the governments are looking for silver bullet.  It's because like, say, I need an argument because the priorities of the decisionmaker, of the policymakers in developing countries is not ICT.  So we need some arguments to get them to do the right thing.
>> The first thing you need to do is make sure is they're all online using the 5 G phone.  When I was at the Federal Communications Commission in '97, the Chairman made sure every member of the staff had the Internet on their desk and told them to waste at least half an hour a day playing with it.  That's really important.
>> MODERATOR: So we have to end at 12:30, yeah?  All right.  So we maybe can go a little over that.  We can't take that many more questions.  That's actually good idea.  Let's take all the questions at once and you have to make them very snappy, then.  So, please, let's just make the rounds.
>> Okay.  Lawrence -- Italian.  It is the second workshop that was mentioned before we organized this workshop on the first day.  I want to get the feeling from the speakers here.  I think that one of the most important issues needs to be looked at is this the implication that we need a stakeholder model.  
I think in my opinion there are three effects.  The first one is this distributional effect.  I think this is really important because we are facing ‑‑ losing up of the middle class and the implication for increasing inequality, jobs and so on.  
The second aspect I think is what does it mean, new policy for this digital market?  This market to some extent are becoming monopoly although in theory should be temporary monopoly.  But this has a lot of implications.  Think about for instant there was this example before of the company that was asking support, can you help us in stop eBay and so on?  I think that regulators should look at are these companies doing predatory pricing -- so it needs to be looked at carefully.  
The other issue is taxation.  In other words, it's important to have a level playing field among all the companies, the digital market, all the companies should pay the fair share of taxation.
The other issue is the implication of the shared economy.  In other words, there are measurement issues but also probably we should try to bring in terms also of solution the spirit of the shared economy.  For instance I was thinking about ‑‑ think about the issue of contribute to pull together some resources like holiday time devoted to leisure or to nonprofit organisation, nonprofit activity in a more structured way, okay?
>> MODERATOR: Okay, thanks.  Let's do like this.  I'll set one minute each so that everybody gets a chance to say something before we have to close the session, okay?  So let's just continue.
>> Hi.  I'm with eBay.  We don't engage in predatory pricing.  So you don't have to worry about that because we don't price anything.
But, really quickly, thank you for the panel.  It's been extremely helpful.  This is a lot of stuff I work on.  I think there's been a discussion of a lot of material at the microeconomic and a dearth of information at the macroeconomic level.  And I would actually argue there's some good stuff at the macroeconomic level.  I just don't understand it.
[Laughter]
So, for example, the U.S. Census Bureau report, 55 percent of manufacturing commerce is actually e-commerce.  I don't understand that figure or what that means.  And so 1 percent -- 10 percent ICT means 1 percent GDP growth.  I don't understand what that means or what the underlying implications of that are.  So that's what I would argue there needs to be more thinking in, is how we actually take the macroeconomic, high level statistic and drill down and create an actionable insight from it.
On the regulatory point another thing that we think a lot about ‑‑ and I would offer the term dynamic performance standard as a slightly better than regulation and would encourage some more economic work on that actually on how a more performance‑based standard could actually improve the performance of the entities.  So the airline industry and the European Union is the first example to look at.
>> MODERATOR: Thanks.  Okay.  Thank you.  It seems like my phone got stuck here.  Not good.  Okay.  So please, who wants to?  There?  Okay.
>> Louise Bennett, BCS, UK.  I've looked at the national statistics in the UK and how they're failing to capture the digital economy.  And it seems to me that national statistics organizations find it much easier to capture manufacturing data than service data.  And essentially the Internet economy is service data and it's cross‑border service data which makes it even harder.
I'd like to ask of if you think that's true and how you can capture at that that?  But I'd also like to ask if 3‑D printing takes off, is that going to make it even harder for people to capture even the manufacturing data?
    >> Yep.
>> I am from Turkish regulator.  I have a question to Miss Roslyn.  I think she explained that there is a shift from network regulation to platform regulation.  So what will be the core issues of the new regulatory framework for the new digital network economy?  And what will be the main concept of the platform regulation?  Thank you.
>> MODERATOR: Thank you.
>> Hello, I'm a researcher from Berlin in Germany.  And I'd like to pose a question to everybody who has called for more data.  And my question is:  Where do you see the role of qualitative data in these macroanalyses?  By which I mean data that allows to explore the how, the why and the meaning of action maybe in the economic realm specifically?
>> MODERATOR: Okay.  Not only numbers but also words.  Okay.  Well, when do we have to close the session?  Does somebody have a time?
>> You should probably finish up in five minutes.
>> MODERATOR: Let's take 1 minute each from the guys on the panel.  Please.
>> So just to follow up a little on a few of the very targeted points.  I think that all the lessons of the first crash have not ‑‑ were not fully learned.  I think people are learning better.  But on the apps, on a lot of these markets, people's business models are changing.  They're finding out what they can charge for the newspaper.  People will pay for quality.  They will pay for in‑app.  
People will learn.  I wouldn't worry about the stock market.  I think people will learn that it's a winner take all market but ways to make money along the tails, as well.  I think it needs to be better understood by everyone when they're getting governments and companies when they're getting into it, what they're getting into.
>> MODERATOR: Okay.  Rudolf?
>> Well, with regards to broadband and GDP, measuring its impact takes a long time.  But government policy has to happen now.  History shows us if you're not connected, you will lose out in the long run.  So more broadband is more better according to everything we know except we don't have the statistics today.  But we will have the statistics in 15 years from now.  
So we could make an experiment and, for example, have Lebanon not invest in any broadband whatsoever and Turkey would invest and then see the comparative differences.  But however 15 years from now you will have a problem because you have not invested and you still need to invest.  And that's the trouble that a policy maker has.  Best practice is invest.  Best practice is get the fastest you can get.  Best practice is you will probably need more of it by the time you're done with it.
>> HELANI:  What happens to qualitative research?  I think it grows in importance in this way.  The big data tells you correlations.  It tells you what happens.  Something increases and the other thing decreases, right?  Some type of sweets sell more when storms come in Florida.  Walmart has seen this so they stack the shelves.  We all know this.  
But to design products, to really design development interventions, I think you need to understand the hows and the whys, how people use products.  So in combination with predicted modeling, big data analysis, occasional surveys, the importance of qualitative research where we observe the said and the unsaid is increasing.  And we see this day‑to‑day in our lives where we are spending more time research dollars on qualitative research.
>> ROSLYN:  So to answer this question, just to the clarify the platform approach, this is not yet the law.  It is an exploration by French government.  If you can give me your card at the end, I'm happy to send you the link that you can read for yourself some of the background on that.  CNN.  I only know it in French.  Concer Numerique -- yeah, thank you.  
One other thing I would look at this approach and I know the case of Denmark because I lived there.  It's a society which has a very high spending rate by the government, also very high taxes.  And they have opted to have the broadband investments to be fully taken over by the private sector because they spend a lot for health and education and so on.  
So I think I agree with Rudolf for the need to invest but which party and under what circumstances.  So for example if you have limited public funds, is it better that you would focus on basic digital literacy skills and ensure you have a lower level, maybe you have 5 megabit, 10 megabit broadband instead of the fastest necessary works.  So that I think is an important discussion in this society there may not be a hard and fast rule.  
Just finally the other challenge is that it's very difficult in areas that are remote that don't have other areas of economic development or education to simply say well we need a fiber network.  And to think that that itself lifts that particular island or community is lifted into development and that is very difficult to show.  So we know when we talk about winner take all that not all regions of any country even have the same endowments that they can operate that way.  So it's definitely difficult question.
>> Just real quick.  I wanted to pick up on the last question which I learned in Washington is always the hardest question.  But actually tie it to the question asked by our friend from eBay about macroeconomic studies.
Our programme at Georgetown as I mentioned includes anthropologists, people who are communications specialists and are really good at distilling from text what people are thinking, why they're acting certain ways, what they're really trying to communicate.  And I think that's an exciting field.  We now have the ability through machine learning and through some of the textual analysis techniques to take a Twitter stream from a meeting like a meeting IGF.  What are the people worried about.  Are they supportive or negative?  This will change not only economics but it will also change politics, I think, because we don't have to do so many phone surveys.  And I just think there's a whole new opportunity there.
But it ties a little bit to this point about the macroeconomic studies because a lot of the times, these studies are summarized in a tweet.  It's the 17 pages, 70 pages of analysis becomes one headline.  And that's what's really dangerous in the policy world, where people are misreading, often intentionally, some very thorough analysis and they're throwing away the uncertainty.  They're throwing away the limits on the research.  They're just making some brash generalized assumption.  And that's my biggest concern about the misuse of economic models.  Thanks again, Dave, for putting this together.
>> MODERATOR: Thank you.  Just to summarize final thoughts here.  It seems that companies are going to sit very much on the numbers.  Like, I mean eBay, you have a lot of numbers, right, that are important for people like Andy from the OECD.  And it seems that there will be a need for more public‑private partnerships around this because also many companies maybe offer services because there is lacking public offerings of those things.  And maybe that gives them this dis-incentivizes them to tell the governments how to do something properly because then they lose their own business.  
We'll of course need much more collaboration between -- collaborative atmosphere between industry and government, which is going to be something new.  I mean, industry don't always say good things about government.  And the other way around.
I think also that the power struggle around these resources are going to be something worth thinking about.  And also that we nowadays can analyze words and not only numbers.  Traditionally we measure numbers and that's something that happens.  And then that ‑‑ then people react with words because that's how we deal with each other.  We say if we're happy or sad and so on and we relate that to numbers.  Maybe we can measure directly if people are happy and sad by looking at their dialogue because all the words are out there.  
So that's -- something else that came across my mind was with Helani here, that you used social network analysis more.  And graph analysis on the statistics to kind of map communities, then, through the film code.  And I think that this type of mathematics might also be on the rise.
Well that's quick impressions.  Well, thanks so much for coming here.  Thank you to the panel.  And I think this is going to be a terribly exciting future to be an economist and to develop new schools of economics.
So thank you.
[Applause.]
(End of session)

***
This is the output of the real‑time captioning taken during the IGF 2014 Istanbul, Turkey, meetings.  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 session, but should not be treated as an authoritative record.   
***