 I think I'm quite interested in what Tony Sharrock has presented. I'm not too sure I caught the methodology quite closely, but it's something I would be very interested to read, and I don't know when the book is coming out or whether there are some papers based on that that are already available somewhere that one can download. The specific question I have is, why do you focus on the billionaires? Because the Forbes dataset, I've used it to basically to try and look at how economic growth in Africa is generating more millionaires, but I looked at the millionaire numbers rather than the billionaire numbers. What happens if you drop the barrel and looking at billionaires, but you have a continuum from, say, a couple of millions to billions, and you try and plot that so that you don't lump Africa as one country, but you distinguish them to see what is happening. Thank you. Could you identify yourself? Who are you? My name is Nicholas Gepa. I'm with Oxfam, and I'm also with the University of Johannesburg. Thank you. Thank you very much, Nicholas. Yes, please. Okay. I have a question for Andrea. You showed that education is the main driving force for Latin American inequality change. When I'm looking at surveys, because in household surveys, normally we don't look so much at technical education, just regular education, but looking in Brazil, I was surprised to see a big jump that no one is talking very much about on the trends of vocational education. And the second thing is when you mention lack of land reform, etc., when you see a Piketty's book, you see most of wealth is housing, urban housing. In the past, it was land, inequality hold, wealth holdings. And as I showed, there seems to be a change in Brazil. I'm not very secure about that. We're just starting to look at this data. But a change in the wealth distribution in terms of housing. So I would like to see your opinion. I would also like to ask Professor Schorach about these household surveys. They are not good to get top incomes, but they are reasonably good to get housing information. So you can estimate housing wealth from these surveys. Of course, not for the top ones, but to get your impression on how useful standard surveys are to capture housing wealth. Thank you. Thank you very much, Professor. So the next question we have, we have two there, please. The microphone is heading towards you. The person who grabs it first gets the next question. Please identify yourself, sir. Yes, John Linden Lane from Rutgers University. My first question is regarding to the first paper, which I'm very interested. During this period of time during Latin America, there was the pursuit of low inflation policies. You didn't touch on it as much, but how much did that play in the drop in the inequality that you observe? My comment is that obviously high inflation affects people differently, and with low inflation, we also get higher quality of measurement. There's sort of a double, two questions here. How much improvement of measurement and the quality of the data that statistical agencies are generating nowadays is causing or our statistical findings and quality is changing. And how much do you think the low inflation policies, that some of the... Low inflation. Inflation targeting like Chile and Peru in the 2000s started inflation targeting. And is there any way that you could, from your data that you've got, could tell us what contribution that has had on your analysis? Okay, thank you, John. We'll take one more in the current round and we'll ask for some responses, so we don't forget. So please identify yourself, sir. My name is José María Larrua and came from Universidad de San Pablo in Spain. My question is for Andrea Cornia. It's just if you could expand a little bit more the cases of Honduras and Costa Rica, because I think that they're both a contra egalitarian trend recently. So what happened in... More widely in Central America, this is not following the example of the connoisseur. Thank you. Okay, thank you very much. We'll come back for another round in a minute or two. So, gentlemen, questions on Africa's billionaires versus millionaires, the issue of land reform and housing wealth in Latin America, the validity or worth of household surveys or not, the consequences of low inflation for Latin American inequality and Honduras and Costa Rica, are they outliers and what do we know? So, Andrea, please. Now, the data which we have used to measure the educational improvements are general enrollment and secondary education. And now I think that you raise an important picture because, including in Europe, there is a big debate whether there should be a general secondary education or vocational education. In Italy, for instance, there is a drive towards the reintroduction of vocational education in the secondary, which is what Germany does on a bigger scale than anybody else. But I don't know. I mean, I haven't really tried to do that. My guess would be that with the very rapid technological changes which are still ongoing and with more limited progress in secondary education, positive but gradual progress in secondary education, going back to some, at least for some of the students' population, to vocational education would have a positive effect. And I come from Bologna City where the most famous school was not the university which is the oldest in Europe. It was a mechanical school which was not able to graduate the students because they would go away, they would be recruited instantaneously. And then the expansion of the educational system has gone in the past towards general education. So everybody has to go to the lycee. Now, the land versus urban housing. I think that is a good point. I'm familiar with the wealth surveys done by Jim Davis, Tony, Lenny and all that. And in fact, I was struck myself by seeing that 70, something like 67% of the wealth is represented by housing. So the argument would be, well, land matters less because in the total amount of wealth, so letting me in my houses doesn't matter if we give them land. That problem for the urban population would be okay. But for the rural population, I don't know. So if I go to Paraguay, Paraguay there is a major issue of land with actually with all these bad Brazilian soldiers that they go in and then occupy Manu Militari, the land to produce social for experts. And I think that Lugo tried to redistribute the land but did not have enough political muscle to do it. And I guess that President Lula has always become, at least in Europe, an icon. It's like Virgin Mary, Lula, and then now also Pope Francis. So he, in a way, apparently probably wanted me, I don't know. I don't know what went through his brain. But I remember well that when he was campaigning he had promised to give land titles to... Now, of course, many people they also say, well, okay, and now Brazil is urbanizing. So it's becoming less of an issue. But I mean, if you go to Guatemala, Honduras, many of these countries, a nice piece of land reform would be quite useful. Now, the issue of inflation I think is a very good point. And inflation, we, basically inflation has been stabilized in the 90s, in the 80s and 90s. So, and of course, if you want to stabilize inflation, normally if you follow the standard IMF package, basically repress the economy, you must adopt the contractionary measures. And that, of course, are in a way the worst inequality. So one could say, well, you know, the Washington consensus did the dirty job of reducing inflation and balancing the books, then let the others now do the good things. Now, during the 2000, inflation has remained pretty low now with the exception of Venezuela. And since 2010, Argentina, when the country, basically the National Statistical Institute has been taken by the government and producing fake data. So the rate of inflation in Argentina is about 30%. And the government continued to say it was 10%. Now, if you want to reduce inflation, if you don't use an administrative method which may be useful or sometimes are not, basically you have to use contractionary monetary policies or contractionary fiscal policies, which will worsen inflation, which will worsen inequality. So now we try the econometrically, it doesn't come out anything significant. Now, the improvement, they did it. Now, let in America, country to Africa, for instance, because now I'm studying these problems in Africa, it has a very good data set. So they said, like, the work done by Leonardo Gasparini and Guigermo Cruces in Universidad de la Plata in Argentina is really to be recommended. They took, they went back, Cepale has done a lot of good work as well, but I think that what the La Plata people did, they took the service and they applied the same procedures for inflation. So, you know, and I think that this difference in procedures can explain up to two, three, four points in genius. So if they use the same methodology, then we have reasonable data from 1990 onwards. So I don't think that there is any problem with that concerning the service, except for concerning financial wealth. Now, with an open capital account, if you are a rich Brazilian, you can export all the money you want. Probably you go to the Bahama, no, to the Virgin Islands and that is the real issue now. I think that the issue of international fixation. Really, Central America has been doing on average less well. Now, and Costa Rica, which was like the icon of Central America, was like the good country. People, even this morning, Amartya Sen mentioned Costa Rica, but Costa Rica is one of the few countries in Latin America which had an increase in income inequality. It also had a shift in political regime towards the center-right, the more conservative regime. And I think that that is not the problem, but the problem is that the economy is moving towards manufacturing and so perhaps the Christmas model here applies. Now, Honduras had bad distribution, worsening distribution until 2005 and then an improvement from there onward. And El Salvador, which fought the civil war, which ended in 2001, I think, in 2002. And basically had a very, very large improvement in equality with the upper party, which is a very conservative party. I think the prime minister was one of the guy leading the death squad or something like that. So it was very difficult to explain how inequality fell in El Salvador. And I think that what the authors of the case studies show is that migration, I think migration has been... And normally in theory, the hump theory of migration says migration is disequalizing, but now so many El Salvadorians have gone abroad. So now even the poor migrate. So remittances tend to accrue also to poor families and also the amount of people who has left have reduced the number of unskilled labour who are available in the economy. So if you have a migration of the Salvadorian scale, then you can have a very large reduction of income inequality with almost no impact of the CCT type problem, which are being very modest. Okay. Thank you, Andrea. Tony. Thanks. First question about why do we use billionaires rather than millionaires. I wasn't clear enough. I used billionaires because that's the only data we have. If there was estimates of the number of millionaires in the world, that's what I would use. But there are, I think our estimates are, there's something like 30 million dollar millionaires in the world. You're not going to get accurate numbers on that. The billionaires though, someone sits down and puts resources and does some reasonably accurate, I think, estimates. I mean, the Forbes, if you read the Forbes pages and so on, they even have people sending in their certified accounts, I mean, to Forbes. Some people are so vain that they want to make sure they're on the list and stay on the list. And it's, you know, these are, if there's a relatively small number, we're talking about, what, 1,500 or something, billionaires in the world. So the, you know, we're talking about modest numbers and presumably they have quite good estimates of people that are quite close to the billionaire range. We use that, we use the billionaire as our observation, but then we use, we assume it's a Pareto distribution, so all you need with a straight line is two points. And one of them is the billionaires and one of those distributions. We just fit it to the, you know, we're just saying, assume it has a Pareto tail at some point and assume that the billionaire data is correct. Then I can generate estimates for everything. We estimate what's the number with 100 million, what's the number of 10 million, what's the number of millionaires for every country in the world. It's what we do every year and we've been doing it. It's just now we've got a more consistent way, I think, to avoid, you know, year on year variations, which are a bit erratic. I also have to say, you know, we generate those numbers, we also generate now median wealth in each country in the world back to the year 2000 in a sort of comparable way. So obviously for African countries, there's very few that have billionaires, but quite a few, pretty well all of them are going to have some millionaires and we will give those numbers clearly. It's the countries which have more, where they have better data, when you get to a point where you've got no billionaires, you've got no wealth distribution, you've got no wealth data, you know, you're limited in what confidence you can have in the numbers, but we try our best to get the best estimate based on what we have, the information we have available. The second question about housing, I think it's an interesting one. The problem, as it turns out, the evidence suggests that people are actually quite accurate in, when you ask surveys and they ask them their housing wealth, they're quite good at responding. Indeed, the estimates, the evidence suggests that they actually even give inflated values, but the numbers you get out are maybe five or 10% higher than the true value of their housing. There's quite a lot of evidence. There's various people that have done studies to compare that. It's not housing wealth that's under-reported, and the whole real wealth is probably, isn't the problem area. It's financial wealth and it's to do with, and even worse, I think, debts. I mean, people do not tell people how much they owe. So you get, this is where you have the problem areas. The problem, though, is if you're looking at that as a trend, because the housing wealth is only, is housing non-financial wealth globally is around about 45% of the total now. Historically, what's happened is in the year 2000, financial wealth was about 55% of gross wealth, 55% to 45%, the ratio then went down until they were pretty well equal in the year around about the crisis. In fact, I think there was one year in which financial wealth went below non-financial wealth. But since 2008, non-financial wealth has been pretty flat globally, and it's financial wealth that expanded. So the gap has gone up to roughly what it was before. It's interesting, because of the, as I told you, these trends that it looks like, inequality went down up to the crisis, and then it's gone up again. This is exactly what you'd expect on the basis of portfolio differences, because richer people have a higher proportion of financial assets. So if financial assets are declining relative, if asset prices are changing such that financial assets are losing value, then you'd expect that to be equalizing. And if financial assets are going up, which they have been. This last year, we're just about to the report this year, stock markets have been booming. This is the last 12 months or 18 months have been probably the biggest rise in personal wealth in the globe ever. Stock markets have just soared. The US, which rose last year by 22%, is the median. It's actually just slightly below the median for the countries in the world. More than half the countries did better than 20% increase in market capitalization. So we had this huge increase in financial assets, but this has been true since the financial crisis. It's just booming. It's interesting because of, sort of Piketty's intervention has raised questions about what is happening globally. And one suggestion is, and this is really quite irrelevant, I think, to the discussion here. One suggestion is income inequality has gone up. So more income has gone to richer people. Richer people don't spend as much. So what do they do with this extra cash? They look around for investments. So they've been channeling this extra money that's been coming along. They've been channeling into investments. What's that done? It's driven down interest rates. You look at interest rates since 1980, they're just declining. We're now at historically low interest rates, as we know. But of course, if interest rates are low, asset prices go up. So housing prices go up, and also all stock markets. So you could think that what has been happening, and this is the macroeconomic effects of increasing inequality, is that it's been driving down interest rates, driving down asset prices, which is then of course fed back to the people that own the assets. Of course, what's happened at the low end, the people who don't have such high incomes, everyone else has been galloping ahead. Their income has not been going up. So what have they been doing? They've been borrowing more. So this increase in debt has been fueled. The people that have benefited are funding. They're actually lending the money to the people who have not done so well and they're just building up debt. And again, you've only got to look at all these things. This is what's happening, and this is really one of the worries in the world, that whether this sort of trends can be sustained. And of course, all of this also is funded a little bit by quantitative easing by the central banks, because they are buying up all these assets and forcing down these interest rates. So when all this comes to an end, we don't really know, but it's going to be interesting to see whether the whole thing unravels or whether or not, you know, whether there's another option. Thank you, Tony. Since we started late, we've got a little bit of time to take out of the coffee break, but we can have another very quick round. Please be very concise and point it in your questions. I'll take three more. So the gentleman over there, please. We'll get the microphone to the gentleman there if you could identify yourself. I've had no female questioners so far, and I've got two males lining up. So I'm going to take the next one over there. So please identify yourself. Hi, I'm Johar Honkla from Statistics Finland. First of all, I would like to congratulate my former boss from 96th Andrea for a very interesting presentation. Glad to see you're still doing fine. My question is to Tony Sharks. Now, I appreciate a lot what you're doing. It seems an impossible task to construct a global distribution of wealth. Now, you have data limitations, and you were discussing about financial and non-financial wealth. My question is on the estimation of real wealth. To my understanding, there are very little countries which have balance sheet data on non-financial assets. And even worse, I've looked at your report. The countries for which you have balance sheet data on real assets are countries which have relatively low home ownership rates like Germany, Netherlands, France. And you use this data to estimate the real wealth for countries which have much higher home ownership rates. So my question is, how do you account for the differences in home ownership rates? Because for this, you have really good data. Okay, thank you, Johar. If you could take the lady now at the back if you could identify yourself, please. Hi, I'm Tanya Pohan Sangra Zhang from Chulalonggong University in Thailand. I have one response to the second presentation. I think the wealth pyramid is really interesting, mainly because I've been interacting with demographers. And I'm not sure whether we can borrow some of the techniques, for example, right now in that pyramid in each layer, not only the number of the population that's shown, also some other categories like proportion of people who finish secondary education or proportion of people with a certain level of asset holding, such as land holding. So I'm not sure how much contribution that would be, but it may be interesting to show other dimensions of inequalities within the pyramid as well. Okay, thank you. I'm going to take one last question. I'm going to go inie-minie-minie moment. I'm going to take James Foster at the very front. So microphone to James Foster, please. Apologies, sir, but you will be able to ask over a coffee break. James, James, there we are. Now, hello. Yes, one thing that I was interested in was, in the first talk, everything comes down to education. And in the second talk, nothing comes down to education. In other words, education or human capital isn't included in your definition of wealth, but as you know and remark in the report, it's three times the value of the wealth that you're talking about. Has there been any discussion or attempt to incorporate education as part of wealth per se, human capital as part of wealth per se? So that's my main question. I'll just ask two other little questions. Is it really appropriate to use the Lorenz curve to measure wealth inequality or should some absolute notion be used instead? So that's just one of my measurement questions. I've been thinking about that for a long time, that wealth is different than inequality being stock versus flow, lots of other things. Yes, and zeros are there as well, and negatives like crazy. So these are really big problems. And then finally, just in terms of education, you are using schooling clearly for your discussion of education and equality. Is there anything on the front of introducing quality into those numbers? I'm actually going to sneak the gentleman there back in because he nearly got the microphone, so he deserves to ask the question. Please identify yourself. Thank you very much. My name is General Wu from Australia. Just a very brief comment on the billionaire number. A, this number is very small in each country. So if someone becomes bankrupt, then it creates a lot of distortion. B, those guys are also very mobile. They can change their nationality anytime. I was thinking it was better to look at the other side of the distribution. They poised 10% or 5% instead of they reached 1%. The number of the poised 5% in each country. Thank you. Okay, so back to the panel for some very brief responses. First, let's start with Tony first. Tony. Thank you. Okay, let's just... First question about real assets and whether there's a bias there to countries which have low home ownership rates. I'm not quite sure how... I don't think we'd really bias that. And I'm not quite sure... I think we might think about that question rather than something which we do throw in other variables. So there's a sense in which we may pick that up. We are all the time, I think, trying to revise our numbers and see whether we can improve. But we are at least limited by the fact by what data is available. So I'm sure in 20 years' time people will be doing much better than we are. But in the meantime, we'll do what we can and we'll just sort of keep adjusting. But we'll think about that and see whether we can add some of the variables in, I think, to see what difference it makes. The wealth pyramid question about whether we can have other dimensions. In a sense, we do in our report. We have, for example, we give a breakdown by regions and so we could easily break it down by lots of other factors. We don't have any... Remember here, I told you at the beginning, we're generating a synthetic database. So we do not have any other characteristics of individuals. You can go back and you can generate for countries where we do have survey data, we could get that, but otherwise we don't. So we're a bit limited there. Of course, if you want to take a country and go and process their wealth data, then you would be able to do other things as they do. So I'm not sure we don't... If we had other dimensions, what we'd like more than anything, I think, is age because we don't... We just have no information other. We're just generating what we think is a number and we don't even break that down by age, but we would like to do it by age. We'd probably like to do it by gender and then we might get on to things like education level, but that's a long way ahead. About James' question about human capital, this always comes up really. I mean, how you define wealth, if you start to sort of add in human capital, you get something which is almost a sort of a neurotized version of income. And so I think there is an advantage for keeping the thing separately. The usual definition is something that's marketable. The questions about definition are tricky. And often the trickiest question is to do with what you do about pension wealth. And because there's different sorts of pension wealth, there's funded pension, state pensions versus private pensions. So that is a really kind of complex area. If you can, if you like, add in... You can add in these other things, but you end of the day, you say, is that really helping you understand the world? On the whole, I think people do have an interest in the wealth of people, what their assets are, what they're doing with it, how they've made their wealth. And that is what we try to respond to. As James also mentioned, though, that the wealth inequality is interesting because we can have negative numbers. And in fact, we do get quite a few negative numbers. And indeed, negative numbers are growing. More countries are having negative numbers. And part of that is there's lots of debts at the bottom, payday loans, student debts are getting enormous in lots of countries. And one of the problem countries we've had is Denmark, because they have 30% of their population, has negative wealth. And indeed, if you look at the Lorenz curve, it doesn't actually go above the x-axis until you get into the 60th or 70th percentile. And it gives you a genie more than 100. At least we did have. That was partly our processing, but if you look at the report last year, you'll find that Denmark has a genie of 105 or something. And we get people writing to us all the time saying, it can't be right. The genie is between zero and 100. Of course, that is true. We're not using the Lorenz curve anymore. I've just told you. So there's a sense in which moving to this sort of diagram in which we're plotting this curve is much below or along the lines that you're suggesting. If you wanted to keep it in terms of wealth rather than wealth relative to the mean at the bottom, then you'd be able to plot what's happening over time and seeing how the curve is shifting. But if you're interested in analyzing inequality and inequality trends and trying to, let's say, combine the use-the-forbes-type top-tailed information for different years, then I think I would recommend that one goes ahead in the way that I've proposed. Thank you. The last question was about billionaires. Yes, there's a small number. Again, we're not interested just in billionaires. I'm just using it as a point, as a sample point. It's used as saying we're treating it as if it's accurate. It is true that, you know, people can move around. On the whole, we use the forbes definition. It is on what their nationality is, not where they're resident. So if people move to a different country, the forbes data doesn't treat Abramovich as a UK billionaire. The Sunday Times does in crude terms. We're taking the citizenship definition, largely because that seems to be more relevant in terms of if we're using that as a predictor of what the wealth is in that country, the few people at the top who happen to move to another country. They made their money in Russia. That's where the other big wealth holders are like to be. Some of them will have moved. That's true, but I think it's a big problem for Russia. It's a big problem for India. I think there's quite a few transient Indians. And I think probably will be, you know, I can imagine it being a problem for other countries as well in the future. So this issue about, as I mentioned, is one which I think will have to be addressed in the future. Okay. Thank you, Tony. So, Andrea, any brief responses before we break for coffee? Yes. I think that the Latin American story boils down not only to education because there are also policies in the labor market, taxation, public expenditure, macroeconomics and so on and so forth. But basically the educational policies already since the 1990s, basically they affected very strongly the distribution of human capital and therefore the ratio between skilled and skilled wages because they increased the number of people with sufficient education. Now, the quality of education, maybe we don't have data. They're not easy to measure because we don't have a distribution. We have tests, but you don't have then a distribution of people with the quality taken into consideration. But I think that one of the argument is that the quality of education has been falling as we, I think, that has been shown because of massification of education and that may have contributed to the decline of the wages of the skilled laborers. Okay. Very well. Okay. Thank you very much, Andrea. So it remains for me to thank our two excellent speakers who've certainly given us a lot of food for thought to adjourn for coffee. Remember that the next sessions will begin shortly and when you next get a call from Forbes reminding you to fill in that billionaire form, please do it accurately. Okay. Thank you very much.