 I'm going to say a little bit about what we've been doing, what we did at Wider, and what we've been doing in cooperation with the Credit Swiss Research Institute. And then, because the topic here is wealth distribution, what next, I'm going to talk about some needs, the need for more data and the need for better data that we keep being reminded about as we do our work in looking at the world wealth distribution. And then I'm going to say a few words and show a few slides about the situation in China and in India, which are leaders in the field of wealth distribution statistics in the developing world. Okay, so yes, this, a wider project produced a book. There's a copy of it on the literature table. It's pretty fat and has a gold cover with stacks of gold coins, I guess representing wealth. And it's very broad in terms of the topics that are covered. And I think that everybody who's got a serious interest in wealth these days should have a copy of that book and it's still available on Amazon, okay? And then something else that came out was an article in the Economic Journal in March 2011, which kind of summarized what we had done in the wider study just on the estimating the world distribution of wealth. That was just one chapter of the OUP book. Now, so last few years, we've been putting out an annual report and data book from Credit Suisse. The report is essentially targeted at a lay audience and there's a data book which comes out each year, which gives the details about how things are done in addition to lots of numbers. So anyways, we've addressed special topics in our annual reports and here you can see what some of those are, gender dimensions, mobility, inheritance. This year, the report is going to come out in October, probably late October, I think, and the focus is on the global middle class this time. I'm going to show you, I'm not going to talk about the nuts and bolts. We did that here a year ago and I think many people have heard this before, but what I'm going to do is show you some of the interesting results from this work. Here we've got a chart of aggregate household wealth from 2000. This one goes up to 2013. You can see that basically household wealth has been rising fairly well around the world, except for the period of global financial crisis. And not all regions have rebounded with equal success from the financial crisis, but world wealth has been growing. And here's a kind of a familiar looking picture that shows the countries in the world divided up according to their household wealth per adult. And so the usual suspects are in red, high income countries, also attending to be high wealth countries. You'll notice that some are white. Those are countries for which we don't have any relevant data. One thing about our study is we decided at the beginning we weren't going to leave out any countries. So despite the fact they're shown here in white, what we do in our work is we assign them the region and income group average. So all countries are included. And we thought that was very important because the countries that have less data are also the less wealthy countries. So if you leave them out, you get a distorted picture of what the world wealth distribution is. This is a very interesting chart that Tony devised, oh, I guess about 10 years ago, and we've been updating it. Along the bottom, you've got the deciles of the world wealth distribution. So you've got the poorest people on the left, the richest people on the right. And the areas show what fraction of each one of those deciles is in each one of these world regions. So you can see, for example, China is really big in deciles 6, 7, and 8 now. And if we go back 10 years ago, it was really big in deciles 4, 5, and 6. It's moved to the right. And you can just look at the shape of that area for China and sea that as if China's economic progress continues, as we all hope it will, close to the rate that it's been going, that large group of people are just going to continue moving to the right. There's quite a different shape, for example, for India. It's kind of long and tapered, right? So there are some very wealthy people in India. There's a middle class in India. But the concentration of population is at a lower level in world terms than the concentration of population in China. So Latin America has almost equal representation in every decile. So it's a microcosm of the world. Another interesting thing is that North America's got some very poor people, very low wealth, and so does Europe. So everybody has a little bit of a share at each level. Here are the numbers that would give you a Lorenz curve for the world wealth distribution last year. Just pick out a few key ones, see if the laser pointer works. The share of the top 1% we estimated last year, the share of the top 1% was 48%. In October, we'll come out with a number for 2015. Genicoefficient.911 is getting pretty close to the upper limit. And this is the average wealth per adult in the world we estimated. This is all in US dollars, $51,600 last year. OK, last year we looked at the trend, which was new for us. We had always just been trying to show the most recent picture. And Tony in particular did quite a bit of work on this. And what you find is that in each region and for the world as a whole, maybe I'll point to the world as a whole, this diagram, this chart is for the top share of the top 1%. From 2000 to 2007, the share of the top 1% went down. You can see it actually was lower again in year 2009. But then after the financial crisis, it went back up. And one of the main reasons for this is that world financial markets have been on a tear since 2009. Stock markets have been going up and up in general with a few wobbles. So if you look at any particular region, you'll see the same thing. Inequality, according to the top 1%, went down to the crisis. And then it went up. So let me move along. As I was saying, we've been impressed by the need for more and better data in certain areas. This should not be overemphasized. Sometimes people think, oh, wealth data are not very good or they're kind of sparse or so on. Two thirds, the countries that have wealth data represent about two thirds of the world's population. We have 95% of the world's wealth. And so we shouldn't. And the quality and the quantity of wealth data keeps increasing all the time. I mean, for example, the ECB sponsored the development of wealth surveys in 15 countries on a standardized basis. This came out a year or two ago. Great leap forward for Europe. We have wealth surveys in China and India. We've had them for some time. So the picture is not that some desperate problem with wealth data. OK, so anyways, having said that, there are two kinds of data that we use. One are household balance sheets. These are part of the UN system of national accounts. They're supposed to be balance sheets for every sector. And the high income countries have these. In the South, South Africa has got household balance sheets. But in general, they're currently lacking for developing countries. So we need more of those. That gives us the level of wealth. Of course, what we do for the missing countries is we do imputations. They're based on a regression approach. And I won't go into the details on that. So in total, there were 47 countries that had this kind of data, although for 30 of them, they only had it for the financial assets and debts not for the real assets. Microstatistics, at the moment, there are 28 countries that currently have wealth survey data. There will be others very soon. Ireland and Uruguay within the next year will be releasing their numbers. And we heard yesterday at a conference in Berlin that Poland is also working on getting their wealth survey data out. So there are more of these countries all the time. But very few developing countries. Fortunately, the two largest developing countries have this data. There is a data set for Indonesia, but it's from 1998. So it's getting a little bit old at this point. Thailand has a survey in South America, Chile. So there is some regional representation, but there's a need for more of this data to be collected. And there's no reason not to do it. OK, the next question I'm going to ask is about pensions. And this is very relevant for high income countries, but I think is increasingly relevant in developing countries. For example, there's a very complex and interesting pension situation in China. I don't know a lot about it, but I'm reading a little bit about it. And this is an important form of wealth for many middle class, upper middle class families in that country, and probably is as well in quite a few other developing countries. The countries, so at the moment, for some reason, English speaking countries have moved into this in a big way. So Australia, Canada, and the UK, they have the employer-based estimates of employer-based pensions. So the present value of your benefits that you're going to get when you retire, that's included in the wealth surveys. The US only has the defined contribution pensions, but that's about half of what's going on in the US at the moment, defined benefit or left out. And then there are a lot of countries that include what are sometimes called private pension plans, tax-sheltered household savings. The impact of including employer-based pensions differs depending on the country. Australia in the top line here, we can see that the share of the, don't have the numbers for the share of the top 10%, share of the top 20% hardly changes if you put this in. On the other hand, in Canada, share of the top 20% goes down by five percentage points. And the genie coefficients, according to the US, the UK estate tax data, which was very good in 1994, putting in employer-based pensions would reduce the genie coefficient from 0.67 to 0.59. So this is something that really has to be thought about seriously in the design of wealth surveys going forward. Need for better data, I would say that more attention needs to be paid to the upper tail of the wealth distribution. And this is not controversial. People who produce the data are well aware of the problems. There's both sampling error and non-sampling error. The sampling error can be illustrated by figuring out what's the chances of getting some of these genuinely rich people in your sample. So I forgot to put down the sample size for these calculations, but in a sample of 10,000 households, say, your chance of getting something to top 0.01% by random sampling are only 63%. The chances of getting at least one billionaire in India in a sample of size 10,000 is only 0.1%. And in fact, it'd be kind of embarrassing if you did find a billionaire who might throw all the calculations off. But the non-sampling error is a much worse problem. Basically, once people are asked a certain point, they're not going to answer your sample survey, right? There are going to be non-respondents. So it's a big problem of differential non-response. And there's underreporting. There are validation studies that show that this is especially important for financial assets. So something to be aware of now, in our global wealth work, we try to do something about that. In this slide, I've listed some other sources of data that can be brought in. What we bring in are the Forbes and other rich lists. And I've said here, this is exhortation. Don't randomly sample to determine the upper tail of the distribution of stars according to their brightness. Go outside and look. If you're going to try to establish the distribution of stars according to their brightness, of course, you would go and you would just look up in the sky. You can see the brightest stars and make a list. Well, that's what Forbes and these other organizations, Sunday Times, the UK are doing. We all know Bill Gates, Warren Buffett, Carlos Slim, Mattel, and so on. These are household names. So we shouldn't pretend we don't know that these are the richest people in the world. And we have estimates that have been provided by journalists. And these people themselves can objective. They think those numbers are wrong. And by and large, they haven't done that. Now here is kind of a simplified explanation of what we do to correct the upper tail of the distribution. We do this for all the countries that have Forbes billionaires. The blue line, as we've got the log of wealth on the horizontal axis, the log of the number of people who have wealth above the given level on the vertical axis. If you have a Pareto distribution, this would show us a straight line. So the actual distributions from survey evidence are not a straight line of the upper tail they tend to droop. The blue line is representative. I think it was from China. And so what we do is the number of billionaires, billion dollars is this 1E plus 09. That's a billion. So for each country, we can say, well, how many billionaires were there? That gives us one point on this line. The other point is found by taking a straight line, which I asked some totes to the blue line that comes from the survey. So Tony Shrocks has done quite a bit of work and has done refinements. This is the basic approach that we use. Here just for fun is the list of the billionaires. 535 of them this year in the US. You can see that in India, which is number 21, it was four at the beginning of our project, 2001. It's now 92. Some other countries have not had very much increase. Russia has had a big explosion, which happened quite early actually in this period. China is number 39, according to Forbes, had 213 billionaires this year. And back in 2001, it had just one. So things are pretty dynamic. And there are lots of countries that had none in 2001. And now they have some. Here's a comparison of the top tail from the survey data. And then after it's been revised in our estimation. So for some of these countries, there's a big difference. Chile is a good example. According to the survey, the share of the top 10% was about 38%. According to us, it's 69%. So it's just that if you go out to sample 4,000, 5,000 households, collecting them in the usual way, your stratified, clustered, random sample, you're not getting a reflection of the true upper tail. Now countries that have registered data, like Denmark and Norway and Sweden, of course, they get much closer to what we think is a good estimate of the concentration. So there we have 69% of their survey, 68% in our stuff. The country that has really excellent well survey and shows that you can actually, these surveys can be very good. As the US, their survey of consumer finance, which is conducted by the Federal Reserve Board, they've invested a lot in this over sampling of the upper tail, which is a standard approach in surveys to try and deal with these problems. They have what they call a list sample of people. And what they do is they have income tax records on the high income people. And taking that and other stuff into account, they try to guess what their wealth is. And then they pick high wealth, predicted high wealth sample of people and they oversample them above. The highest wealth category, only 15% of those people will respond to the survey. They actually leave out what's referred to as the Forbes 400 from their sampling frame. They just say, well, it's no point, you know? Forbes is telling us what their wealth is and they're probably not gonna respond to our survey. So they're not even in the sampling frame. So they don't deal with those top 400 families. At any rate, you can see 74% in the survey, 74.6% in our estimates. China and India, okay. Well, when we first started thinking about could we estimate the world distribution of wealth? We didn't realize this. We looked into it and discovered that there was very good surveys in both China and India. And then we knew we were in business because there you have about a third of the world's population. So under the umbrella of the Chinese Academy of Social Sciences, a group of researchers in the Institute of Economics, of whom at Li Xi, who's here today, is one of the leaders, have conducted a series of wealth surveys. The ones where the results have already been published are for 1995 and 2002. 1995, the wealth genie was 0.45. 2002 was 0.55. The 2013, I'm sorry, actually, this number should say 2013, I believe, right? Those results have not yet been released, but Li Xi was telling me that according to their preliminary work with the data, they think the genie coefficient is gonna be approximately 0.65. So the wealth inequality, of course, has continued to increase in China. There's another survey conducted at Chengdu University, 2012, and they actually have a 2013 version as well, seems to concentrate on higher income people, and it may actually be that, relatively speaking, the financial assets are overestimated in that survey. I want to know more about it. We would like to know more about this survey before we make use of it in our work with Credit Suisse and don't have enough confidence in that at the moment. India, well, this is the distributional data from 1995 and 2002 for China. One of the big things that happened is that the urban per capita wealth shot up while the rural per capita wealth did not increase very much at all. India has had a wealth survey conducted every 10 years with actually a very large sample, earlier I was talking about what would happen if they had a sample of 10,000 households in India. Actually, the number's more like 130,000, and the response rate, the reported response rate is very high, it's about 95%. And so you've got a relatively small increase in inequality, well, actually looks about the same in those first two years. But the 2012 data, which is just out, the Gini coefficient leaps up to 0.716, and this is without any adjustment for the upper tail. So my conclusions are, first of all, that world wealth inequality is high, it fell before the financial crisis and increased after. We need more data, household balance sheets and wealth surveys need to spread to more countries. There are examples to follow, the South Africa's taking the lead there on the household balance sheets and China and India have taken the lead on the wealth surveys. And there's some issues that need to be considered carefully, pensions and what to do about the upper tail. Thanks.