 I'm going to shift that tension a little bit from inequality to structural transformation as this is one of the issues that we work on in the IMF these past few years. In particular, I'd like to focus on two dimensions of structural transformation. One is diversification, which is very intrinsically almost related to structural transformation. And I'll explain how here. And also the kind of policies that we need to worry about policies on structural reforms in helping countries move along the different stages of development. Now I wanted to start by saying that when we started to think about these issues about ten years ago, what struck us was that especially for developing economies, the data sets that were available were very few and very thin, especially for low-income countries. When you think about export data set, you immediately have in mind something like Comtrade. And for those of you who know Comtrade, you also realize that the focus is on advanced economies. So immediately we had a problem. We wanted to create data sets that are comprehensive enough going back in time, but also cover more than the OECD countries. And that was one of the priorities of this work, create data sets that are comprehensive enough to go back in time, but also include a lot of emerging and low-income countries. Starting with the work on diversification, we wanted to think a little bit more deeply about what that meant. And we went straight to some definitional issues. So first was the idea of diversification having basically two dimensions, one that had very little attention by the literature. The first one is basically export diversification from the perspective of how many types of goods you sent out the country. That's very easy to measure given Comtrade data, especially for advanced economies. What was missed, and we tried to cover in this data set, is the other dimension, which is the quality dimension. So the idea that for every type of good you sent out the country, there is a quality dimension that could be measured by its price. And basically we take advantage of Comtrade because Comtrade has not only volumes, but also prices for every one of the goods that you sent out the country. And we were able to develop this additional dimension. I'll speak a little bit more about this as we move along. The other dimensions, which are a little bit second order, but also important when you start thinking about diversification, is beyond products. Who are you sending these goods to? For example, you can think about the shift of trade from Europe to Asia, to China in particular, when you think about many developing countries over the last 10, 15 years. So with that in mind, we set for a very ambitious project to develop these comprehensive data sets both on diversification and structural reforms. First I'll talk about the data set on diversification. And then I'll spend 10 minutes or 5 minutes on the data sets on structural reforms. So we set out basically to fill a major gap for developing economies and low income countries starting from a very good starting point, the Comtrade data. For you who know about the Comtrade data set allows us to look at the source and the destination of any bilateral trade, which basically allows us to control for the quality of what goes in and out of a low income country in terms of trade. So for example, if Tanzania sent something to the Netherlands, we are able to check whether the Tanzanian trade statistics are accurate by looking at the imports in the Netherlands. And that's quite a bit of a comparative advantage to have compared to anything else, any other data set, especially on national statistics or anything like that for low income countries. So what I show at the top is the export diversification data set. As I said before, this is not difficult to put together given the volume data that you have from Comtrade. What was more difficult, and that's what we spend most of our time doing over the last five years, is to match these volumes with prices, which would give us an indication of the quality of any one of this bilateral trade. And again, for all of these developing and low income countries, this is what we call the additional dimension that was not there or was there in theory, but was not there in terms of empirical work. The outcome is this comprehensive data set that covers 187 countries, including most leaks, and goes back to 1962. Of course, as you may realize as we go back in time, the work becomes more difficult. There are more holes, but we are convinced that what we have at least, the observations that we have are quite accurate. So I want to spend only two minutes just providing a flavor of what you get out of this data. We have a lot of stylus facts in working papers and publications. Here I'll just give you only a couple of slides on cross-country facts, and maybe one slide on when you zoom into a country-specific facts. What you see here is using our data set, where in its most aggregate level, one observation is a country-time pair. And what you have on the vertical axis is a concentration index. The way to read this out is as you go up the index, you have more concentration, less diversification. And what you see here, using our data set, is the gold standard in this literature. This idea is based on an IMPS-Warziak AER paper that basically revealed this U-shaped curve. The idea that low-income countries first diversify up until they become middle income, and after that they concentrate, they specialize, basically. Just to preempt what we will be doing, what we will be arguing later on, is that without the leverage of knowing what's going on in terms of quality, you don't have a good story of why countries start re-concentrating. With our new data set on quality, we basically make the argument that after middle income countries start re-concentrating because they focus on quality, because that's where they have the comparative advantage. That's where they get most return from. Now, notice that at the very top you have some observations which are outliers, if you want to call them. These are all the commodity exporters in small countries. These are basically the countries that don't diversify. Okay, so why we care so much about these issues is because we have shown in publications and the latest is out just a month ago in the Canadian Journal of Economics by TUICare that shows that diversification is positively related to growth and negatively related to volatility. Here I'm just showing you some correlations focusing on low income countries, but in these papers they establish causality as well. Now, it would be an omission if I don't dive into the country-specific examples. The data set is there for research, for cross-country analysis, but a lot of our country teams in the IMF use them for their mission and discussions with authorities. For example, some of them are focusing on what we call here diversification episodes, extended period of times where you see diversification either in terms of volumes or quality. You have interesting cases like Chile, Malaysia, Thailand from emerging markets, but also a number of low income countries. Then these teams are trying to connect this with volatility and growth, trade and so on. Let me very quickly now move to the most important addition to this work, which is this new dimension that we uncover on the quality upgrading. Again, for every observation from Comtrade, every volume observation, you have its price. From that, we are basically relating that price to quality with some econometric method and we have a measure of quality for every bilateral trade transaction in the millions. Taking this, we aggregate and again what you see here, every observation is a country-time pair. That's the most aggregate level here. What you get out of this, where on the vertical axis you have quality, is that quality matters a lot for development, but it matters most when countries go from low to middle income. It's binding. It has its most effect when countries turn from, I don't know, something like Cambodia, Vietnam, Laos to something like Philippines. That's the green area here. Notice the variability that exists in low-income countries is huge, but as you go and you become advanced, there is much less margin in terms of quality. Basically, you hit the frontier. What we can do, again, zooming into some country specificity here, I wanted to show you what our toolkit of diversification can do for some of these country-specific examples. Look at, for example, Tanzania. We can produce these nice quality ladders, as we call them, where they identify the ladder, the possibilities that exist in progressing in terms of quality, and where you actually are as a country. The ladder is indicated by the red vertical line, and where you are on that line is identified by the yellow dot. Here, you see an example with SITC level 1 trade data, where is the most aggregate. So, for example, you have categories like food, live animal, and manufacturing, and so on. So, if we just stay in one of those bar charts, say, the food, live animals, you can see on the left-hand side, the percent of exports. This is the export share. For Tanzania, it's more than 30%. You see the quality ladder on this red line, and then you see where Tanzania belongs in this line, which basically tells you that in terms of quality, there is a lot of potential for Tanzania to progress in this dimension. Again, this was non-existent before we constructed this data set on prices and evaluated diversification in this additional dimension of quality. Now, one thing I wanted to mention here, and I will move on, is to recognize that this work can become much deeper as you dive into more disaggregated sectors. SITC level 1 is the most aggregate with 10 sub-sectors. Then you can dive in for every one of these sectors to another 20, and then more and more to the fourth digit, with 857 sectors. So, here is an example of the analysis that one can do here. Same picture, same quality ladders panel, as I've shown you previously for Tanzania. It's presented here for China. If we say we focus on the miscellaneous manufacturing articles for a minute, we can dive in in this particular sub-sector to the second digit and identify where China belongs in that particular ladder. Here is what I mean. So, this is now SITC level, one level of disaggregation further. You identify now all of these categories that belong under miscellaneous manufacturing articles. You have the construction of this ladder, which is this red line, and the yellow dot where China is for every one of these sub-sectors. You can keep doing this up to the fourth digit, which is this 800 sectors. You can also do this dynamically. What I have shown you here is only for 2010. You can start from 1960 and see the evolution of this until today. All of this is summarized in what we call the diversification tool. It is very easy to see. It's publicly available. You Google diversification, IMF, and you will get this. There are many interesting examples, a manual, and so forth. This is one dimension of the work that we do on structural transformation. The second, the more recent, the one that we are still working on is not finished yet, is in constructing a major dataset with different types of structural reforms. You may know of some work that has been done by OECD on these issues. We consider ours to be a lot more comprehensive. Usually the OECD focuses on labor product. What we have, what we are focusing here beyond labor product is financial, trade, and also dive much more deeply into the labor market's reform. Notice that one of the key advantages of our reforms dataset is, again, that it covers not only advanced economies, but pays a lot of attention to emerging markets and many low-income countries. The entire sample that we are working on at this point is 90 countries, but we expect to increase this to about 120 in the future. Again, one advantage of our reforms database is that it's coming from staff reports. Most, it comes from IMF staff reports, so we can identify exactly the source of the index number that we produced. So it's fully transparent. We can identify why a particular country has a particular index value for a particular year, and even try and debate this if need be. So all of this will become eventually publicly available when we have the data ready. Okay. So again, just like I've done with diversification, I'd like to focus with a couple of slides on cross-country patterns, and then maybe a couple of slides on country-specific facts. So this is very preliminary results coming only from one year, 2014. And what you can see here immediately from this spider graph, you can do very interesting things. Two minutes. Yes, I'm finishing some, thanks. You can see very quickly how you can map the different types of structural reforms, including labor, financial, product, trade, current, and capital accounts for all countries. Here is an example only for advanced grouping countries into advanced, emerging, and low-income countries. Very quickly, you can see that the advanced economies are doing very well followed by emerging and low-income countries, not a surprise. Perhaps something that is very surprising that I will mention a bit later as well is labor reforms, where the priority is that there would be many differences between leaks and non-leaks in this reform. We find that actually that's not the case. And of course, the two extremes between leaks and advanced economies regard capital account and product markets. Okay, I'm finishing up Max, two slides. So very quickly, just to focus on a few of those reforms, trade regulation, trade liberalization has a pattern that is very much expected. What you have on the vertical scale is trade liberalization and on the horizontal development. You have more or less a positive slope, meaning that developed countries are more liberalized than low-income countries, although not as the heterogeneity that exists for low-income countries. What is surprising is when we go to capital account regulation is the sheer heterogeneity that exists, even when you focus on low-income countries. For example, look at what's going on in Mozambique compared to Uganda. Uganda has a complete open capital account and Mozambique very, very close. And this continues to be the case for emerging market. It's only when you get to advanced economies that basically most of these countries are completely open. And here is the result. I'm not going to speak about this. Again, is what I mentioned before that there's not much difference when you look at our data set in terms of labor reforms among leaks and advanced economies. You can zoom in and do very interesting work with country-specific work. Here is an example with Malaysia, where you identify what has been going on in Malaysia going back in time. And you can identify waves of reforms. For example, here we identify trade liberalization from 1985 to 1995, followed by capital openness and banking system, followed by labor. You can do this for all countries. The hope is that you can not only do this for the aggregate indices, but you can do this for the many disaggregate indices that we will have. You can also dive into the particular aggregate indices and decompose them in the many sub-indices. Here is an example with our financial reform index. And the colored bars show you the different components of the financial reform. I didn't have time to talk about the different sub-indices, but when we get the data out, you will see that the aggregate indices are very rich and additive. So you can go and do analysis for any of the sub-indices that we provide. So two more slides and I will be done. There's a whole idea here. The goal is to conceptually think about a framework. Perhaps a post-Washington consensus framework, where we try to advise countries on what the right reforms or waves of reforms are given their stage of development, their stage on structural transformation. Maybe the broader goal is to have a unified theory where everything is pulled together. You have a story of diversification that feeds a story of structural transformation as countries move from agriculture to manufacturing and services. And we advise countries on what reforms to implement as they converge to middle income and advanced economy status. Thank you very much.