 So when Gary asked me to be a discussant for this session, I was pleasantly surprised because I don't work on South Africa, not Latin America, but also I was very pleased. I was very pleased because the three papers we've seen are from bodies of literature, which I greatly admire. Carlos's work with Luis Felipe, Nora Lustig and others have been very important at the World Bank and live up on the same level markets. Gary's work, of course, not only the Latin America book, true UN wider, but also the work, the book that he mentioned earlier and everything else he's done, has been fantastically influential in my own thinking of the living markets. And of course Haroon's and his colleagues work in UCT is the best work among the best work I've seen on living markets in the global south. So a chance to reflect on the three papers is a great opportunity really for me. I want to make some overall remarks and I'm going to try and get to a synthesis and then think a little bit about what does it mean for other middle-income countries. This amount of sort of, you know, we clear about the comparative dimension here. We have three presentations with examined the patterns and determinants of inequality in middle-income countries, Latin America, several Latin American countries in Gary and Carlos's case, and South Africa and others in Africa. And of course, you know, these two regions, if you wish, or one country in one region, provides interesting contrast really because we see, as we can see from both Gary's and Carlos's papers, that there's declining income inequality in Latin America and an increase in South Africa. So clearly the question is why do we see these differences? So what role did patterns, economic growth, national policies, labor market solutions play in this divergent trends? And of course, what lessons can other MIC countries in particular learn from these country experiences? I was asked by a guy not to get into paper-by-paper, but I can't sort of just now say one or two comments about each paper for a methodological point, even though I get to the synthesis and other things. And so the methodological points. I was not sure, Carlos, in the decomposition he did, that essentially when you're looking at a wage decomposition, a kind of blind hole, how could he come if you wish, if you're saying residual inequality explains most of the wage inequality, where are the coefficients coming in there? Because you already shown that a lot of what you're seeing in Latin America is a decline in education premium, a decline in experience premium. So pretty much isn't that what's really driving the decline in inequality? And so what is residual inequality here? Is that the famous residual we have in the TFP case, which can't explain anything, I would like a little bit more structure on that. So I wasn't sure methodologically, you know, that was very clear to me. On Gary, the point here I would say is that, you know, if economic growth is not really driving poverty declines, we see a very weak correlation there, but mean in earnings is what's driving mean-label earnings? I mean, what else can be the causal factor behind the increase in mean-label earnings? I wasn't very sure how that comes into play here. On Haroun, I love the RIF approach. We are great admirers of circuit-flippers work, but I just wonder the circuit-flippers and the approach you used kind of puts, you know, the very nature of the approach puts everything in the coefficients. In a sense, I mean, so when you mimic the U-shaped distribution, the weight distribution using the RIF approach, and you get pretty much a clear kind of fit between the coefficient effect and the actual distribution, nothing much on the characteristic side, I think why would that characteristic effect have any role to play? Because you can't imagine composition changes in characteristics really being important in the weight in the U-shaped US distribution. So almost, is it by default, tautologically, that you're putting everything in the coefficients? So I'm not sure how to think about that, but it's worth thinking a little bit about, is there a kind of problem here in the way we use RIF regression? So I'm not an expert in this, but perhaps it's worth thinking about. Okay, so on the synthesis side, what we really see clearly are labor market institutions matter a great deal in explaining inequality movements. Again, there were still things I wasn't very clear about. Pattern of growth mattered more than the rate of growth in South Africa. Very clear argument here about the fact that we have this service sector growth that's fueling polarization, at least on the top end, and also manufacturing hasn't really done well. But in the case of Latin America, I wasn't very clear because the rate of growth doesn't seem to have a fit as Gary shows, but what kind of pattern of growth do we see there? We haven't seen manufacturing there at all, and it's been mostly commodity-driven growth in Latin America in this period in most countries. So I wasn't clear about the pattern of growth story when it came to Latin America. And a similar problem with the skill bias technical change story here, because it seems to play a different role in Latin America than South Africa. Why? I mean, skill bias technical change should really be driving up inequality in pretty much every country where it was opening to the world. Latin America has been globalizing too. So I wasn't sure why is it that we see the decline in education premium happening in Latin America while clearly in South Africa is what we expect has been increasing inequality. So it's something that I wasn't very clear about. And I also feel that even the question of education premium falling in Latin America needs to be thought through a little bit more carefully. Is it due to less demand for skilled workers, or is it increasing supply of college-educated workers? Very different implications, right? So again, we need to think a little bit more carefully about why do we see this counterintuitive decline in education premium, which is not what we see in most other countries where I've seen the data on. So that's something I couldn't really understand. But finally, I would have liked a little bit of sense of the role of trade. Half-stout stopper-samson effect has been more favorable in Latin America. In the 2000s, we've seen earlier work by Nina Pavnik and others where they show the stopper-samson has actually had unfavorable effects on inequality in Latin America in the 1980s and 1990s. So what is trade playing in this story for South Africa and also for Latin America? It wasn't very clear. And we know globalization is a big, big thing here, both in driving true stopper-samson effect, but also true skill-biastectical change. It wasn't very clear. So I wanted to get a sense, get a sense of that. Okay, so possible lesson for other MICs. You know, it's interesting, the Latin America story is saying that it's possible to increase in, possibly reverse increase in quality, the appropriate government policies, particularly minimum wage increases, which we see certainly have seen in Latin America and South Africa also, as everyone has shown to us. But a whole word of caution here, you know, it's possible to increase minimum wages when the economy is doing well, which is what we saw in Latin America. We had reasonably strong growth in most countries. It's also clear, like right now, as you see growth essentially declining for many of these countries, is it possible to keep on following the same policies? So that's not very clear to me. So again, I mean, on the same point, while national policies may have matter of the decline in inequality in Latin America, strong growth, fewer but the commodity, more improvement in most countries, not all Latin American countries, may have provided a favorable environment for the dispute policies that they followed, whether it's also familiar in Brazil or progress in Mexico, and also minimum wage policies. Now, maybe the world is different now. I mean, if that's the case, what can governments do in this region? Something that, again, I think it's worth thinking about. The other thing I thought was interesting, especially coming from a possible aspect where I work mostly on India, that the pattern of growth seemed to matter really in the South African story. You see a stagnant manufacturing sector, along with the growth of the two-tier service sector, a high-end service sector, and the low-end service sector, which has been fueling which were all around South Africa. And if I had to think about what I would take from this presentation of the countries that we are, and the other MICs, I would say the South African story is probably more what we might expect because many of these countries in the emerging world are also going through premature deindustrialization, or tertiaryization, or financialization, whatever you want to call it. And so in a sense, the South African story may well be mimicked by other emerging economies. We don't know because we don't really have the kind of data sets that Haroun has been working with in the case of South Africa. So if I have to sort of think what are possible indications, while I think there is a role for government policies in the right moment, in the right economic environment, you could do something like increase minimum wages if that's possible to implement in your country, but also I think that what we might expect to see is more and more the story that Haroun has been saying in South Africa in most other willing countries. Thanks. Thank you.