 Okay, Luis. I wanted to make some comments on the first presentation. First of all, I guess I find it a little bit discouraging that you say it's not clear how the official results were done. About two years ago I was actually working with that survey data and helping the World Bank team help the government to come up with those results. I'm not going to fully defend the methodology. I will say that the reason why they did a fixed share with one national poverty line in 2010 was that was done in 2004. And the view at the time was that if they then brought in improvement in the poverty line like having, say, four poverty lines for the different regions instead of one, they'd have to recalculate the 2004. The NSO was firmly against recalculating the 2004 because they felt it would look like the books were being cooked. So I think it's important to think about it as we think about different ways of doing poverty lines and how to do it. It's not that I'm persuaded that was the right decision. It's just an important consideration. Now, in terms of the – I am, however, a little bit surprised by your results. Prima facie, if the food share is going up, that is an indication that people are poorer. Not perfect, right? But, I mean, we start with an angle curve and it has some indications. So I'm a little bit surprised about your bundles. The second thing I will say is that this may not be exactly on – this particular story may be not exactly on the NSO website. But what I know is that when we first looked at the national price index, the national consumer price index, and the food prices that were obtainable from the survey, the 2004 and the 2010, we noticed a huge gap between the food inflation as measured by the prices in the survey and the food inflation as measured by the prices in the market and the national consumer price index that was computed. Now, that's not unusual. The size of this gap was huge. As a result, however, some technical assistance was provided by South Africa who came in and re-estimated the food price index over that period and it was adjusted substantially upward. And it was that new price index which was used to compute the poverty line. I am surprised that you were coming up with a lower food inflation using the prices in the survey than the restated national prices, which – the poverty line done by South Africa, which was significantly below the prices that we were able to compute from the prices in the survey. So that is kind of a puzzle to me. We also did a survey to survey mapping between the two and we, of course, using mostly assets. And one of the things we noticed is that the – rural inequality ridened substantially and the increase in assets in rural areas primarily went to the upper-income farmers. So the people who got more sheet roof, for example, tended to be the upper-income farmers and the lower end of the distribution did not show an increase in assets. So that also – that survey to survey mapping should be available. If you haven't seen it, you can talk to Tala – Tala Gillip or anyone on the World Bank team that can provide it. So that also gave us some confidence that regardless of whether the levels were right, that the direction suggested not too much poverty reduction between 2004 and 2010. So I am a little surprised by your results. Finally, I would suggest that this experience with the consumer price index maybe perhaps might lead one to question the GDP figures. So perhaps it's possible that GDP did not increase at 7% per annum. I know that there was going to be further work to strengthen the national accounts. I don't know how that proceeded. There's been a lot of discussion about the quality of national accounts data in South African Africa and I won't repeat it. But it may be – it might suggest why the poverty reduction wasn't what you expected. Maybe growth wasn't actually as high as you got. Thank you for two very interesting presentations. I had a question about the first one. I just – a comment and a question. So the question is – it looked to me by the numbers that you put up there that urban poverty had fallen more than rural poverty. Is that right? So I find that puzzling given that I thought the program was targeted at the rural poor and which kind of leads to my second comment, which is, you know, you posited this as a puzzle why poverty isn't going down. But as an onlooker observer that's not intimately familiar with Malawi knowing about the food price hikes and so on in the letter half of the 2000s wouldn't have surprised me if poverty increased over that – the time period that you're looking at 2005 to 2011 because food prices spiked in 2008, right? And they stayed pretty high. So I'm just – it's just a question. And that relates to the second presentation. Have you found any way to include in your overall – your general equilibrium analysis trade? And I've heard stories about maize being locked up in warehouses and going to rot because there was so much maize. I mean, is that true? And if it is true, is there something that can be done about that? And more broadly, like, what about this FIS program as it relates to other countries in Africa? Because if everybody does this, if every country does this, right, won't there be a surplus of glut in Africa? I don't know. Maybe that's a pie-in-the-sky scenario. But anyway. Again on Richard's presentation, I wanted to ask if your inflation dates are credible. Malawi's not such a big country. You have much, much higher prices in the rural north in the first year and you have much, much lower inflation in the rural north. Why is that – I mean, is that plausible? And I mean, you said your fixed basket results give something similar to the official. They don't. The difference between the different rural areas is very different. And again, the rural north comes out here. So I sort of question the basic inflation dates of the two years. Thank you for the comments and questions. I will try to address some of them. I think my colleagues can step in as well. On the urban poverty decline that is bigger than the rural decline. I mean, we come up with a story which is very speculative that actually because urban areas are net consumers of maize and because of a massive increase in production, prices are significantly lower. They're benefiting from the reduction in prices. Also, the logistical – it's urban households that are going to benefit from transporting this fertilizer or maize or whatever to the rural areas. So it's kind of speculative, but that's one of the arguments that we make in the paper that maybe that is driving this decline in urban areas which is bigger than the rural areas. On inflation, I mean, inflation figures are dodgy. I agree with you that they're dodgy. And actually just to come back to what we're talking about, we have two sets of conversion factors that we are given by the NSO, right? So we used both. The results that we get are totally different, okay? We tried to play around with them coming up with averages of the conversion factors using conversion factors for one region. We tried different ways of fixing this conversion factor problem, but finally we ended up with what we have here at the moment. It's an issue. It's a big issue. We talked to them. Actually what we have presented here, some of the findings were also presented at the NSO to see – I mean, to have a chat with them and have a sense of what is going on exactly. But again, it wasn't clear as to what was driving this issue of conversion factors. Actually, conversion factors are a key part of what we are finding here because it seems to me that whatever we choose to convert our different unions are kind of driving what we end up getting. I've noted what you've raised about redistribution towards the upper-end rural areas. I think it's something that we can also include maybe in the paper. We have interactively started it. I think we had a meeting to remit. But through email we have kind of interacted with him. But actually we got the conversion factors from him. Do you make up your stats at only one or four? Yes. Otherwise that's the issue that I've raised. I think we make all those things. Yes, thanks Louise. I also really appreciate your comments and insights. I mean, we've been to the NSO and it's sadly – it's quite clear that they don't really know what's going on with the conversion factors, with the revision done by stats of Africa. We haven't been able to get answers from them on that all-technical documentation. Talip, as you know, is a very busy guy so it's quite hard to get him to meet with us. So it's been a struggle to kind of fully understand what went into the NSO analysis. I'm a bit concerned, frankly, about the revised inflation estimates. You have widely differing rural and urban inflation rates for food and non-food components. Particularly in the end, when they combine and estimate a national average inflation estimate, it's exactly the same for urban and rural – 128.9%. It looks like a bit of magic, to be honest. So we'd love to get to the bottom of that. So thanks for your comments. Andy, on the rural north, it's a puzzle. Again, with some of those conversion factors, the north has its own unique set of conversion factors which we decided to throw away because it's just too strange. In the north, they consume a lot of fish and cassava. And some of those conversion factors for fish and cassava had been clearly changed and we need to get Talip maybe again to explain to us. But some of these are for fish 45 times lower in the north than in other areas for dried fish. And it's a food consumption item that's consumed, I think about 40% of calories come from fish and cassava. And those two conversion factors have been manipulated to, I don't know, achieve what may be, you know, not get a 16% reduction in poverty. So there's some really strange things going on in those conversion factors. And that is something that we continue to work with. We have a colleague in Washington who's doing stuff on nutrition and he's really spent many, many hours on this and actually sent a new set yesterday which we need to start looking at. The Maggie on trade, yes. There's been lots of questions asked about where is all this extra maize. The production increase that we get of about 300,000 tons is believable. The official estimate said we produced over a million tons extra, 1.5 million. And everyone's saying, well, where the hell is the maize? It's not in the silos, it wasn't exported. You can't export that much maize on the back of a bicycle. So I think it's simply a case of over-inflated maize production estimates possibly to make the programme look good. And that then ultimately drives the national growth estimates as well or the GDP estimates as well for agriculture. So absolutely, I think the 6%, 7% GDP growth is an overstatement. So the truth probably lies somewhere in between. Moderate growth, moderate poverty reduction rather than... Okay, can you please take the last set of questions because we only have five minutes? Please go ahead. Yeah, I guess the issue with the food and the non-food. Did you look at the national CPI numbers for non-food inflation and how that relates to... I mean, you've got your unit values that some of us are a bit concerned about because you're looking at apples and oranges with chairs and whatever else. But it would be helpful to see that. The more I hear these conversions factors and the way you set up the paper about trying to understand the rationale for growth and poverty reduction, I don't know, maybe you have too many degrees of freedom but that's being a bit naughty on my part. Okay, my name is Haruna and I'm from Uganda. So my question is to the last presenter, Mr. Carl. So I'm much interested in the way the picture of the analysis which seems to have serious policy implications. And from that analysis, it much shows that much of the increments and the reductions in poverty much depend on the productivity of the fertilizers. Perhaps if the fertilizers still work well and maybe there is increment in misproduction and so on then you can see the reduction in poverty and so on. But my skepticism lies from the fact that here is one point or one factor which actually stands out to be responsible for the positivism which the analysis shows. And my understanding is that the earlier level of mid-welfare had a contribution of perhaps several factors. I know there can also be other factors which can also influence the productivity of the fertilizers. For instance, the quite of the soil and the weather and so on. And to me, if all these credit is given to the fertilizer section, then I would wish to know your opinion on if perhaps would respond that you would recommend the continuity of the program which is only dependent on perhaps a single kind of factor to drive an economy. And I think that's how we see so many scenarios and people think it is all this production and it's not being exported and so on. And with much of the policy makers in Africa you've got to tell him that perhaps he's going to continue pumping money into the subsidy and maybe something else needs to explain the difference but maybe not only the fertilizer. So my interest is in your opinion, your own opinion as to whether perhaps we recommend policy to actually sustain the program for some time or maybe you would limit the sustainability of the program to the same duration of time or perhaps maybe if you're not going ahead to assess how sustained or for how long this policy due to the fertilizer can go, maybe you would still also guide in that way. Thank you. I said a couple of quick questions, one for each of the presenters. On the Richard's presentation, I was wondering if you worked, I know you mentioned IHS-1 and I've dealt with it, data set. It was a little difficult but as a robustness check it would probably be a bit of work but if you could see what your poverty lines look like if you change from one to two and then two to three because it does kind of come across as you were maybe were looking for something or you had questions about three and so one to two might help giving a better full image of what's going on. And then for Carl, I'm wondering if you made any assumptions about targeting of the program. I know that there's been a lot of questions about FISP and whether or not it actually was given to kind of poor households and if those assumptions would affect the poverty reduction measurements that you came up with. Thank you. Only using IHS-1 and IHS, using all the three waves of the IHS, that would be great but the problem with IHS-1 is the coverage of the consumption model is not as extensive as IHS-2 and IHS-3. So that limits actually what you can do. And this is why actually if you look at IHS-2 and IHS-3 in terms of the consumption model it's directly comparable. IHS-1 is just totally different. You can't, there's very little that you can do I think with IHS-1. Only non-food component of inflation. I mean in the paper we talk about the official inflation figures in quite some detail. Sir? Thank you. So there is the CPI inflation rate which is quite low and then there's 128% which is higher than the CPI inflation and which was used to inflate the poverty line and then there is our inflation rate. So there's three different inflation rates going on. I think that's maybe the issue. Not the one that they used to adjust the poverty line. The official one is 77, I think, 0.3. The one that they used to adjust the poverty line is 128.9. Again, like I said, it's not clear why they used the other one. It's just... Are there too many mistakes? Yeah, sorry. I'll start with Ben's question on targeting. Very good question. It's not something that we can deal with adequately or easily in the CG model. Because this is a program targeted at essentially what we will call producers or activities in our model, households are only linked to activities indirectly via factor market. So if we want to target households it's kind of tricky because you have to target them indirectly via the activities and via the factor market. So we simply assume that the benefits from the fertilizer or from the subsidized sectors are distributed among households in the same way that any maize income would be distributed. So it's kind of neutral in that sense. It doesn't change our income distribution. The question from the frontier, I must admit I did lose you along the way. So I think maybe just to very briefly comment, I think, yes, I fully appreciate your point that it's not all about fertilizer and there are other things also that impact on productivity. That's certainly true. What we set out to do is not to... Because this is a parameter in our model when we show the sensitivity of outcomes to that parameter, what we are arguing is that we need to really understand how we can measure marginal returns to fertilizer use, which is a very, very tricky area, as you mentioned. It's about other technologies as well and weeding and rainfall and all kinds of things. And I think thus far because of all these multicollinearity problems in estimating that parameter econometrically, the work that's been done so far falls a little bit short. So we still don't really know what the right number is. What we then do is just ex ante that kind of Ketris paribus type stuff. Obviously we can recommend that farmers should adopt certain techniques as a way to improve return, but I don't think that's really the ultimate aim of the study. In terms of continuing this or having a kind of a sunset clause on this policy, again, we evaluate the impact of the program and we see that it has positive returns. A next phase would be to suggest policy alternatives and look at the impact under alternative programs. So again, I don't think what we've done here necessarily helps us understand the question of whether we should stop doing this a few years down the line. I'm afraid I think we only have one minute. If you can be quick, that would be great. Just the financing in your model of the fiscal expenditure, because you said you talked a little bit about how you took account for absorption. How is that 3% financed? Do you have a fully specified model then that does that? Yeah, we just have a closure on our government budget where we would raise hustled income taxes to finance the extra. That's not really hard to happen, but it kind of accounts for the opportunity cost. I think what government really did was reorganize their budget to pay for this. So slightly different model, but by taxing hustle, do you kind of capture the opportunity cost of having this program? Okay, our session has come to an end. I would like to thank all the participants, all comments and views and insights. It will make sense of Malawi's poverty puzzle. So our colleagues, I guess they will go back and look into all your comments and inputs to improve on their papers. So thank you very much for coming.