 Right, thank you, Paul. Thank you Richard. As in Paul said, this paper looks at the economy-wide impacts of the and risk associated with the farm-input subsidy program in Malawi and the way we see it a very important part of this growth poverty puzzle that we try to tackle in the growth and poverty project. The study was really conducted as part of another UNU-wide research program on development and the climate change. Co-authors James Thurlow, who until very recently was with UNU-Wider and training on to you who is also very closely affiliated with UNU-Wider. Now, as Richard explained, agriculture is a key driver of national GDP growth in Malawi. We've had massive or very rapid agricultural growth and as a result sustained national GDP growth of about 67 percent and one would have expected significant poverty reduction. Now the gap study that Richard just talked about raises some questions not only about the national growth figures that be or the national accounts figures that we have but also about the official poverty outcomes. But if taken at face value then it's certainly disappointing and one would have expected a much bigger impact from the farm-input subsidy program. So what we try and do in this study is to first of all add a piece to that growth poverty puzzle but then secondly also contribute to the first evaluation literature by conducting an economy-wide assessment of the program that really tries to identify all the impact pathways and spillover effects rather than just looking at impacts in the sector alone or impacts for beneficiary households on their own. And through doing this hopefully we can kind of isolate the impact of the farm-input subsidy on growth and poverty in Malawi and see whether that's consistent with what the official trends are saying. So the farm-input subsidy program has been implemented annually since about 2005-06. It's very widely targeted. Over half of small holds in Malawi you're talking about 1.5 million farm family households. Probably about 40 percent of the total Malawian population receive this subsidy and it's also very generous. The fertilizer and the seed inputs that are being provided under this program according to my kind of back-and-on envelope calculation should produce enough maize to satisfy a family's annual maize demand. Maize is a very important staple in Malawi accounts for about 60 or 70 percent of calories consumed by the household. But it's also very sorry and along with that the fertilizer is not entirely free. The seed is free but the fertilizer has a redemption fee attached to the voucher of about 500 kwacha. That's about $1.50 to get 100 kilograms of fertilizer. So it's not very costly to the household to participate in this if they're lucky enough to be selected. It is very costly about 3 percent of GDP, 70 percent of the agricultural budget. So clearly some important opportunity costs in terms of spending that we need to think about. Initially there was a great deal of support also globally. I mean the program was on the front or there was an article on the front page of the New York Times. But as time progressed I get the sense that there's growing skepticism about input subsidy programs generally but also specifically in Malawi. The obvious implementation issues are on procurement and logistics and corruption but then also bigger issues that need to need to be addressed. First of all fiscal sustainability. What about policy alternatives? Can we achieve the same goals with other policies that are less distortive in our economy? How well would the program perform under weather risk? In 2008-09 we had a doubling or even more than that about a 140 percent spike in world fertilizer prices which had tremendous implications for the budget of this program and eventually was one of the causing factors of an economic crisis in Malawi. So the program does come with serious risk concerns that we need to take into account as well. Evidence that we have at the moment ranges from suggesting marginally positive returns to the program to more recent evaluations, looking at current prices of fertilizer and maize, suggesting relatively high returns to this program. Not only in terms of the grain output but also in terms of hustled income of beneficiaries. The implications for the rest of the economy is less clear and that's where we try and come in. So what we're doing here, and I lifted that top paragraph straight from the paper simply because it explains it quite well, it's kind of a mixed methods approach where we harness the strength of both exposed evaluation data. Not anything that we had done but a lot of the analyses that have already been done, the data is there, we use that exposed data. We triangulate that with other information we have about input use and technologies within agricultural sectors in Malawi and we then build this into our ex-ante modeling framework which is an economy-wide framework to really get a comprehensive picture of and what we think is a fairly unique approach to program evaluation. So essentially what we do is we have a CG model, we calibrated it with the 2003-04 SAM. The reason for that is that that is a good base here for a farm input subsidy program evaluation. This model includes a traditional maize sector which looks like maize produced, the way we produced maize before the input subsidy program came along. But then we build into that new sectors which we're going to call FISP sectors and we're going to have a FISP maize sector which is where all the subsidized maize is going to be produced. We have a FISP fertilizer sector through which we import the fertilizer needed for this FISP sector and we have a FISP seed sector as well and we take program design elements and farm level evaluations of the program to carefully construct these sectors and what we then essentially do is in our simulation we exogenously allocate 500,000 hectares of land to these new sectors, blow them up to scale and as they grow they pull in the resources that are needed in order to produce the maize that we expect to be produced under FISP and it ultimately then exactly replicates the impact of the farm input subsidy program at the sector level. So we can then measure these direct production effects of the program but the usefulness of having this embedded inside an economy-wide model is we can also look at the spillover effect into the rest of the economy specifically those associated with price and income transmission effects, resource allocation, I mean there's tremendous competition for resources when you when you introduce a big program like this, also financing implications and balance of payments effect. None of these kind of indirect effects have ever been evaluated so as far as we know this is the first paper that attempts to do that. We can also then because we have this exante model framework now set up with a simulated farm input subsidy program we can look at things like what happens if we change our assumptions about marginal returns to fertilizer use which is a very important parameter in our benefit cost analysis. What about playing around with program scale or with fertilizer price shocks or with with the variability? So it becomes a very useful tool not only for traditional impact evaluation but also playing around now with alternative program designs and risks. This is basically how we build the new FISP sectors. So there's FISP maize, we have two types of seeds that are being distributed under the program and one is a composite open pollinate variety other one is hybrids. Back in 2006 or 2007 the program that we specifically evaluate or the year that we evaluate about 60% of the seed that was distributed were hybrid seeds. So they have slightly different characteristics in terms of yields, in terms of drought tolerance but also in terms of recyclability. Hybrids need to be purchased every year. We just do a static analysis so we don't actually look at the this this option of recycling the composite maize seed varieties which is something that we may need to doing in the future. You'll notice that we assume based on recommended fertilizer application rates there's a much higher and and because fertilizer is being given for free a much higher fertilizer use in our FISP sectors compared to our traditional maize sectors. Also exclusively use of improved improved seed varieties we specifically look or take into account the the seed planting rates. Slightly more labor needed and then ultimately you'll see the big yield differences between these different maize varieties. So essentially as we shift maize away from traditional agricultural sector shift land across to these these first these subsidized sectors we do expect average maize yields to increase. What do you mean? Now this yield can be decomposed into a component associated with the seed itself so the seed characteristics but then also component associated with the marginal return to fertilizer use which ultimately comes from this assumption down here that for every kilogram of nitrogen that you add to the soil you're going to get roughly 15 kilograms of extra grain for composites and 18 for for hybrids and these are the same numbers that are being used in what we can call the the official valuation of the farm input subsidy program in Malawi being done every year by by Andrew Dawood and Salas. Right so looking at some of the results the first one that's of interest is is the top one here maize production. We have two types of funding scenarios the first one was really just to kind of see well what if this was a free gift and donors paid in full for this program. The second one is the tax funded one where where the Malawian government as is currently the case pays about 80 85% or so of of the program costs so focusing perhaps just now on the final column we have an increase of about 300 000 metric tons of maize which is a little bit less than what the official program evaluation suggests and the reason for that is we take into account the allocation of land so what's happening to total maize land initially we have about 1.5 million hectares planted to maize that's your that was your kind of traditional maize production we then shift 500 000 hectares into into fist maize and we tell the model well you know don't reallocate land across all the other agricultural sectors based on on returns and profits in those other sectors and ultimately what that happens is because we're producing more on less land it's actually profitable for farmers to reallocate land away from traditional maize and into other sectors and particularly export sectors in a minute my average average maize yield increases from about 1.32 in the economy to about 1.81 so 0.49 increase metric tons per hectare because of this reallocation of land we have an increase in crop diversification fairly significant declines in maize prices which is consistent with some of the more recent kind of micro-level evaluations that have been done a real exchange rate effect obviously when it's donor funded there's a there's an appreciation in the exchange rate but there's a depreciation when it's funded internally and when we have this depreciation we actually see down here a significant boost in exports as farmers reallocate land into export sector specific so in terms of GDP affects roughly 1.9 percent impact on on GDP and absorption effects etc and this absorption number is is what we use really to come up with our own benefit cost ratio now the official evaluation for the same year 2006 came up with a with the benefit cost ratio which we term a production based benefit cost ratio of around one depending on assumptions but it was roughly around one we we we get a very similar benefit cost ratio 0.92 and the reason ours or we think the reason ours is slightly lower than under Dawood's estimate is because again because of this land reallocation that we permit right so farmers simply decide well we don't need to grow that much maize buying the market but the important thing is our what we call our economy-wide benefit cost ratio and this now takes into account all these indirect spillover effects from the program and that's about 60 percent higher than our direct production-based benefit cost ratio and so we see there are really significant indirect effects that none of the the other program evaluations have been able to capture right just something on on factor returns in poverty fairly significant increases in in returns to land because of increased productivity and then average farm wages and also increase and that that ties in with this assumption of ours that you know the subsidized maize production is a more labor intensive activity so we there's an increase in demand for labor um decline in poverty of about 1.78 percent or percentage points I should say right so um not massive right um but certainly not insignificant um so the program does have um some poverty um effects but also fairly significant effects for for urban households even in this scenario where we primarily tax urban households and get them to pay for this program and that is that that's that's linked to the um to the fact that urban households actually benefit a lot from declining maize prices as well and obviously um you know from from logistics you know from the transport contracts that are that are linked to the program which we also account for um now we get to this issue of the marginal return to fertilizer use and our reading of or our average considering the 15 and the 18 that I had up before um and 60 percent as hybrid so your average you know return roughly crudely speaking is about 16.8 kilograms of grain per kilogram of nitrogen that we add to the soil right and that's what yields our benefit cost ratio of 1.62 now some of the some of the evaluations that we've that we've seen particularly one by by um Jacob Ricker Gilbert and Thomas Jane and others um their number suggests that the that the marginal return to fertilizer use is actually closer to to 12 which um and door would questions but if our return was that low we'd no longer see these positive economy wide this positive or a benefit cost ratio greater than one um so there's I think there's a lot more work needed in this area and a lot of people are doing work in this area to really fully understand you know what our farm is doing with this fertilizer that they that they're getting are they using it appropriately and you obviously have this trade-off issue as well um if 70 percent of your budget is going into supplying fertilizer there's not much left to actually pay for extension services and maybe you know you know the moment this return starts slipping you start losing the gains from this program right we also looked at fertilizer price risks um nothing um of the order of magnitude that we saw in 2008-09 but clearly a significant decline in these benefit cost ratios as the fertilizer price goes up um and as the fertilizer price goes up we have an exchange rate um depreciation right which is expected again and we kind of you know our economy is forced to to generate more foreign exchange revenues um in order to pay for this more expensive fertilizer so what happens in our model endogenously is is quite a significant relocation of land into the tobacco sectors tobacco being the the major export sector in Malawi um the now I you know I I haven't had a chance to to look into this carefully but you know thinking about the 2008-09 price shock it wasn't long thereafter where Malawi suddenly oversupplied tobacco and it's a it's a major supply of burly tobacco the tobacco price collapsed and that was the start of the financial crisis in Malawi so maybe something to this effect did happen it was you know possibly um you know you know the start the fertilizer price um increase was possibly the start of of the financial um crisis eventually um also much smaller poverty implications as well when when the fertilizer price does go up now the last thing we did was look at um whether variability and how this program performs we do have draft events and here we draw on on some earlier work that that James and I had done on on draft and and and economic or production maize production in Malawi and we actually have some data working with with a group of meteorologists from from India showing the production losses associated with drafts of different return periods and these return periods tell us about the not only about the severity of the drought but also about the probability of that event occurring and we noticed that local the red line here is the local varieties local maize varieties much more um vulnerable when it comes to droughts um so much smaller losses for hybrids and our composite varieties actually supposedly the most resistant to droughts so what we thought about doing here was well actually if we are shifting into subsidized maize production which includes and you know importantly includes a seed component and increases the the the amount of improved seed varieties then we're actually buying not only higher yields but we're also buying greater drought tolerance um when a drought does come along and so the the dashed lines down here show your benefit cost ratios for droughts of different um severities this vertical line is a roughly a a one in seven year drought which um i won't go into this but that's kind of your average annual expected loss which we talk about a lot more in the other paper um but essentially what we have here is you know so for the actual program which is the red line um we have a decline um in the benefit cost ratio and only when we reach a roughly a one in 13 year drought does it drop below one but then um just very briefly say what we really should be thinking about well what if we didn't have FISP right we would have had much bigger drought losses right um so this is an adjusted um or a baseline adjusted um so looking at the benefit relative to what we would have achieved if we didn't have FISP and and still you know had a lot of local maize varieties in our in our seed mix right so conclusions um FISP is reasonably pro-poor has the potential to generate substantial indirect benefits and we we strongly believe that um you know that this approach that we that we've developed here um strongly complements uh some of the uh partial equilibrium or micro level survey based methods benefit ratios depend very strongly on the marginal return to to fertilize a use so this important area um of you know for intervention and to make sure that the program remains viable and we need to um make sure that these that these marginal returns do not drop off and and really we need to get a better sense of what they actually are in the first place um we did some work on real fertilizer prices um and um saw that macroeconomic constraints in terms of paying for the fertilizer then really come into play um the benefit cost ratios that we said they they they fall during a drought year but the FISP does generate double dividends of higher and more drought resilient yields so there's a kind of a hidden benefit to this program as well thank you