 So there aren't going to be, we originally hoped that by this time we would have estimates of poverty impacts to present. We don't have that, but we do have a taste, a somewhat simplistic taste of some of the results coming out with regard to sources of yield gains. So the latter part of my presentation, if you hold tight to the end, you will get to see some preliminary findings about what the big winners in terms of yield gains will be. We haven't had time to process the input-output data in order to look at technologies that improve efficiency, that are cost savings, that have other more complicated effects. But there are some interesting results at the end of this presentation. To begin with, I'll start with the process and how we will use these data and how we will get to estimating impacts for the poor and on the environment. So just to review where we are, for those who may have just stepped in or if they weren't here before the coffee break, where we are right now is we have about 55 research solutions that have been assessed by participatory scientist working groups where they've looked at the resource requirements, probabilities of success in terms of having research solutions develop the time frames for those being developed, and then the expected adoption levels and on-farm input-output effects. These 55 solutions are still not all the solutions. Some of them are still being discussed. There are interesting points where we do have particular uncertainty or divergent views where we're still discussing. We also have probably on the order of something like another 20 that we still haven't yet done, that are ones that are particularly more complicated or maybe even areas where we do not have sufficient in-house capacity to make a good assessment and we need more to bring in external expertise. Those assessments of the research solutions have been backed up by analyses of problem prevalence and trends in those problems, and that's been complimented by what was presented in the first presentation, and that's a spatial analysis of trends in actual yield and attainable yields to 2035 so that we identify how the base yields that we should assume for one different technologies come in, what the base yields will be, and so we can look at what that scope for all technologies that help to close the yield gap is collectively over time, how that has evolved. We will use that to basically triangulate the information we get from the scientists. We don't want to be coming up with estimates that say that we are actually exceeding the yield gap through technologies that close the yield gap, so it's a good reality check on what we're doing. We'll take those estimates which Boss gave you a much more detailed picture of. Those are generally in relative terms in terms of effects on input use, different elements of production costs. We have an extensive in-house survey database covering many different Asian countries that we're going to use to get the baseline values in terms of input-output use by operation so that then we can translate those relative changes into absolute changes and we can look at what the unit cost of production is and how that's being shifted by the particular technologies. When we do that, we also will take into account the adoption trajectory over time and the adoption trajectory over time that would be likely for the solution without yield. So that will all be reflected in the effect that's estimated on the unit cost of production. Now why is the unit cost of production important? It's important because A, it's the basis of farm revenue changes and B, there's an important dynamic effect that when the unit cost of production goes down, the general market price will go down and you need to incorporate that in order to see what the farm level benefits are. So the price effects of changes in the unit cost of production are being modeled via our own global rice trade model that's been developed by Sam Mohanty. I won't go into the specifics of the model but I'll just illustrate a few aspects quickly. The model itself then has elements of the cost of production that are used in terms of determining the supply for a given country and basically the changes in the unit cost of production will be shocked into this element of the national rice supply. The national rice supply will then be shocked into the equilibrium of supply and demand and trade for an individual country which then will be equilibrated across the different countries in the model. And so the model includes these 18 countries and most of the countries are sub-nationally disaggregated so we actually get sub-national price effects as a result of the change in the unit cost of production from adoption of the technologies. So once we have the change in the unit cost of production and we have the data on the price effect we can then look at what portion of production is occurring by those under the poverty line and what their general characteristics are compared with average producers, particularly in the area of rice that they have cultivated because that will determine what kinds of farm level benefits they can receive from the technologies considering the equilibrium price response. So we have data sets on the spatial distribution of poverty. We can use that to approximate the number of poor in the different agroecologies and based on that we can look at the proportion of benefits occurring to those poor households when we take into account their differences particularly in production area. Then on the consumer side it's also very important to look at benefits to the poor. Through the trade model we get these price effects and we know that poor consumers spend very high proportions of their income on rice in Asia. They spend about 50% of their food income on rice since about 25% of their entire household income for the population surviving on less than $1.25 per day in purchasing car parity terms. So that's also an extremely important pathway to impact for the poor. So this was a simple run that we did before looking at an aggregate contribution to productivity and rice supply. It's a very modest set of assumptions we used. We basically said that we will continue productivity contributions at the same level as we've been documented to do historically up until from 1960 to 1998 that we would continue the same level of contribution. With that same level of contribution you end up with these kinds of benefits. It was basically a net contribution of 15 kilograms of rice on average per hectare per year for Asia. With that kind of a contribution it's about 8.5% over 25 years. You end up with a price reduction that ranges from 9% to 18%. The price reduction exceeds the supply increase. So that price reduction then considering how much the poor actually spend on rice it has massive benefits for poor consumers. So if you look here that kind of a price reduction over 25 years would lift 125 million people above the $1.25 a day poverty line if you count the reduction in their expenditure on rice as income. If you look at the proportion of the population that's food insecure if that expenditure savings were used to purchase rice to close the caloric gap potentially 62 million people could be lifted out of hunger. That's a big assumption but it's a possibility. So taking into account the consumer side is very important. Also when you see that these price effects are so pronounced it's important that we consider that when we come back to the producer benefits. So economic and poverty benefits are not the only benefits we're looking at. We also will look at environmental benefits. In particular when you have such a big price response you get an area response. So through the productivity enhancing technologies you actually have less area under rice. And part of that decline in rice area is also an avoidance of expansion. So given that there's a lot of land pressure in Asia and competing land uses some alleviation of that pressure on land we'll save natural land cover such as tropical forests. So taking into account the environmental benefits of that savings of natural ecosystems is very important and something that we're trying to reflect here. The spatial disaggregation of the model in terms of the supply response can help because then we can look at where the contraction in area happens relative to the proportion of safe forest cover. We will also then look at the more direct environmental benefits such as Boss mentioned for AWD in terms of the reduction of greenhouse gas emissions. And we will supplement this with analysis of the health benefits from our biophortification work and potentially some work to reduce the glycemic index of rice. And then we will put together all these different data on expected impact by different impact sources and we'll compare it with our patterns of resource allocation to look at where are areas that are in need of additional investment in order to realize much greater impact. So those are the methodological steps going forward. I didn't want to go into great level detail because this is the last presentation of the day and I want to leave some time for discussion. I'll now present some very preliminary results based on sources of yield gains using the parameter estimates of the sort that Boss was presenting. Now I have to of course give a number of caveats because these are very preliminary. These are not final estimates. We are still in the process of refining and reviewing our assumptions in these sheets. We are though looking at this as an opportunity for feedback so we do want your feedback. Your feedback still can affect actually what we're estimating given that we are still in that process of adjustments. This what I will present is only on the sources of yield gains. So it's a subset of all the research solutions. It only includes those that we have more confidence in our estimates and it doesn't include a whole array of technologies such as those that increase efficiency, those that are quality oriented, post harvest technologies, policy oriented research, etc. Those are incorporated in our priority setting exercise but they will not be incorporated in what's presented right now. This does not take into account the actual effect on the Unicostar production. It's solely based on yield and it's only using the lower bound parameters from scientists on adoption.