 and these workshops and think through these topics. And really excited to have him kick off this morning. Thanks very much, Beth. It's great to be here. For those of you who haven't met already, I work as a consultant in a group team called the Capital Science and Policy Practice at Willis Towers Watson, which is a global risk advisory company. But my background is I'm a volcanologist. So my knowledge of the science of floods is not very good. But if you want to know about erupting volcanoes, then I'm the guy to come and see. And but most recently, the last 15 years or so, I've been working in the developing parametric insurance solutions for developing countries on the whole. And so I'm going to talk a little bit today about that and the relevance of Earth observation data. I'm not going to talk a lot about flooding, to be honest. I do have a couple of slides which are specific on flooding. But I'm going to give you a feel for why I think the time is ripe right now for moving things forward. There was a lot of potential end users out there, to David's point from early yesterday. And I'm going to try and give you a flavor of who those are and why those end users are there, and also some examples of where we've put parametric insurance solutions in place. And then just kind of end with a few kind of key points, which I think will be good to take into the rest of the day. So what's the global setting for natural disaster risk management and risk financing? There's certainly a hugely increased political profile of natural disaster risk. I think you're all very well aware of that. Climate risk in particular. And there's a need, therefore, for practical tools to manage the consequences through building resilience. And insurance is part of that because the insurance industry does this as its day-to-day job. It's not the whole solution, but insurance as a discipline and as an institutional form that deals with risk, analyzes it, and can put a price on it, which is extremely important. It's not necessarily the answer to all your money questions, by the way, but it does bring risk into the real economic and financial sphere, which is critical. Example is the global risk outlook from the World Economic Forum. This is the, I'll show you on the next slide, the 2018 version, where environmental risks, the three furthest into the top right corner, which are both the most impactful and the most likely. So the biggest global risks are all environmental risks. So definitely high on the OECD kind of agenda. Last week in Barilochi, there was the first ever insurance forum for the G20. Argentina are the chair of the G20 this year. And so, you know, and various other places where it's coming to the higher and higher up the agenda. And this is both in emerging, both in developed economies, but also in emerging and developing economies. And in the latter, the sustainable development goals. As David said yesterday, Sendai framework, the Paris Agreement, the first time in the UNFCCC process, actually kind of, I felt the big thing in Paris was the identification that actually, we're talking about risk management in the climate space and what are the tools, insurances obviously mentioned specifically, but there's much more to it than that. And then there's this thing called Insure Resilience, which is a G7 initiative under the German presidency, the chairmanship in 2014-15. And I'll say a little bit more about that as well. This is just the world economic forums, global risk outlook. And you see the top right is the really big global events or the green ones are environmental ones. So you can see the top right corner is the, or the top right quadrant is the area of real interest from a global perspective. And you can see all five of the environmental risks that they capture are in that quadrant. You'll all be very, very well aware of the Thai floods in 2011, that was another wake-up call, I think, more from a commercial perspective of the interruption to global supply chains. So all of this is kind of feeding into the narrative that climate risk in particular is becoming more and more real, both on the political side and on the commercial side. We've also developed over the past 30 years or so a massively increased ability to assess and understand risk. You know, we talked about this, others talked about it a lot yesterday. And from the insurance perspective, that's captured in the catastrophe risk modeling world, which starts in the late 80s coming out of academic science and has been now absolutely completely mainstreamed into the global, certainly the re-insurance market. So the global, the ultimate risk takers of catastrophe risk, it's completely mainstreamed. Some catastrophe modeling is done for every single deal of consequence on a global scale, and even at the insurance level, certainly for the big markets, it's a normal day-to-day tool that's in use. What this has done has led to much greater stability and efficient capital deployment in the global economic system, can manage big risk events much, much more effectively and efficiently than it was before. And it's also, you know, it's opened up a new investment class. So insurance-linked securities have been growing year-on-year since 2008. They started before that. The collapse of Lehman Brothers actually put an interruption to it because Lehman were involved in almost all of the cap-on deals that were in place in 2008, guaranteeing the return, actually not to do with the risk side. The primary risk side. But ever since then, it's been growing and it's still a pretty small asset class from a global perspective, but definitely growing. And that's an indication that risk as a commodity is kind of more, much more accessible, if you like. The deployment of cap models, I think, is going to expand, and it could do that fairly quickly. The TCFD, the Task Force on Climate Finance, Climate-Related Financial Disclosures, which is a financial stability board, which is a G20 body, which was set up after the global financial crisis in 2010, I think. The TCFD is very active, it's chaired by Mark Carney, and their outputs are starting to feed into regulation of the asset management and banking sectors, for example. They're starting to put tools in place that the insurance industry has been using for 20 years, also, into a much broader range of economic activities. And we see this particularly quickly in happening in Europe, but it's also happening in the US as well, and I think will grow in its import over the coming years. Another interesting thing, this is a relatively recent development, or certainly recent to me, is that I've come across it. The force-major concept in contract law is starting to be coming under a lot of questions, particularly related to climate risk, and this is all part of the overall, well, wait a minute, we can quantify this to some extent, so we need to, we can't just use it, say an act of God has gone the way, so, and I think, again, that along with the TCFD and other things is gonna really have an impact. So I think the understanding of risk and the quantifying of risk, particularly climate risk, is going to have a lot more end users in the coming years, and a lot of those are big financial actors. This is just to illustrate the role that cat models have played in insurance, in global risk pricing, let's say, and this is an index which a guy happened to do every year, it's basically a rate online. So a price index for global cat risk, if you like, you can see both the volatility and the, and most recently the pricing has been falling, and this is independent effectively of insured losses. So it's not because we're having less insured losses, we all know that we're not, and certainly uninsured losses are growing even more quickly, but so this is a result of us understanding and deploying capital more effectively. So what about in the development context? Well, I think we've, disaster risk in a pure development context has been, talked about quite a lot, 29 billion a year in the 77 poorest countries in the world. I think this was 2016 numbers, I believe, and of which international aid, so humanitarian support generally is about 8%, insurance about 3%. So from an insurance perspective, but also the tools and services that go around insurance have a very low penetration rate. And the thing that I've always tried to convey when I'm talking to ministers of finance in developing world governments is that just because we've identified a risk doesn't mean that we've created it. It's there already, it's having some impact on your ongoing economy. It's somebody is paying for that risk. And so when they compare the price of an insurance policy to what they're paying now, which they see as zero, it's always gonna look expensive. But actually, if you look at the underlying impacts on development on individuals, et cetera, then it can be incredibly cost effective to manage it more effectively. The obvious things that we see in the developed world as well or even more to the fore in the developing world, financial selling of assets in Africa, I'll talk a little bit about African risk capacity later, but one of the big driving forces for ARC is not because insurance in and of itself is a great thing, but if you can get in after a drought before people start selling their assets, then their economic, their fall down, the economic ladder is massively reduced. So there's a four to five times benefit, we think, and maybe in some situations even higher than that of acting quickly and more quickly than the standard humanitarian response. So, but we all see, I've spent quite a lot of time in Haiti over the years and the level of exposure to natural disasters there is huge, we all know that, but there's, until recently, there haven't been very much of a link made between that and the very slow economic development that's going on in Haiti, but to me, they're intimately linked. So, but then there's also, I think over the last few years and probably less than five years, this kind of thinking has come actually into the humanitarian space in a very compelling way in terms of both understanding the causes of humanitarian disasters, but also responding more quickly to them. And for those of you who are not aware and are interested in this space, I very much recommend a very accessible book called Dull Disasters by Daniel Clark, who was at the World Bank for a long time, now works in the British government, and Stefan Durkon, who was the chief economist at Difford for in the UK for many years, a professor at Oxford, very, very, very accessible, very readable. And basically it's saying we need better planning for post-disaster action, we need evidence-based decision-making, and we need financing, but also the appropriate responsibilities are assigned between the different stakeholders. So it shouldn't be governments de facto are liable for everything or donors in the humanitarian space that we have to sit down and think rationally before these events happen, how the response and who's responsible for which elements of the response. This is very much what we're trying to do with ARC. It's starting to build the kind of infrastructure, if you like, that will enable this kind of approach. It's not gonna work all the time, and it's certainly not 100% of the answer, but when you have somebody like Stefan who's worked his entire very illustrious career, academic career, and applied side as well, I think it's worth listening to. And then there's the bigger global development policy context, and David spoke a little bit about this. This slide is probably a year and a half old now. So it was, I didn't update it, but I mean, these four 2015 and early 2016 international agreements and events really have shaped probably the next 25 years of potentially of action, maybe 25 years is too long, but certainly the next 15 years. Sendai, Addis Ababa action agenda, and then that feeds into the, that's the financing part of the Sustainable Development Goals, Paris Agreement, and the World Humanitarian Summit, particularly on the humanitarian side. I'm gonna skip that. I think you're probably all aware of the Sustainable Development Goals, but almost all of them have some risk element to them. And this is just a highlight. You don't need to take in the slide, but there's something called the Ensure Resilience Global Partnership, which is a German government led initiative which came out of the G7, Ensure Resilience Initiative. And then there's something called the London Centre for Global Disaster Protection, which is a DFID funded program. Both of these have basically sprung up in the last year and both of them have significant funding and are gonna be supporting development world efforts in this particular space. So those of you who are not aware of those two initiatives, I very much point you to get to know them a little bit and I'm happy to provide introductions as well to the people who are running those two programs. The other thing I wanted to, before we kind of move on to the next section, I just wanted to kind of highlight that there is much more engagement in the developing and emerging sovereign and sub-sovereign space on these issues, and particularly on the financial aspects of them or the quantitative aspects of them than has been growing significantly over the last 10 to 15 years. I think disaster risk management, when I first got into the space 25 years ago, was all about kind of response and response actions. It started to go into hazard assessment, but it's now, I think there's a much better understanding of the full risk assessment, the full economic consequences, et cetera. I personally have talked to probably 35 Ministers of Finance or Ministers of Finance for 35 countries, I would say, over the last 12 years, 13 years maybe. And all in developing world countries. And so there is a hook into almost every developing world country which can be expanded fairly quickly to be talking about what we're here to talk about. That includes both not only insurance as a kind of a tool, but also a assessment of disaster risk, ownership of risk, and then different risk management tools. And the sharing of that risk between individuals, communities, local, regional, government, et cetera. At least the concepts of those are starting to be understood actually implementing, kind of joined up risk management plans across the whole country and including all of those different elements is far from being mainstreamed, but I think the understanding, at least those conversations are starting. I'm gonna skip over that last one, but just a couple of practical examples. ARC I've talked about a little bit already is covering a big chunk of the kind of theory of identifying risks, planning, financing, decision processes, and then delivery and beneficiaries. That kind of, this is more in the humanitarian space. And also the Red Cross, Red Crescent, has been doing some really interesting stuff on forecast-based financing. And similarly, triggering even pre-disaster action. And if you'd gone into anybody in the Lloyds market or anywhere else in the insurance space five years ago and said we want to do an insurance policy which triggers before an event happened, you would have been laughed at a town. But we are having those conversations now. So, the reason why we're having those conversations is because of something called parametric insurance. And I'm gonna kind of, I'll skip through this fairly quickly, obviously the slides will be available, I presume, if they're not. Then, so you can kind of delve back in and I'm very happy to answer questions. But I did want to spend a little bit of time on it because I was at a volcano conference a couple of weeks ago and talking about parametric insurance. And while volcano colleagues, very technical people, all of them didn't have a clue what I was talking about. So, I've assumed that you guys, most of you guys don't have much of an idea what parametric insurance is. But I think it's important to understand. So, traditional insurance, you wait until a disaster happens and then you go in and work out what the losses are. You argue about what the losses are actually generally. If you're a homeowner and you have a loss of justice you're not gonna come to the same conclusion about what your losses were, there's the loss of justice. So, it's not quite cut and dried but that's generally the way that indemnity insurance has developed. And it's been around for more than 300 years and it's certainly served the first industrial revolution extremely well. I haven't put the plot here but if you look at the plot of UK GDP growth in the late 18th, 30th, 19th century against insurance penetration, they're tracking exactly. A mill owner would not invest in a mill if they didn't have fire insurance. But in the modern, the third and fourth industrial revolutions we don't see that, it's not happening. But what we have seen is a start, the development of this new form of insurance it started in the weather risk markets in the US for energy companies mainly who wanted to be able to hedge out their additional costs for either cooling in very hot days or heating in a number of cold days more than a normal number of cold days in a row for example. So, there's a pretty active weather derivatives market in the US and increasingly in the European Union as well. Little bit in Japan where there are about 20 standardized indices these run off ground stations on the whole in big cities and they can be bought, they're traded on a daily basis. So, it's mainly through the Chicago mercantile exchange but so weather risk has been traded for probably 20 years and is pretty much mainstream but in a fairly small niche market. What we started to do in the about 13 years ago was look to see how these same principles could be applied to bigger catastrophe events, hurricanes and earthquakes were the two that we kind of started with and I worked with the World Bank in 2000, started working with them in 2004 and five to set up the Caribbean Catastrophe Risk Insurance Facility which used parametric insurance to insure against for the governments to insure against hurricanes and earthquakes. So payments for a parametric insurance policy are made on an agreed scale based on the movement of an index. The index is selected so that it's a good proxy for loss damage or impact. Basically it can be in its form most similar to regular insurance, it would be a proxy for loss as you would adjust that loss but it can be very much kind of looser impacts which makes it a really interesting tool for in the development space. So as a very simple example there, less than a certain amount of rain falls over a certain amount of time than a contract triggers. And what you need is that real time monitoring of that index and as I'll come on to say later, Earth observation is a very, very obvious source for that information, particularly in the developing world. And if the trigger is threshold is met then the payouts released. There's no second guessing it, there's no going in afterwards and seeing if the trigger actually did what it, or the event did what we thought it would do on the ground. The contract is based on the index itself and that means that the payout can go up very quickly. So that's the other huge benefit of parametric insurance is that in the Caribbean for example, the governments have received their payouts and 130 million plus of payouts so far in the 10 years, 11 years that it's been operating, all of them have flowed within 14 days. Nobody gets an insurance claim sold within three months generally after a big natural catastrophe. In the Caribbean, Sean knows this very well, there's a big hurricane in the Caribbean, the regular loss adjustment community is about 10 people across the English-speaking Caribbean. 10 people after a NACCAT in any of the islands, let alone one that's hit multiple islands, is the work of 100 plus. So you just cannot adjust claims very quickly and that goes for everywhere in the world actually. If you look at Chile after the earthquake in 2012, was it, there were more than half of the claims were outstanding a year later. I think Maria, I'm not sure of the actual numbers but I know that the Maria actual loss estimates are moving all the time because the adjustment process is still ongoing and this is, we're more than a year past Maria now. So, and then just add the indices are almost always measured by or at least verified by a third party. I say almost always because I have a case right at the end where we have managed to put together deals where the measurement is not totally external to the client but that's the norm is that to take out the moral hazard of pouring water into a rain gauge or something, you wanna have it, the measurement made by a third party and public domain sources are obviously perfect really. I mean, they're ideal because they provide fantastic transparency. So it's very difficult for either the markets, the risk takers or the clients to argue when you say, well, I'm just gonna go to this website and take that data from the USGS, from NASA, from NOAA, whoever it is. So I've run through some of these already. The parametric is very good for disaster response financing because of the speed and as I said earlier, money quickly is much, much more valuable in all disaster response, but particularly where you avoid sales of assets like a slow onset events like droughts but also in hurricanes as well. There's very good evidence that injection of liquidity and ability to both the government to spend money and also the communities to spend money maintains the markets and aids recovery hugely. The contract settlement is very clean as well. It takes a lot of the political angle out of insurance and off the record, I have many stories of where kind of insurance is used as a political tool and with parametric, you can eliminate that effectively. And so the Minister of Finance has to sign on the dotted line for a contract and then it's a contract and it's on an index which is verified or measured by a third party and there's not much wiggle room there. And as I said, there are some now being, we're not thinking about doing forecast-based financing and triggering on either a forecast or certainly the obvious case is upstream flooding, which is gonna inevitably move downstream and getting a week advance in the big river basins, for example, in Niger River Basin, those floods take two weeks to move down the basin. So that really is, there's great potential there and for no regrets investments really. The other thing about parametric insurance, I just wanna mention and I'll give you some examples of this is that it's very bespoke. So it's really pretty flexible. If you're buying the insurance for your house, you basically buy insurance for your house and it's the value of your house. Governments ensuring themselves, what value do they want to insure from? Yes, they can insure all of their assets but most of the governments I've worked with don't have any clue about where their assets are, let alone what they're worth. So whereas with parametric insurance, you can model out what the impacts are gonna be across a broad range of government activities including property loss but also loss of tax revenue, all of those sorts of things can be looked into and then you can find a proxy for those in an index and they can buy the index and they don't have to buy 100% of the index, they can buy relatively small fraction of that index but it will respond in the way that they need it to. So if there's a bigger disaster where they'll need more money, they'll get more of a payout but it may not be anywhere near all the money that they need so parametric insurance is unique in that way that we can effectively, we can effectively tune it to exactly the requirements of the client. I'm gonna skip over these slides a little bit just in the interest of time but those are the two end members, a simple parametric which I just described right at the front what we would call a binary so it meets the trigger threshold of an index and it pays and indemnity is obviously the loss adjustment side but there are actually a couple of additional forms of parametric insurance in the middle getting increasing in complexity but also reducing the effective difference between I've called it basis risk here but the effective difference between what a parametric would pay out and what an equivalent indemnity policy would pay out if you look at a kind of a normal property basis for the insurance and the modeled loss basis is effectively we use we use a catastrophe risk model it's just locked in advance and we generate the pricing from using that catastrophe risk model with historical or stochastic storms for example for cyclones and then when the real storm happens we run the track data from the NHC through that same model and what comes out as the modeled loss is what we pay against so that's quite equivalent and actually from a contractual perspective effectively what we're saying is that we're gonna agree to not go through the whole adjustment process we're gonna accept that the model is gonna give us a reasonable estimate of the loss and we agree to pay on the basis of that reasonable estimate so we can get quite sophisticated with those kind of models but we also have found that explaining a modeled loss versus a parametric index for example can be quite challenging in the development context so those are just the takeaways I've talked about those already so a few practical examples of earth observation in particular being used for parametric risk transfer in the development context and these are all projects that I've personally worked on there are others but I think I've covered all of the main forms that the parametric insurance has been used in again I'm not gonna go into detail I did have slides for all of these different schemes but the slide that was getting pretty heavy so we've used Prim for extreme rainfall as probably the outside of earthquakes and hurricanes which I haven't covered here because they're fairly mainstream I would say extreme rainfall was always the first thing that I got asked about after we launched the CRIF every single minister of finance everybody in the Caribbean said well what about rainfall? Hurricane models so you guys probably know do not capture rainfall very well and so Trim was pretty much all there was ground station density is nowhere near good enough and doesn't capture hurricane rainfall very well in a way so we used Trim in the Caribbean but this particular thing we did was a micro insurance program in Haiti for Foncose which is the biggest MFI in Haiti has women micro entrepreneurs who are the repayment rate for Foncose is absolutely incredible it's 99. something percent I mean these women are incredibly good clients of the micro finance institutions the one time that they can't handle is when they get blown away by a hurricane washed away by a flood shaken down by an earthquake so this was a program I was working on before they're actually right before the earthquake and we tested it with the earthquake as though the clients had insurance in place even though they actually didn't but we had the whole thing pretty much set up so they were given payouts or had loan forgiveness as though the insurance was in place and building on that experience we put together a program and it was in place for a couple of years and I'm not gonna go into the detail but I mean it worked as advertised the main problem was that it was because the risk is so high in Haiti it was just too expensive to add it onto the loan the lender's cost was they just couldn't bear that cost we got some money to pay for it for a few years but then when that dried up it just wasn't sustainable and this goes to well who owns that risk in Haiti the development actors have put millions and millions tens of millions, hundreds of millions of dollars into Haiti year after year after year in the development context so they're paying for it to support this would be a heck of a lot better use of that, a small chunk of that money to be honest and supporting the economic engine of Haiti to get past events that they can't control I think is a pretty good model but so that's one example we did NDVI for dairy farmers in the Dominican Republic they were worried very particularly about a loss of production for when forage levels dropped and so we got it something for them and then we've done a WRSI based drought modeling in ARC and Elka here at the back there and there's a few great posters on the flood stuff that we're doing but ARC started out as the first policies were droughts using WRSI based on satellite rainfall so those are kind of rainfall related for both high rainfall and low rainfall and using both direct satellite rainfall estimates and greenness in the seas a couple of a few that are kind of on the drawing board volcanoes, we are in the middle of doing a project with the World Bank on just exploring what's possible for volcanoes volcanoes are pretty complicated there isn't any one signal that is going to tell you that it's about to erupt there are a multitude of signals many of which can be obtained from space and so give you the kind of coverage and kind of uniformity of data that one would need these are interference rings from INSAR on a volcano in Alaska and so ground deformation has definitely increased in its ability to be routinely monitored but is not there yet and then other signals are being starting to be processed routinely on a global level so there's definitely potential there I think we're not there yet but I hope that, David we've spoken about this you know as a potential end user parametric insurance I think can drive the kind of right developments in some of these spaces wildfires is another obvious one I don't think there's been a wildfire parametric yet but it's certainly something that I think we're starting to get to the point where it's we have enough of a history of what it looks like and then the routine monitoring which is very well developed now of course so there's plenty of wildfire and indemnity insurance but I don't know that it's been done on a parametric basis yet and then another project I'm involved in is ensuring coral reefs and there's been a lot of interest growing interest in the ocean risk space it's not something that the traditional indemnity markets have really paid very much attention to there being those who owns the reef whose value is from it but cleaning up the reef very quickly after a hurricane is incredibly important so just having a relatively modest policy which triggers on a hurricane index which can generate some funding to have a clean up team it can be really valuable and then as I was looking for some stuff yesterday I just came across this headline which I think was after the Hawaii actually it was after Florence but in the context of what happened at Kilauea which I just thought was kind of amusing that volcanoes even though incredibly challenging to model is actually covered in on Hawaii insurance policies but that flood is not I'm gonna skip that, that's in the Philippines I'm seeing your two minutes I'm gonna just very quickly run through actually I'm gonna skip the CRIF one but I've said a little bit about that it's a potentially really good vehicle it now covers both the Caribbean and Central America and a really good vehicle for channeling some of the risk quantification as well as just insurance it's called an insurance facility but it's really talking to a much broader range of issues our arc I could talk for hours and hours about arc and would love to but I think the really important thing about arc is that it's really taken the kind of things that Daniel Clark and Stefan Durkan talk about in their book about the preparedness, tying the preparedness to the financing, having the early warning and the insurance product being linked together so that there's continuity between the messages that the governments are getting about risk about how it can be managed about the importance of preparedness in arc the money the government has to come up with a detailed budget on how they're gonna spend the payout that detailed budget and action plan is audited so we're holding their feet absolutely to the fire in terms of how they're using that to make sure that they're using it to the not only to avoid corruption et cetera but also that it's going to the things that where it has the most value because it's arriving very early what are the investments that we can make right now to reduce the overall impacts? Then one flood example before my I think is what's gonna be my final slide is the flood example, it's not in the, it's in the US but I just thought it might be interesting I'm not gonna go into the details but it was a post Harvey Houston thing that came up it was from a Levy Improvement District which those of you who know the US well these are tax raising entities who have who are mandated to look after the mainly the engineer aspects of Levy systems and so what this particular Levy district which is an extremely good Levy district it's a very, very good reputation and but they had flooding outside of the Levy because they couldn't pump the water high enough quickly enough during Harvey because two things one the rain rate was too high and secondly the water level was too high in the river so those two things combined were what they wanted to be covered for and so their engineers came up with a table of okay this is where we're gonna be in trouble these are the conditions of gauge heights and rainfall rate and so we put together a parametric deal for them and they ultimately they didn't buy it unfortunately but it was very interesting to go through that process and have them think about what it would take and also what they would spend that money on and they said yeah absolutely if we'd had a relatively modest amount of money a couple of million dollars to give out to the community so it wasn't insurance for the individual householders it was an ability to generate some funds that the Levy district could spend in the immediate aftermath of this to help the community and interestingly I'm a very quantitative guy when we took this deal to the international markets we went to a few markets and the pricing differential that we got even though it's US, the data's pretty decent we got rainfall, long history rainfall flood height not so long but the gauge I think was 20, almost 20 years worth of gauge data, so pretty decent data and the differential between the pricing we got was three times differential between two kind of mainstream re-insurance markets it was a fairly small deal but even so the amount of kind of wiggle room when a broker like Willis goes into the room and says okay this is the risk it's all quantified for you it's a parametric, what price can you write that at? There's still a huge, huge difference so there's a lot of stuff around the edges which is important to pay attention to as well and then just my final slide just a few kind of out there questions which I hope we can address the rest of the morning in the breakouts to some extent The one thing I always highlight to the EO community in particular is that to do insurance you need to have a probability of events happening and you basically build that up or you need to have some way of knowing of rooting that in history you can model it out, you can do simulations but when you're talking to a client if you can't point to some historical events where which would or wouldn't have happened would or wouldn't have triggered then they won't believe what comes out of a model so yes having the best real-time data is great but if you can't have some kind of representation of what that data is now capturing for historical events then it's very difficult to use that best data if you like in an insurance setting so that's a flag, I'm going to... Yeah, so and I think the GPM and David I've mentioned this to you I think and other people have as well the GPM is fantastic, you know we know that the real-time stuff is great we've done the comparisons with Trim and it's hugely better we can't use it yet for rainfall parametric because we haven't got the history so we are literally, we are ready to go with three deals I know of in the developing development space with as soon as that history comes out within a couple of weeks we'll be able to switch it over so just and I know that you know that but it's really important in this setting to have that reliability of data so and by the way putting that together is absolutely incredible and the word that there's a poster there by EAR by John and co-authors on the flood modeling that the AAR have done for ARC the footprint modeling our biggest thing on our to-do list when we were looking for somebody to do this for us was we needed to have a history we needed to be able to go to our African governments and say we have these flood footprints going back to 2000 and where is John? 2007? 98? So you know to have that I think we said 2001 to start with I think because that's where RFE2 goes back to but yeah, further the better and that's so I'll, Elka can probably answer that question better it does contribute but because we're using flood footprint not from a model basis but from an actual what you're seeing then that would, I think that would contribute to understanding how that flood footprint got there but we need, we wanna see that flood footprint and know what context that flood footprint happened in to be able to demonstrate to go back to a big flood in 2001 for example and say look, this is what the tool that will be working in real time this is what it came up with when that flood happened that's really, really important and so it's that, the second point I wanna highlight is that I really think that EO data has massive potential because of it's global it's global footprint, it's uniformity of coverage it's ability to be objectively processed or all of those things standardized delivery and NASA, I have to say NASA not just because they're sponsoring this but and NOAA as well, I mean they're really the availability of those data sets the incredible sciences behind them is absolutely fantastic and I just, without that it would have been very difficult to do what most of what we've done particularly outside of the earthquake and hurricane space so more of that and definitely you've got an end user which I think is gonna expand potentially very quickly David and potentially also beyond the developing world I think that some of the stuff we've been doing in the developing world actually has huge application in the developed world as well and then the last thing was the forecast based financing which I think has a lot of potential and I think we'll see a lot more of those kind of deals happening and initiatives in the next few years thank you very much for your attention. Thank you so much Simon, what a clear description of many use cases that we can really I think dig into as a community and think about how to move forward. We're a little over time but we can take some questions for Simon before moving to breakouts, lots of hands. So that's a good question. I think depends on the sophistication of the parametric instrument that we're trying to develop in some concepts context, reanalysis is really important. An example would be the best track hurricane data set which is a reanalysis and I think without that we would be, the parametric insurance for hurricane risk would be a lot less used and a lot less useful and so and then if you go to the model lost parametric then we're using a lot of the same tools that the CAT modelers are using anyway. So we just I think concentrating on the hazard side of it rather than on the ultimate risk side. So we wanna know what the ultimate losses are gonna be or have a good estimate of them so that we know that they proxy what we're trying to cover but really the concentration is on getting the hazard probabilities right because that's what the, so it's something I didn't really emphasize that the global risk takers love this stuff because the probabilities and therefore the risks that they're taking is really down to the hazard and so and that's an area they feel comfortable with if they don't know about it then they can find out about it. Dealing with stochastic hurricane track sets and CAT models is something that they do and they're very comfortable with but particularly at the hazard end of it. So and if you can define the risk more tightly then you're gonna get better pricing from the market. So African drought risk for a country in the Sahal if you'd taken that even to the most adventurous underwriter 10, 15 years ago they would have said, no, we're good, thanks, right? Just because they wouldn't have known how to quantify it not necessarily because it was a high risk just they didn't know how high it was. We've now we're now transferring African drought risk into the international markets. We have 24 markets on the ARC re-insurance program. It goes to 40, 45 re-insurers effectively. So we're sampling the entire market and we're getting the rate we're getting is absolutely incredible in terms of the small margin over the expected loss cost. So it's an incredibly efficient way of transferring that climate risk. But yeah, I think the re-analysis and the modeling is very important especially in the helps to build the understanding on the ultimate risk takers. So you can start to do things that are really cost effectively. More burning questions, yeah, Guy? No, I think that's a really good point and I think most people you taught to on the insurance side would think that it was a given. So I think that's a really useful point and I'm happy to act as the point person to make sure that we join our voices to you guys in that. I should quickly mention that there's something called the insurance development forum which my boss at Willis has helped to put together and that's basically an initiative between the World Bank, the UN system and the global insurance and re-insurance industry and it includes all of the industry. It's not just a few key players, it's everybody. And they have actually just appointed a secretary general anyway, they now have a secretariat after kind of getting into spin up mode over the last 18 months, two years. And so this is something that, this is kind of thing that they would definitely be interested in taking up and being an advocate for. So we're starting to develop some of the institutional forms rather than it being coming from individual companies or individuals, it can be from the industry as a whole. All right, last question and then I'll give you. Yeah, and I just let me mention that the, I mean, what I think what you guys, NASA did with the trim was just incredible. You're putting out data until we were really worried about that because it would have, if you'd switched it off, we would have had to cancel all of the policies that we had based on trim. But the fact is that you, we had the time to make the necessary arrangements. So yeah, that was a great example of something that, yeah, it's gonna come to the end of its life, but you, yeah, so, but a good point. All right, last comment or question and then we'll move to breakouts. I think six years is pretty difficult from a global perspective. I think you could probably start to apply it in particular cases where you could build a history of, you know, in a particular area that was now part of that global monitoring. You could build a picture of what history would look like. So I certainly don't want to stifle innovation, absolutely not. I was just making the point that until we can deploy, there's a barrier to deploying a new tool immediately and that barrier is that what that tool, how that would have behaved or what it's trying to capture. I had a conversation with somebody yesterday about, well, what about if, you know, new technology could, you know, could say whether a pixel was flooded, you know, a 10 meter pixel was flooded in the middle of a city now, would that be useful? And the fact is that you, you know, if you put your mind to it, we could probably put together a history of when that pixel was flooded from various sources. It wouldn't be uniform, but if we could do that, then we can use that data because we can say, well, now we can measure that down to that level. We couldn't before, but we've, you know, we've put together a history. And then, you know, the point I just made about the, you know, the price, different pricing in the markets, that's very much about the story that you tell around the data. And the easiest story to tell is, well, here's 30 years of history from this bird. And, you know, that's what we're going to use in real time. And, you know, so, and that's a very easy story to tell. But I think we've gotten better at telling stories around, you know, putting together data sets, which are a little bit more complicated and multiple, multiple sources as well. And I personally feel pretty confident that, you know, you can start to do that, but it's very much on a case-by-case basis. And some of the, you know, one of the drawbacks to that is that it does end up being kind of costly in a verticomas. And, you know, when some of these deals are fairly small, that cost is difficult to bear within the, you know, so as a kind of research challenge, yes, we could definitely do it, but a practical mainstreaming of these kind of tools into insurance, you have to be careful about those additional costs. Thank you so much. Okay, so we're running a little bit over, but may I suggest grabbing coffee and taking a quick, you know, five to 10 minute break on the way to your breakout session?