David Stevens - Group Chief Executive Officer Geraint Jones - Group Chief Financial Officer Cristina Nestares - Chief Executive Officer, UK Insurance Lorna Connelly - Head of UK Claims Andrew Rose - Chief Executive Officer, Compare.com Milena Mondini - Chief Executive Officer, Europe Insurance

Kamran Hossain - RBC Thomas Seidl - Bernstein Ravi Tanna - Goldman Sachs Arjan van Veen - UBS Andy Hughes - Macquarie Janet Demir - Morgan Stanley Dhruv Gahlaut - HSBC David Bracewell - Redburn Andreas van Embden - Peel Hunt Andrew Crean - Autonomous Greig Paterson - KBW

Good morning, welcome to Admiral's Half Year Results Presentation. For those of you who don't recognize the picture in the background, haven't been to Cardiff, that's the modest atrium in our reception in Cardiff. You are all very welcome to come and visit, come five minutes earlier and enjoy the plaque awards for staff behind the chair there, it was an interesting read. I'm going to start with an introduction, and then Geraint, our CFO, will dive into some of the more financial aspects of the business before Cristina and Lorna, respectively Head of UK Insurance and Head of UK Claims, will talk about some of the more operational aspects in the UK. And then Andrew Rose has joined us from across the water to talk about price comparison in general, but Compare.com in particular, before Milena from Rome talks about our International Insurance Operations. So a very brief few words of introduction from myself. It's been six months of continued growth with 14% more turnover, 13% more customers than we had a year ago. By the half year, we were just short of 5.5 million customers across the world, we are now over 5.5 million customers. Year of growth also in the underlying dividend, up 10% from 51p to 56p, reflecting our strong capital position. A year of some growth, modest growth in our profit, 1% up profit before tax, 3% earnings per share reflecting some continuing headwind from the Ogden change announced in March. When we did our full year 2016 results, we announced Ogden would cost us around £150 million post tax and that we are effectively taking around two-thirds of that in the 2016 results, so the other third carries into 2017 and beyond. So, different rates of growth, growth everywhere, but different rates of growth, steady in the UK, car insurance business up 7% year-on-year, particularly pleasing in the context of a hiatus in growth in the first three months of the year following our decision to increase rates quite substantially in December on the announcement of the Ogden review, and the fact that the market didn't follow until March. Much more dramatic growth in other parts of the business, 43% notably in our household customers insured in the last 12 months and 27% in the number of vehicles we insure internationally. So, to expand on that, over to Geraint.

Thanks David. Good morning everyone. I'm going to talk briefly through the Group results, then go along to the UK Insurance profits, and then talk about capital and the interim dividend. So first up, let's look at the Group profit. This is our usual chart, it shows the contributions to the overall result that come from the different parts of the Group. As David mentioned, the first half profit is £195 million, that's modestly higher than H1 of 2016. That includes an improved result in international insurance; loss was down to £10 million from £13 million, that's in the green, and notably that includes a much lower loss in our U.S. insurance operation. H1 also saw an improved price comparison result, that's in the light blue, you'll have to trust me on that, you can't see that I'm sure; again improved numbers there from the U.S. operations being a key reason for that change. Now, those positive moves were offset by higher Group or central expenses, which is in the yellow. As you can see from the text on the side, that number includes the debt interest, share scheme charges, our business development costs, and the key reasons for our higher number this year, higher share scheme charges, higher headcount, but also higher costs in business development notably from Admiral loans. Not forgetting of course, the UK that's in the dark blue. And as you can see, the UK insurance profit was broadly flat this year compared to the last half year. And the combination of all those factors gets you to a slightly higher H1 result compared to H1 2016. So, I'm going to take a few minutes now to talk about the UK Insurance profit. On the top of the slide, we show the UK Insurance profit, which now covers private car, household, and van over the past three half-year periods. And on the bottom, you can see the breakdown of the UK Car Insurance combined ratio, which is obviously critical to the profitability. Cristina is going to cover household shortly, but as you can see that business made another profit in H1, slightly above 2016, and that's despite adding more than 40% to the customer base, and we continue to be conservative in booking loss ratios on the household business. On car insurance, the H1 profit was basically in line with last year, so that's explained what's going on there, so you can see that one or two things in the text distort the 2016 result. We flagged last year that includes the connotation. And as we said back in March, we will have some residual hangover from the Ogden impact this year, and for the next three or so financial years. The impact for this half year is approximately £10 million that's in form of a higher current year loss ratio and lower profit commission. Probably the key factor is a lower reserve release this half year versus last half year. As you can see, it is a 21% premium, it is still substantial in relative terms, but had that percentage been more in line with 2016 and 2015, then obviously we'd have seen an increased UK Car Insurance profit. So, why wasn't it quite as high? Understanding the development of the projected ultimate loss ratios over the past few periods is part of that answer, so let's take a look at that. This chart should be at least half familiar to you, it is the projection of the ultimate loss ratios by accident year. This chart shows the position at the end of 2015 and obviously the brackets show the change in those projected ultimates compared to six months prior. What we tend to see over time is improvements in these initially cautious projections as the level of certainty over the outcome grows. So for example on this chart, you see 2012 accident year improving by 3 percentage points in six months. If we roll that on six months to the middle of 2016, we again see some pretty material improvements to some of the years. Then we come to the end of 2016 and actually here the position is largely stable, one or two small improvements, but don't forget that's in spite of the new Ogden discount rate minus 0.75% being reflected in these projected ultimates. Roll-on again to the middle of 2017, the current period, and after updating all our case reserves for the impact of Ogden and seeing that going through, pleasing to see generally stable outcomes in the projected ultimates, a small improvement on 2015. And here, we've added the – in the redline, the end of an era , maybe the last time we'll see on a comparable basis, Admiral against the market, an institution of the graph. As you can see, the assessment of the ultimate cost of the minus 0.75% Ogden rate remains unchanged and obviously these projections continue to reflect that rate. It is appropriate to say that there is prudence in these projections, and particularly in the most recent years. At this point, we don't give current year numbers you know, but certainly on an underwriting year basis, at least I would expect that 2017 will be better than 2016 on an ultimate basis. As we've said in the past, the size of Admiral's reserve release is driven by two factors. Firstly, the release of the initial margin built against those first projections, but also from the improvement we see in the projected ultimates over time. What we've seen in H1 is stability in those projected ultimates, which in light of Ogden I think is pretty positive and that stability leads to a lower albeit still substantial reserve release in H1, that's 21% of premiums. Let's take a look at the releases in a bit more detail. This usual top chart shows the release percentage, percentage of premiums focusing on the release on the original Admiral net share of the reserves over the past five half-year periods, and a couple of observations on the numbers. Firstly, the first half of H1 – first half of 2015 and 2016 benefited from the improvements in the projected ultimates that we talked about earlier, and so very large reserve releases at 29% of premiums. The second half of 2016 that was obviously heavily impacted by Ogden, basically no release at all. And finally in the first half of this year, this 21% is well above the long-term average which is somewhere around 15%, but it's lower 21% compared to H1 2015 and 2016. As is always the case, there continues to be a prudent and significant margin above the projected ultimates in the booked reserves. And as we usually say, we'd expect continued significant reserve releases if claims develop as we expect them to. Just to finish the story on releases, this bottom chart splits the total release into the element on Admiral's original net share of the reserves which is in the dark blue on the bottom. On the top, the light blue shows the release that comes from the portion of the reserve that was originally reinsured under quota-share contracts. And just a reminder that 2016 as you can see was impacted by the commutation of 2014 which adversely impacted that number, but no impact – no repeat of that impact this year, no commutation in the first half, mainly because of Ogden, but just important to note that we do continue to project profitability on all underwriting years. Change in topics to look now at capital and dividends and firstly capital. A pretty similar picture here to six months ago, with a strong solvency ratio at 214% which is largely in line, slightly higher than the position at the end of 2016. The only really significant movements in the half year were the economic profit offset by the interim dividend, so a much more normal six month period than the second half of 2016. Just to give an update on the calculation of the capital requirements and the internal model, the basis of this solvency capital requirement remains the standard formula plus the capital add-on, and the work on the internal model development continues. So as you can see, we have decided to increase the scope of the model beyond just UK motor insurance risk to also capture market risk. We believe that will lead to a more complete model of the most important risks to the Admiral Group. Now inevitably the work involved in building, developing, testing, validating and so on that bigger model has led to a delay in the expected application date, which we're now expecting to be in the latter part of 2018. As we said over the past few results announcements, we'll formally confirm our target solvency range once that internal model position is clear. Moving now to the interim dividend. As you can see the strong capital position has meant we've been able to move the interim 2017 dividend up by 10% compared to the interim 2016 payment, that's 56.0p per share up from 51.0p. Just to note, obviously we paid the 11.9% – 11.9p in additional return of surplus capital on top of the underlying dividend in H1 2016. 56.0p that represents a payout ratio of 98% of the first half earnings that's slightly higher again in the first half of 2016. We repeat the dividend policy on the right hand side of the slide and we reiterate that we'd expect dividends to be in the order of 90% to 95% of earnings for the foreseeable future. And also to repeat the point that we don't expect to talk about any potential return of surplus capital until we get the internal model certainty. Dividend dates are also in the bottom of the slide. To summarize the key messages from the first section; very pleasing to see continued growth across the Group, especially again in household and international insurance; Group profits were slightly ahead and improved results in international insurance and comparison; and we continue to report a very strong solvency position, even after a fairly full payout of earnings for the first half, 56.0p interim dividend being 10% ahead of interim 2016. That's it from me. I'll hand you over to Cristina and Lorna to discuss the UK in more detail. Thank you.

Good morning everybody. I want to give you an overview of our UK Insurance. On motor, Geraint has talked about the loss ratio and the overall results, and Lorna is going to cover specifically claims inflation. So I will talk briefly about changes in prices. On the household, I want to talk about the market, because it has gone through a period of change and also our results. And I will answer the question of why are you growing when other competitors are not growing? And finally, I want to talk about our new product which is van. So, when we look at changes in the motor market in terms of prices, we all know that since March and due to the changes in the Ogden rate, there has been continuous price increases in competitors. However, our strategy was quite different, because in mid-December last year, when it became obvious that there was going to be a change, we increased our prices significantly, that meant, as you can see on the graph on the right, that our times top decrease and also did our sales. We were surprised to see that competitors didn't follow and we for the first two months of the year we were selling less than expected. Now since March, competitors started putting prices up and we also continue doing so. Let's look at the household market. What we have seen in terms of prices is that for a number of years it has been quite stable. I think this is partly due to the fact that we have enjoyed a fairly benign weather. Now this is changing and we're seeing clear signs of prices going up in the first half of this year. We think that behind this it's the Flood Re levy, which was pressured on the expense ratio and also claims inflation, especially in escape of water. So what have we done in Admiral? Well, we put prices up in Admiral already in Q4 of last year as a reflection of the claims inflation that we were also experiencing. What we saw is that, since around February this year, prices in the market stared to go up and that meant again, you can see that our times top increased and that meant that also our sales did so. Now since 2010 we have continued putting prices up. So overall what we are seeing is, in the past nine months we have put prices up by about high single digits and we have been able to grow, because prices in the market have continued to increase. What I think is also interesting to talk about the sources of growth. On the graph you can see a very strong growth trajectory. We grew 43% year-on-year, but there are three key sources of where we get this growth. First, the growth of the price comparison market that continues to grow in this period and we are leading players, so when it grows the price comparison market, we're also able to grow ourselves. Secondly, the growth of our direct sales direct; when we say direct sales, we mean sellers that we acquired through direct marketing, so not price comparison. But it also means cross-sells that is selling to our own motor customer book. And finally, another important source of growth for the household book is the strong retention. In terms of the results we were quite pleased to see that both the expense ratio and loss ratio improved in this period. Particularly significant is the outperformance of the expense ratio versus the market. So, overall with the growth and the progress we have made, we have seen profits increase in the first half to £1.6 million. And then let's talk about van and the launch that we did in May this year. And it's interesting because I talked about this new product, but truly this is not new to is. We have had a broker in the van business since 1998, called Gladiator. So what we did this year is that we changed the model from being a broker to being an underwriter and that has started in May as I mentioned. So, by the time that we finish the broker, we had about 10% of the market and we had 175,000 vehicles on cover. So now that we have become an underwriter, the focus is – for this 12 months, it's migrating that book of business from the broker to the underwriters, so you should expect a similar level of a size of the book at the end of the year. We are going to focus on learning about the market, learning more about the claims experience and about pricing and for the future, we expect to achieve similar results to motor, especially on the expense ratio, Now before I finish, I want to talk about the customers, because delivering great service is at the core of the Admiral culture and therefore I was disappointed that we sent inaccurate information to our customers during April and May on our renewal letters. Now, all the letters that are issued today have correct information. But I think it's also important that we focus again about this great service that we try to give to our customers and I thought I'd talk to you about two examples on how we try to constantly reduce the level of complaints. You can see at the top there is a graph that shows the result of the complaints referred to the Ombudsman, and we can see that these results are better for Admiral than the average of the market. On the bottom, you can see a graph on the claims complaints that we have on the – sorry, on the claims complaints that we have versus new claims and you can see that period after period we are constantly increasing or improving this level from an already good level. And finally, we are proud to have won The Best UK Car Insurance Provider for four years in a row. Now I'm going to hand it to Lorna to talk more about claims inflation.

Good morning everybody. I'm pleased to be here today to talk to you about claims. Now I don't know about you, but it certainly seems to me that not a week goes by these days without picking up a paper and reading some headline about claims inflation. So we thought we would take the opportunity today to just touch on some of those headlines and perhaps on what we think are the key drivers behind them. So, to start us off, this first graph is an illustration of some recent ABI motor claims data. And you can see from the blue line that the overall claims frequency has been falling since 2015, this is great news and the reasons for that, we think primarily through the vehicle and road safety improvements that we've seen. However, in contrast, looking at the green line, you can see that the average cost has been increasing for that period and this is driven by inflation across both damage and bodily injury claims which is affecting everyone. So what have those recent inflationary pressures, Ogden included, done to our reserve mix? You can see from the second graph that the top light blue portion representing large bodily injury claims has increased significantly post Ogden reserving. I want to emphasize you that this is outstanding prudent reserves and not actual or ultimate cost. If we turn to the next slide, whilst we just saw that frequency has been reducing, the cost of repair in cars, and certainly newer cars that is with enhanced safety features and advanced technologies is outstripping any frequency saving. And to illustrate this point, even on some of the most simple of claims, you can see here some of the stock differences in the parts prices between older and newer car models. And parts like headlights are just evolving all the time. If you look at Audi and an A4, moving from halogen lights to xenon, this represents a 421% increase in parts cost. But frankly, I think some of these are at that extreme, but it's not just companies and manufacturers like Audi where we see inflation, it's across parts of Ford, Citroen, Mazda and the like. So what are we doing to contract this? Well, we have an innovative approach to procurement, which is very closely aligned with our claims experts. We are building relationships with manufacturers and we are investigating further pricing opportunities. And I'm also pleased to say that recent industry benchmarking shows that our inflation on damage is below the market average. Moving on to bodily injury and whiplash in particular, it really wouldn't be a results presentation, I don't think without whiplash reform coming up. It certainly seems to have been a constant feature over the years. When I stood here in 2014, I said that the claims notifications received through the Ministry of Justice portal, we are at the highest since the portal had launched, but today in contrast we can see that June 2017 was the lowest volumes coming through since the last bill reforms, which is particularly pleasing. So, what's driving that reduction in frequency? Well, I'd like say, just to start with, it's further consolidation of the bodily injury lawyer market, but more than that it's far greater success I think for insurers in successfully defending late reported, spurious and frankly fraudulent claims that plays a significant part here, which is great news. We've included a reminder here for you of what we think will be included in whiplash reform. A big question will they actually happen? I think those that are close to it say that despite Brexit the government still seems to have a strong will to see these reforms through and the fact that it was included in the Queen's speech probably gives a little bit of greater weight to that. However, there is still no clarity on timescales for implementations, so we'll have to wait and see. We also note here that market expectation is that the cost of claims and frequency following these reforms will drop and that's not necessarily our long-term view. So that's more BI, but of course all of that pales into insignificance when it comes to large bodily injury claims, and of course Ogden. So again when I stood here in 2014, I talked about the significant inflation that we've seen across large bodily injury claims in the market and I certainly didn't expect to be stood here just three years later talking about an even greater inflationary element of those same claims which is the change in the discount rate. We have put this slide together for you as an example of the main elements that go into receiving a large claim, the future loss elements that are affected by that discount rate. And in this example you can see we've used a moderate brain type injury for a young person and you can clearly see the difference in reserve between the original discount rate at 2.5% and the minus 2.75%. I'm sure you can imagine the impact of those rate changes on a catastrophic claim, when I say carriers, we needed multiple carriers 24/7. So we can understand the impact of the change in reserving, but what about the real impact of the change day-to-day in handling claims, is it really that stark? Well the discount rate is just one of many factors that come into the negotiation process and those truly catastrophic claims that I just mentioned that are most affected by the Ogden rate, they tend to be multi-factorial. So, what do we mean by that? Well, often these cases will involve lengthy disputes of the liability and/or high degree of contribution negligence, so experience, investigation and application really here are key. In addition, as lack of consistency between and medical and statistical experts to accurately predict life expectancy is now the area of contention between parties. So I think the combination of these two frequently encountered issues amongst many others, means that the reality of the situation when it comes to settlement often means that you talk about lump sum offers rather than nailing down to the nitty-gritty of all of the individual issues. I think that's what enables us to still carry on and settle claims. And this is where we feel our experience and approach really wins through and actually in a minus 0.7% world, it's even more valuable. On this final slide, we've tried to summarize our key strengths in claims handling and I'll leave it to read most of it at your leisure, but I will just call out a few points. If I start at the top right on tactical edge, I'd say that every large claim has a clear action plan for the strategy for that particular claim and the most successful and revered claimant barrister we've ever come up against is now our chosen defense barrister. And the technical edge, I'd say, not only is it our intelligent use of robotics is helping us become more efficient and accurate, but actually frees up our handlers to concentrate even more time on this technical aspect of their roles. And for collaboration, I'd say that not only do we achieve fair outcomes for all parties, but we are readily commended for our approach for the claimant lawyers, which is not easy. And finally, I'd really like to emphasize that the overriding strength I think for us in our claims handling is having staff who are not only highly experienced, but care passionately about the outcomes for us on to the customer. And of course all insurers will have some handlers who care like this, but I think it's our ability to consistently achieve this right across the board that really stands out and that's down to our culture. So, what will happen to the discount rate? Well, we don't know. And much like whiplash, it does – it still seem to be will the government change it, but again we have to wait and see. In the mean time, please remember that we are fully reserved from minus 0.75% and we are continuing to work hard to settle out claims in the best possible terms. So I'll hand you back to Cristina.

Thanks Lorna. So to summarize the UK insurance section, we have continued to grow in car, but especially in household and we have done so while increasing our profit. Now – what we are focusing now at the moment? Well, we have finished that rollout of our new system, Guidewire, that allows to increase the base of change. So, we're focusing on creating a bigger or a more intense customer journey online, both at sell stage – self-service and online. We are also focusing on increasing our telematics book and our telematics offering and the products that we offer. And finally, we are also improving our pricing capabilities by adding new external databases and also improving the tools that we use. So, for the future, you should expect that our strength in pricing, our cost control mentality and the strong claims negotiation puts us in a good place to continue our growth. And I'm going to leave you with Andrew to talk about the price comparison.

Thank you Cristina and Lorna, good morning everyone. Let's kick things off with Compare.com. First and foremost, Compare.com hit its objectives in H1, hitting marketing breakeven, revenue exceeding marketing spend in two of its core states. The team at Compare worked very hard to achieve this goal and I'm obviously very proud of them. We extended that win however to 20 plus states, achieving overall marketing breakeven countrywide in each month of the first half of the year. This was beyond our expectations and was made possible by meaningfully reducing our acquisition cost metrics and delivering even more rates from even more carriers to grateful customers. Now, this was not done by slashing quote and sales volumes. On the contrary, we increased our clicks and sales but on reduced overall marketing spend. The added volume pleased our existing panel members and makes it easier for us to attract new ones. As we reflect on H1, the team achieved so much. We've made great progress on our metrics and maintained a strong growth trajectory of improvement on on top of all these existing gains. More volume at lower cost is one of the keys to success for this business and while we cannot yet claim victory, we're pleased with our progress. Consumers obviously understand and prefer this approach and the more rates from well-known brands we show them, the happier they get. We now serve all 51 markets in the U.S., recently adding Alaska. Carriers, the other half of our customer equation remain quite satisfied as well. They like the flexible quoting model we provide, as well as the increased volume of customers available to them. These achievements do need to be caveated as we remain on a small relative scale in the U.S. All that said, challenges do remain. As a reminder, the U.S. is a very large market, $208 billion, both a challenge and huge opportunity. And it's a fragmented market, 51 discrete markets with carriers catering to specific segments, nonstandard and standard, rather than quoting all comers. Competition is fierce with insurance advertising reaching $6.5 billion last year alone. The four biggest players dominate both the airwaves and market share and while they are showing greater willingness to participate on our platform, none have yet fully embraced this channel. Complex coverages, agents and bundling are all reasons why this market will not develop as quickly or likely as far as the UK market did. Now that I've given you the optimistic and pessimistic views of our business, I'll return to our overall economics. Operations last year and this year continue to yield better than forecast overall losses and we hope to continue that in H2. We believe the comparison model is now part of the U.S. distribution ecosystem and we plan for it to remain that way for years to come. I will be remiss as an Admiral business if I didn't highlight that compare.com now regularly shows up on various top workplace lists. It's all about the team. The inevitable question is, when will we achieve overall profits? The short answer, we don't know. We don't know the real answer. And when it becomes possible for us to do that, we could choose more scale, further delaying profits. Reaching it depends on our ability to do two things, as you can see above, scale the marketing and improve the ultimate conversion of our carriers. We're scaling marketing now with our most successful TV campaign to date. Progress on conversion is being achieved, both with internal and carrier-driven efforts. As you can see from this chart, on relative carrier conversion rates, there remains a great opportunity for improvement and that improvement goes straight to the bottom line. For compare.com, I want you to take away a message of nice progress. We're encouraged, hopeful, even optimistic, but it's not time for celebration as much more work remains. More time is required before we can claim that ultimately desired outcome. Leaving the U.S. and returning to where it all started, let's talk about confused.com. In a market that is struggling to differentiate one comparison business from another, confused.com has made a strategic shift to focus on drivers and all the things necessary to be the number one site for car savings. This move has required investment in both new product verticals as well as a new position in consumers' minds. These investments include a 30% increase in marketing spend, with an ad campaign featuring James Corden. It's yielded greater top of mind awareness at the expense of near-term profits. We remain encouraged by the early results of this strategic shift and hopeful that this will yield higher profits and market share in the years to come. As you know, there are no guarantees and time will be the judge of this shift in strategy. Moving to mainland Europe. Let's discuss our comparison businesses in France and Spain. Overall, the two businesses showed both growing revenues and profits. LeLynx faced greater market competition, but still yielded double-digit revenue growth and panel improvements. The team in Paris remains focused on increasing brand awareness and preference, while improving conversion and customer experience. Rastreator ended another excellent half year and remains the clear leader in the Spanish market. They are using this position to diversify into both telephone and finance, with promising early results. The team in Madrid remains focused on improving price accuracy and conversion, while beginning to offer Data as a Service to the market. Overall, Admiral's comparison businesses remain strong. Compare.com is growing in a challenging market, delivering better than forecast results. Confused.com is investing in a new product vertical and brand position that holds the promise of differentiation and greater long-term profits. Rastreator and LeLynx have different market positions in different markets, but both continue to grow. With that, I'll hand the floor over to Milena to talk to you about the International Insurance businesses.

Thank you, Andrew. Good morning to everybody. I'm really glad to be here today to talk to you about the progress in Spain, Italy, France and United States. And I'm glad because it is a positive picture overall with substantial growth everywhere; improvement on our key metrics; a reduction on the overall losses from GBP13 million to GBP10 million, as Geraint reminded us before; and also a slightly more positive outlook in the market than a couple of years ago. So with that in mind, I will start to talk to you about Europe and then move to U.S. So as just mentioned, the first positive notes come from the restart of the economy and the opportunity that the local market present to us. Everywhere, we are in the worst part of the cycle. The cycle is worsening. And after years of massive fall in rates in Spain and Italy, we finally some signs of price increase. And this is quite relevant to us. Historically, increasing price has been a strong push for customers to shop more through price comparison site and direct channel. And it's worth to remind that differently from UK, in Europe, the direct channel is still quite small and has quite small share of the market, from 15%, 20% in Spain depending on how we account for hybrid models, to a very tiny 4% in France. Another restraint to the development of the direct market in the last year has been the reduction of car sales. Actually, in the last 3 years, the trend is reverting and now it's almost fully recovered. In Spain, for example, we're now back to the same volumes of cars sold than in 2008 after it almost halved in 2012. If we added to this an incessant increase in the adoption of smartphone as a way to look for car insurance and buy car insurance, you see our overall outlook of the market is slightly more positive than it used to be. We're now already growing at quite substantial speed. We passed the market of 200,000 customer in Spain and 100,000 customer in France. Brand and conversion have been the main drivers of this growth. You may remind that six months ago, we aligned our focus on increasing brand awareness. And so I'm quite pleased to say that we're improving year-on-year. And actually, in the last 12 months, we increased on average 10%, even with reduced media spend. Very similarly, 12 months ago, David mentioned that European insurance had a clear top one priority, and this was to increase our share on price comparison site. And again, conversion on this channel is improving between 10% and 20% in France and Spain year-on-year. We are also enlarging our product offering. For example, we're starting to offer van in Spain as we are doing in UK. And we're focusing very much on improving the customer journey, particularly online for more demanding multichannel customers. Finally, to support the larger scale, we migrate our IT platform in Italy to Guidewire, that is now adopted largely everywhere in the group. And as a collateral note, it's very nice to see how this partial convergence in technology is actually helping us to share more and more best practice at group level. What – which was the cost of this growth? So the growth came at a cost of additional GBP3 million in Europe. That is quite fair investment considering that in the last year, we had more than 170,000 customers and more than GBP36 million of turnover. And just in the last six months, it was 90,000 additional customer and GBP23 million of turnover. On the bottom of the slides, a couple of consideration on the results. On the positive sides on the left, you see how we're clearly starting to reap the benefits of the economy of scale. Our expense increased by 7% and corresponded to a turnover increase by 30%. So this means that we're actually growing more efficiently. And indeed, we did reduce our acquisition cost substantially. On the less positive side, on the negative side, our loss ratio combined deteriorated. And this is partially a reflection of a substantial growth in the worst part of the cycle and partially a reflection of moving our reserve methodology in France and in Spain toward the same methodology we are adopting already in Italy and in UK. And in the bottom right of the graph or the slides, you can see a graph that shows our movement in the book loss ratio for Italy as an example. Now before to move to U.S., I think it's worth to remind that ConTe had another half year of profits, continued to show an underwriting advantage versus the market and more particularly towards direct competitors and at the same time is growing quite substantially. So we move now to U.S. The trends in the market, as you can see, are very similar to Europe, but much more impressive. This is because we are used to see a very flat combined ratio in U.S. historically and such a strong upturn in the cycle has never been seen, at least not in the last couple of decades. And the results of U.S., of Elephant, are quite impressive. We did increase our turnover and our customer base, respectively, by 19%, 17%, and at the same time reduce our losses from $15 million to $6 million. The reduction and improvement on these results are mainly due to improvement of the technical results and improvement of the loss ratio in U.S. And again, this was the main focus that David highlighted six months ago for the United States team. Now said that, we've also been lucky because we experienced a more benign weather in Texas, where, you may remember, we had the big hail last year. And given I mentioned Texas, I think worth to remind that we are now trading in six states in the United States: Indiana, Illinois, Texas, Tennessee, Maryland and Virginia. And that Texas alone represent more than half of our customer base. Actually, as an insurance market, Texas alone is of a comparable size to France or to Italy, much larger than Spain and even larger than UK. So it has been really key in order to achieve the results for us. Now to summarize. Again, we had a very good growth in every single market. We improved our metrics. More notably, we improved our loss ratio in United States. And overall, we feel we have a more positive outlook of the market compared to the last couple of years and stronger foundation. And this put us in a quite good position to continue to grow and to develop our business outside UK. Thank you very much. And I now hand over to David to talk to you about other group opportunity and wrap up. Thanks.

Thank you, Milena. In my CEO comments on the release, I said that Admiral was ambitious in pursuit of immediate growth opportunities and long-term growth opportunities. And my colleagues have talked to you about those immediate growth opportunities. I want to focus more on the longer-term growth opportunities. We've invested around 3% of our profits during the course of the first half in initiatives that are irrelevant in terms of return in the short term, but I think potentially, very, very important for the medium-term and long-term health of the business. And I've chosen 3 examples to talk to you about today. Admiral Loans; in the first half of 2017, we originated our first loan on a dedicated end-to-end, mobile-friendly loan system. We are currently active in a very small way on price comparison sites, offering a fairly vanilla unsecured personal lending product to prime and near-prime customers. But we are laying the foundations for a healthy long-term business, good team, good system. It looks like a good expense base is going to be built here. It looks like a good response from the Admiral customer base to the loan product, so very encouraging. Another thing we've been working on is offering nontraditional forms of car insurance. The vast majority of the UK market is very happy with car insurance as it stands, but there are some people with more specific needs. One example of that is a site we've opened up for people who have short-term needs, 1 hour to 90 days. Learner drivers who want to do a lesson. People who haven't got a car, they got a license, but no car, they want to use a friend's or a family member's car. That site's growing faster than we anticipated. The acquisition costs are proving better than we anticipated. There are a number of insure tech brands who are start-ups in this sort of space. It's very hard to know for sure, get market data, but from the data we can get, we think we're substantially bigger than any of those brands in the short-term insurance market. And more encouragingly, this product is bringing new customers. The majority of the users of this product are new customers to the Admiral Group. And the last of the initiatives I'll highlight is expansion into other areas in nonlife insurance. So we've gone from car to home, from home now to van. Before the end of the year, we will be launching a travel insurance product underwritten by Admiral as well. Now one of the key questions is how do we make sure we don't go from being a company that's excellent in one or two products to being mediocre in a number of products? And I think the key to that for us will be to maintain the elements of our culture that are key to our success in the past and will be key to our success in the future. Which elements am I referring to? There's a lot I could talk about. I could talk about frugality. I could talk about commitment to finding a way of doing better by the customer in any market we enter. I could talk about data-driven decision-making. But I've chosen to talk about three elements of our culture, which are the three that I talk about when I stand up in front of new members of staff, as I do all new members of staff in Wales. The first one I talk about is people who like what they do, do it better. We want our staff to enjoy what they do. That's important. The customer feels that when you call and make a midterm adjustment. You can hear it on the end of the phone. But it's also important in terms of keeping our staff. We have excellent retention, and Lorna talked about the quality of some of our claims handlers. They've been with us for a number of years. They will stay with us because they enjoy working for us. It's also important in terms of attracting good new employees. The loan team is a mixture of long-term Admiral employees and high-quality experienced lending managers who've chosen to join us. And one of the reasons they've joined us is for the culture. The second point I make to everyone is we want people to feel like they own the business, so we make them owners. So when you're having that phone call with someone to do a midterm adjustment, the person that answers your phone, 20% of their remuneration is equity and it's Admiral shares. Their manager, the team manager, typically 33% of their remuneration is equity. Why do we do that? We want people to care about the outcome of everything they do for the company. We want them to care about making sure the customer has a good experience, that we don't waste money, we want them to be adaptable to change, and we want to encourage, very importantly in the context of new products, cooperation across the whole of the group. We are big believers that there is a good payout in terms of cooperation. Car insurance can learn from lending, lending can learn from car insurance. House insurance can benefit from sharing with travel insurance and vice versa. The common ownership and interest in the shares is important. And that's also reflected in the last point we make: The team, the team, the team. It's all about cooperation. In Admiral, you don't get on by elbowing your way past your colleagues. You get on by finding ways of cooperating with them. So those are the elements of culture, among others, that we think are key to helping us in the long term make a success of these new ventures. Thank you. And I'll open it up for questions. A - David Stevens: Will you have a microphone?

Yes.

Thank you.

It's Kamran Hossain from RBC. Two questions. The first one is, it feels like you're – I don't know how whether – it's just me, it feels like you're doing a lot outside of UK mostly. Lots of things you've done in the past, but that you seem to be mentioning more and more new initiatives. Do you think on the UK motor side, we're kind of at the peak of where you could be on market share? So that's the first question. And the second question. In terms of the expansion into new products, you mentioned 3% of PBT at the first half. What should we expect going forward on that? Thank you.

In terms of costs, something similar going forward, that sort of level. In terms of long-term prospects for UK car insurance, Cristina?

Yes. I don't think we're at the peak. The fact, that our ratios on pricing strength and cost control are still some of the leaders in the market, we believe we can still grow. I think if we look long-term, 20, 30 years, there is a lot of pressure about the size of the market, but that's the type of timeframe we're looking at. And when we look about initiatives on other products, we also think that we can give a better offering to our customer. So when we offer a car and then a multicar policy, partly it's because we learn more about the customer, we can offer a better price. And that’s similar things we want to do with our household, our van, and our travel offering, it's just to offer a multicover proposition to the customer.

Should we come in and go across; this gentleman here at front?

Thanks. Thomas Seidl, Bernstein. Three questions. First is on pricing. In the UK you mentioned you put through price increases in December, also year-to-date, but average prices per car seem to be flat year-on-year. So does it mean you changed the mix? Or what – how would you explain that? Secondly, on capital guarantee, you said you're still strong on Solvency II standard model, so why withdrawing now from giving back extra capital just because you delayed the internal model a couple of months? And thirdly, on the compare.com in the U.S., if I'm not wrong, the carriers you have listed up for compare.com are stable. So you have not signed up new carriers recently and may have even lost one or two, if I'm not wrong. So the question is, is this business plan now based on keeping the carrier list stable and you just play the advertising and the conversion thing or are you still hopeful to increase the carriers?

Should we do that in order then? Cristina will take the first one.

Yes. So, yes, the mix has changed. First, I think we need to say that when you put a price increase, you shouldn't see a translation fully of that price increase because there is normally a change of mix. But in the case of what we have been doing in the past few months, there has been a change against some of the larger segments exposed to BI due to the [indiscernible] change. So, yes.

Caps also, I think it's fair to say that 214% is not the long-term sort of level we'd run at in terms of solvency. I think I'd almost say exactly the same as I said six months ago. We think it's appropriate to wait until we get through it and deal with the internal model position before we commit to any potential returns of surplus that might exist at that point. And the reasons for that are the position – obviously, we have a fairly good idea of what we think the position is and should be, but it's not fully in our control, so we think it's appropriate to wait until we've gone through that process, get that certainty. And then we'll obviously communicate any potential return of surplus at that point. I might also point to uncertainty over the Ogden consultation, where that might go, particularly with respect to PPOs and the impact they might have on capital, but the internal model certainty is the key thing.

And on compare.com, we do continue to add carriers. We do continue to add carrier states. Because of the 51 discrete markets, a lot of it is less about adding specific new carriers as to getting density in specific markets. And some of what we're doing is mining our existing carriers for additional state opportunities, but there continues to be carrier additions as well as product extensions moving into home, rent, or a condo to complete the bundling part of the portfolio as well.

Thank you. It’s Ravi Tanna from Goldman Sachs. And I have three questions, please. The first one was just on the Admiral Loans venture. It strikes me as a potentially quite capital-consumptive initiative, and I was just wondering if you could elaborate on thoughts around that and the shape of the group from a capital perspective, has that built on? I know it's early stages at the moment. Second one was on the capital front. It's more a point of clarification really. Can you give us a sense? When you talk about being comfortable with the internal model, does that mean when you get approval, does it mean when you make the application or does it mean some point in the interim? And secondly, why have you kind of expanded the scope at this stage to include the capital risk module? And then the third and final one was just on small bodily injury claims. The Slide 20 chart showed quite a big drop-off in claims through the portal. Last time you experienced that, post LASPO, it took quite some time for the favorable reserve releases to flow through. I was wondering if you could comment on where we are in that journey this time around. Thank you.

I'll do loans and reserves feeding through, and Geraint will do the two capital points. On loans, at the moment, the size of loan business is immaterial in terms of balance sheet strain. Our heritage as an insurer is to establish good underlying economics and then partner with reinsurers who provide capital support to our initiatives. Our hope is that we will demonstrate that, those good underwriting economics, and be able to partner with external sources of capital to achieve some efficiencies there. In terms of the impact of the fall in small bodily injury on ultimate profits, it is the case that when we report profits, we tend to report profits that are a function of what happened two or three years ago, so you would not expect, in the Admiral way of reporting profits, that the surprisingly low bodily injury frequency we've seen in the first half would necessarily immediately have fed through. Geraint?

The timing; so we expect that the application will be in the latter part of 2018, so somewhere probably in the fourth quarter. I feel a very good process would take six months with the regulator. I think it's unlikely that we would commit to any returns with – as the application goes in. I think that's probably premature. So at some point, after approval is given, or certainly when we get a lot more clarity on the position, I think, before we would communicate further on potential returns of surplus. And the scope, there are a large number of reasons. I think, over the past year or so, it's become reasonably clear to us that the standard format doesn't deal particularly well with very volatile interest rates, particularly in the environment we've been in. So that's one factor. We think, actually, there's quite a lot of interaction between inflation and interest rates and the insurance risk as well. And I think to properly capture those, we believe that we need to develop a better market risk model. There's also regulatory interaction considerations as well, so we think it's appropriate to go with a more complete model rather than go with the first application and then go back 12 months later, then 12 months later, then 12 months later saying, can we have this bit, can we have that bit, can we have the next bit. So we're trying to balance between getting the model approved and getting it used as quickly as possible, but also getting a robust as possible model in – at first application.

Thank you.

Now very unusually, we actually have a question coming in from – outside, from the call. Do we know how to – this is so unusual. I not even sure of the logistics. How do we handle that?

So [indiscernible].

Okay, we'll take that question after this one if we can sort out how to take it. I'm going over there.

Arjan van Veen from UBS, just coming through again. Two follow-up questions on UK motor pricing, please. Firstly, going back to the original question, your GWP was up 7%, policy count was up 7%, yet you indicated to us six months ago that underlying claims are facing this kind of tracking, is 3%, 4%, 5%, plus Ogden 8% to 10%. So can you just square that a bit and then give us some indication of what the risk adjusted price increases were during the period? And then, secondly, when we look at your slide around the competitiveness of UK motor, you show it’s falling, and then obviously coming back as other were people pricing up, but it's below where it was, end of last year. So can you give me a bit of commentary around has – therefore, your policy can start to improve towards the end of the second – first half into the second half? And then, also, can you comment on your gross versus net in terms of pricing? So do – you may have adjusted for net pricing now, but do you need to put prices up further to fully cover Ogden on a gross basis? Thank you.

Yes. So what we have done so far this year is take into account the full extent of the Ogden change and the full impact in our book, and that means gross arrangements. So what we expect to do is that if there is no further clarification from the Minister of Justice of how is this rate going to change going forward, it will stay at the same level as we are today. We expect the reinsurers to adjust the pricing level quite significantly in the program that is going to start next year. And I think both Admiral and the rest of the market are going to adjust their prices in January next year to reflect that. So what have we done already is taken into account our current program, adjust the prices. So Ogden is fully embedded in our pricing increases, and we have taken also some steps regarding the inflation. I mean, remember, prices have gone up since March this year, but they have also been going up quite substantially last year, and that they were already reflecting the increase in negative inflation in the prices. Also at the moment, even though there was a significant increase in March, we've seen that the market continues to put price up month-on-month, so it's not yet clear when is this going to finish.

Okay, great.

[indiscernible]

Fine. Thank you.

[indiscernible]

Building the suspense.

We have a question on the phone lines from Andy Hughes from Macquarie. Please go ahead, Andy.

It's Andy Hughes. It’s good to know I’m just asking difficult questions. So I just have a question about the ultimate loss ratio movement, the projected ultimate one, from the end of the year to the first half. If I look at 2014, it's 74% in both cases, 2015 has gone from 76% to 77%, meaning neither of those years has particularly moved very much due to the impact of Ogden or maybe it was already factored in to the ultimate best estimate loss ratios. And I guess what I'm interested in is, the ultimate best estimate loss ratios have developed positively in the past. Presumably, some of that has been because you've settled things at 2.5% rather than at the implied rate in the reserving and in ultimate best estimate. So I'm curious as to how much impact that had in the past as you settled those claims through. And I guess the point that Lorna made about the – how these claims are settled and how you have an advantage that's even stronger going forward, post Ogden, than you had before. Are what you're saying that these debates are really about the quantum? So you sit in a room, you say, okay, this claim is GBP5 million and then you allocate it back to the different lines. So actually, when you change Ogden discount rates, maybe then you sit down and you have a different conversation, but it's not quite the same as saying, what we've played last year has gone up by the impact of the Ogden discount rates. Thank you.

Shall I take the first bit on ultimate?

Go ahead.

So I think the question was relative to the comparison of the June 2017 position against December 2016. And what we – what I said earlier was that we'd seen general stability in those projected ultimate loss ratios apart from some reasonably small movements, including the small improvement on 2015. The position at the end of 2016 did include the rebasing of those ratios for the minus 0.75% rate. And so actually, the process we've gone through since we reported our results in March is to, obviously, put the impact of that minus 0.75% through the case reserves, and then, obviously, reprojected the ultimate at the middle of the year. And so seeing stability in that context we think is actually a pretty positive outcome. But the December 2016 numbers did include minus 0.75%. The claims question?

And in terms of big bodily injury settlements, Lorna, which you've won?

I think we're – the point I was trying to make on the – is even more valuable thing is, if you get a case where – an easy example would be somebody who fails to wear a seatbelt and has been badly injured, and you're arguing a 10% or 15% contribution on that. If it's 10% to 15% of a GBP3 million claim originally, what's 10% to 15% of a GBP 5million or GBP 6million claim and at the new rate, then it's – that investigation is even more important to be able to prove that. And that comes even more into play when you have things like contributions. So as people get into cars with drunk drivers, that kind of thing. So I think our advantage there in being able to look at things like CCTV and establish relationships throughout the night for a town, just to be able to show that one driver was actually with that passenger throughout the night and they knew they'd been drinking, is quite helpful. When it comes down to settlements on these types of claims, I was in a settlement meeting a few weeks ago, and the other – the lawyer started off by saying, look I only talk about a 0% rate. And our approach to that will be, well, we're not even going to talk about a rate, we'll just talk in actual sums of what we think this claim is worth. And that has been our approach so far which seems to work on these claims, where there is contributed negligence. We don't get many where we are absolutely 100% at fault.

Thank you.

Thank you.

We're going back and then well work our way forward.

Janet Demir, Morgan Stanley. Two questions from me, please. So just to be clear on the rates. You've said that your competitiveness has now improved since March. Does that mean that you're increasing rates in line with the market or below the market now? And then just on the commutations, I note there was no commutation in 1H, presumably from the 2015 year. How do you think that will develop? Are you waiting maybe a little bit longer just to see how Ogden impacts that? Or what's the reasoning behind that, please?

On the rates, on Page 13, I think the page that shows Admiral times top, it tells the story, really, of a situation where our rate increases were earlier and substantial. The market then moved up substantially in March. It has continued to move up. We've also moved up. We are less competitive than we were a year ago. So by implication, our rate increases, versus a year ago, have been higher than the market, which is not irrational in the context of a temporary competitive disadvantage due to the impact of different reinsurance structures during 2017. In terms of commutation?

Commutations, basically you answered the question. I think, so normally, we would commute 2015 in the first quarter of this year, and that decision point coincided exactly with the point in which the new Ogden rate was announced. And so even though we were projecting that 2015 would remain profitable at that point, it didn't feel appropriate to commute – cover a third of the reserve for that year at that point. In terms of what we do in the future, it will depend partly on what comes out of Ogden, the consultation. I'd expect that we would commute it. It appears to us to be a very profitable underwriting year, so I'd expect us to commute it at some point. But the clarity on the Ogden consultation will be a key input into that timing.

Okay, great.

Thank you. If we go just forward one. Thank you.

Dhruv Gahlaut, HSBC. Three questions. Firstly, the reinsurance contracts would come to an end next year. You're fairly confident you guys were last time when you presented in terms of nothing should change. Is that still the view on those reinsurance contracts? Have you started engaging with those partners? And when do we hear more? Secondly, on the U.S. aggregator business. Could you remind us as in how many states now you guys are breakeven in, in the first six months of this year? Also, as in what are the number of panel players you have currently? Similarly, for the French aggregator business, if we can know how many players, panel players, are there on that website? And has – have players like AXA fully engaged with you guys there? Thanks.

[indiscernible] first, we'd expect to get into negotiations and discussions on extending those contracts beyond 2018 in the UK shortly. We've had some discussions with some of our partners and we believe that the conversations will be very similar. There is a clear interest there, and we'd expect to be able to renew those deals if we want to on – at least comparable terms to lift the current contracts.

Andrew, do you want to do U.S. and I'll do France?

Sure. Most of the states that we were aggressively advertising in, we were better than marketing breakeven. It was north of 20-plus states. There were a variety of states that we don't have a panel that would be sufficient for us to aggressively market. So some months, those would be profitable; some months, they wouldn't. But it wasn't relevant in the overall scheme of things. The number of carrier is north of 60 contracts signed, and they bring north of 100 brands.

In France, the engagements of the traditional players is mixed in contrast to, say, Spain or the UK. So the big mutuals don't tend to engage a price comparison with the exception of [indiscernible], who was the market leader and is now a distant third, which is owned by the big mutuals. AXA itself is an important panel player under its Direct Assurance direct brand.

It's David Bracewell here from Redburn. Just a couple of questions on the U.S. price comparison business. You gave some stats on, I think, click for rates or the number of clicks going up significantly. But I just wonder if you could perhaps give us some absolute numbers or some indication of what perhaps your market share there is of the new business. And then a kind of a follow-up question to that really is, given the combined ratio in the U.S. has increased significantly and the pricing, as far as I'm aware, has increased significantly, I'm just wondering if that's led to significant increases in the churn or in terms of new business churn rates? And if that's helping you grow business in the U.S.? Thanks.

For competitive reasons, we haven't released absolutes, and we choose not to. From an overall market share standpoint, we remain quite small. As far as the aggregator markets, we remain quite large. So it's sort of on the overall scale of things there. I come back to the 51 discrete markets. The dynamics in the market are highly variable. You will get some states, Georgia, California, as example, that have proven to be very volatile in the last six months. Carriers coming in and out with appetite. People feeling like they've got their rates adequate and then others feeling inadequate. So the panel will move based on competitiveness there. And the consumers shop in response to what the carriers are doing. If they are raising rates, you'll see more shopping in different markets. But it's a state-by-state phenomenon. In some cases, it's a DMA by – or city-by-city phenomenon.

Thank you.

Andreas van Embden, Peel Hunt. You mentioned in your press release, a reduced level of relative reserve margin strength in UK car. Just thinking about Ogden and assuming no change in Ogden future, reserve releases, would that mainly be driven by the improvements in the ultimate going back to the 2% to 3% per year, which you showed in the past? And you want to keep your reserve margin where it is today? Or will it be again a combination of both? Thanks.

The relative size of the margin has come down since the end of the year when it was basically at its peak because of – largely because of Ogden. If Ogden stays the same, minus 0.75% and nothing else changes, then we would expect significant reserve releases. Because the margin we have is significant, so the unwind of that margin as we move towards the ultimate outcomes obviously gets released into the income statement and the most recent years the margin gets replenished. In terms of the direction it goes in the future, it's difficult to comment because, obviously, it depends on the conditions and circumstances at that point. And we said in the past that we would expect to live in the middle of our range, we don't give these kind of tolerances of the range. But at the end of the year, we were right at the top end of it, and we've moved back down towards the middle of it. But where it goes in the future will obviously depend on circumstances at that point. But we would sort of – I would expect our reserve releases to run well ahead of 2% to 3%, that was much more of a market number, I think, rather than Admiral.

We missed out at the very back. Can we get Luke back at the very back, and then I've only got one more after that, which would come to the front.

It's Andrew Crean for Autonomous. A couple of things. Could you say what's happened to the Ogden reforms, there was a number of dates in early August suggested for a release from the independent commission. And secondly, could you talk a little bit about your diversifications over the years? I mean, clearly, I can see that you're trying to build a big business, which rivals that of the UK major business. You have been at this since 2006 in some cases. At what point do you say, we are not going to create a great business, and perhaps pass it on to somebody else?

So what's the first question, Andrew? Ogden, yes. Okay. So Ogden is in a state of flux. The government realizes that they – it was a bizarre and eccentric decision and that was reflecting the fact that immediately they announced it, they announced the fact that they were reconsidering it. They were then due to announce the result of a consultation – not the result, but some feedback from the consultation in early August and they haven't as yet. The general view of the market appears to be somewhat optimistic about a move towards a more rational outcome on Ogden. I think the challenge – and that's reflected, I think, for example, in the number that esure gave for the increase in its excess of loss, which to me was implying that reinsurers felt that some – there would be some step back. The challenge is the government's ability to deliver our legislation at the moment. And therefore, I don't think anyone should take it for granted that there will be an improvement in the Ogden situation in the current parliamentary context. In terms of investments, I think it would be bizarre to step back our investments in our international businesses and compare at a time when they seem to me on the cusp of actually really demonstrating long-term value for the group. I think the profit number is only a partial reflection of what we're building. In the international insurance business in Europe, when you take into account the growth of business, the time of the cycle and the fact that we are introducing more caution and more reserve redundancy in France and Spain, we're very, very close to something that's clearly valuable in profit terms and is even more valuable in underlying capital accumulation. Compare, it looks like an exciting business where if you look at the GBP6.5 billion being spent on marketing, if we can only get 10% of that marketing spend switching into price comparison, it's a huge opportunity. And one thing we've certainly learned in America is it isn't easy to do price comparison. So there's a really defensible position if and when we make a success of this. So I think there are lots of exciting things going on. I think loans is a huge potential business that gives a diversification for 15 years or 20 years time when autonomous vehicles might become a threat to the core business. So I'm very positive about what we're doing. I recognize shareholders – we've been asking shareholders for patience in terms of return in the terms of – in terms of profit. But I think, ultimately, they will prove their value to the group.

Good morning. Greig Paterson at KBW. Just three quick questions. One is on van, I wonder if you – I don't know if you've got a net written premium number. In other words, if you converted your entire stock, how much should we add to the net written premium by the end of the year for van? Second point is, in terms of written, the difference between booked and ultimate for 1H for 2017, I wonder, have you increased or decreased that official margin versus the 2016 booked? And then a third one, usually – I mean there were 3 questions on rate. Usually, you provide some kind of number on rate, so let me try again. If you look at June year-on-year rate, assuming no change in mix, what will be the number? And then if you adjust it for the fact that your renewal business has increased relative to a new business, what is the effective rate that you get, please?

Do you want to do van, Cristina. Do you want to rates as well?

Yes. On van, I think the best way to look at it is that we had a size of the broker book of 175,000, and we're going to migrate the book starting mid-May. So if we manage to migrate all the book and finish with a similar size, it will be seven months of the year, truly June to December, and the average premium that we have at the moment is quite similar to the one that we have in car.

What is it?

North of 500, but not much.

Yes.

So 500x the 175?

But only 7 months of the year in our book.

That assumes we get – we migrate them all.

And we grow our written sales. So at the end, it's more or less stable.

And do you want on comment on rates again?

I wasn't sure exactly the point...

I think Greig's after a sort of precise number in terms of...

What was your base rate – to use your technical terms, what was your base rate increase in June year-on-year? And then how much was that distorted by the fact, when you actually look at your effective rate that came through, how much was it distorted by the fact that there was a reduction – or increase in your business? And so that we can – we've got a number that we can actually put through the loss ratio effectively. And the June one, so year-on-year on June.

I think our base rates, on average, across new business and renewals, were double digit, but only just. The change in average premium is mostly a reflection of mix within new business, but partly a fraction of an increase of renewal within the whole mix. The middle question I think was a Geraint question, was it?

Yes. What was the question?

So let me be technical. If you look at 2017...

I'm sorry. If the...

Booked versus ultimate on a written basis, what was the official margin there? And how did that – have you changed that from what you booked end of last year?

Well, I'm not sure we actually give you any of the numbers to get to that official...

Well, I can look it out at the end of the year. I was just wondering where we are at the half year.

What we're saying about the margin is, in relative terms, it's decreased. We think about margin across all the reserves, not just by underwriting year. In general, the more recent the underwriting year, the bigger the margin is. We haven't fundamentally changed that approach at all. But the relative size of the margin is less now than it was six months ago, but it remains significant.

That's the official margin or the margin included in your ultimates as well?

Now we don't give any clarity on these – the margin. They're a best estimate, albeit with some caution.

We have investors waiting for visits, so we'll have to cut it there. Apologize for those who didn't manage to make a call. There might be some of us around briefly for follow-up one-to-one. Thank you all for coming. See you in six months