Afternoon, everybody. Is this on? Great. Thank you. I'm Brian Meredith. I am the Property and Casualty Insurance Analyst here at UBS. I want to thank everybody for joining us here at our UBS Financial Services Conference here in Chicago. Love all of your feedback also on it. We've really made a big effort this year to make it a much bigger conference here. Please, feedback would be great. Our keynote speaker is here. It gives me great pleasure introducing him. Peter Hancock is the CEO of AIG. He's been there since 2010, and prior to that, he was with JP Morgan, set up their global derivatives group, CFO. He's got an unbelievable past within the financial services industry.
A lot of great experience, as well as the vice chairman of KeyCorp. He's taken a lot of that experience and actually brought it to AIG and I think made a lot of tremendous changes within the organization. A lot of them that I know a lot of outsiders have been skeptical about. The last quarter proved that some of these are actually working pretty well. I want to give him a big hand for that one. The other thing I was just noticing actually, as I was looking at AIG stock price, I think AIG stock price has just about doubled since he joined the firm back in 2010. Clearly he's created a lot of shareholder value for people at AIG. With that, I'm going to turn it over to Peter.
He's going to give us some comments. Then we're going to open it up to a Q&A.
Thank you, Brian. It's a pleasure to be here and have a chance to hear from you and to share with you why I am in love with AIG. Who would have thought that at this stage of my career, I'd fall in love with a company? I feel unbelievably blessed to be asked to lead this company two years ago. It's an extraordinary company of very talented employees, all 56,000 of them in 90 countries, that have unbelievable relationships with their clients, distribution partners. I just want to use this as a platform to thank all of them for the extraordinary job they've done to make my job a whole lot easier.
To me, an important part of my message today is what we are doing to reshape the company so that more of our fate is a function of the hard work of those individuals and the customers that depend on them than on exogenous factors like interest rates, the property P&C pricing cycle, Brexit, or other factors that happen to us. We want to make our mix of earnings more a function of the hard work, innovation, efficiency, and customer intimacy, the relationships that we have with customers that date back decades, that are in multiple lines, in multiple countries, that give us an edge that is a source of long-term competitive advantage.
Yes, we are on a path to improve our profitability, and yes, as Brian says, the second quarter was a very helpful indicator that we are on track to meet the goals that we laid out in our January strategic plan that highlighted four important metrics. One, returning $25 billion of capital to shareholders through 2016 and 2017. Two, to improve our return on equity. Three, to reduce expenses, and four, to improve our underwriting margins as measured by accident year loss ratio. On all four of those targets, we are ahead of schedule. We returned almost $8 billion to shareholders through buybacks and dividends in the first half of the year. We've improved the accident year loss ratio year-on-year by four full points.
We've reduced expenses for the first six months of this year versus the first six months of the year before by 11%, and we've improved normalized return on equity by 200 basis points. We think that the second quarter is good progress, but we set out an eight-quarter plan, so we're literally only 25% of the way there, and we certainly don't want to declare victory. There's a lot of hard work ahead of us. As I said, the mix of earnings is as important as the quantity, and the proportion of our earnings that is coming from mark-to-market assets has been reduced substantially. In our earnings call, we highlighted the reduction of market sensitive assets as a proportion of total assets from about 11% to 7.5% year-on-year.
We continue to redeploy capital away from hedge funds, which at its peak was about $11 billion allocation out of our $350 billion general accounts, to less than half that. As more of the other legacy assets that we have are sold or just wind down, the sensitivity of our results to factors outside of our control gradually diminishes, giving you, I think, an easier line of sight on our long-term earnings capacity. Which brings me to an important lesson I've learned, which is that it's really important to be transparent about this industry. It's a complex industry, for sure, and we participate in a lot of elements of the insurance industry, from consumer lines to commercial lines, and in many different countries. We've added a lot of transparency to our financial results and given much more clarity about our long-term plans than in the past.
If you trace it back to my first letter to shareholders in February of 2015, I laid out the basic outline of what we're doing today. In January of this year, I became much more specific about timelines and amounts, importantly, divided the company into two broad segments. The legacy segment, which has about a quarter of the company's capital, and the operating segment. The important distinction is that in the legacy segment, we have both property casualty and life activities that have below-market ROEs and that are basically in runoff. If I could wave a wand and get rid of those at book value, I'd do it in a heartbeat.
It was really important to let people understand how much of the capital was tied up in what is essentially a runoff activity so that we can shine a light on the operating portfolio, which is profitable and growing, and where our future lies. We have very different performance metrics for those two segments. The leader of the legacy portfolio, Charlie Shamieh, has a clear objective to liberate capital through a mixture of divestitures and reinsurance contracts, as well as portfolio transfers. Remember, this is an asset and liability portfolio, often in regulated entities. It's not as simple as selling non-performing loans or something else. It's not like a good bank, bad bank. It's not a perfect parallel. I'm happy to say that the legacy divestitures are also ahead of schedule.
We've done over $4 billion of capital liberation from the legacy portfolio in the last three quarters, we have plans to contribute quite considerably to the $25 billion total capital return through wind down of the legacy. The objectives of the operating portfolio is to improve ROE and to maintain an ROE comfortably above our cost of capital and grow. Grow profitable businesses where we have comparative advantage and focus. While the company is a lot more focused than it has been in the past, to give you a sense of the scale of that refocusing, we've sold over 90 companies over the course of the last seven years. Many non-insurance activities, consumer finance, aircraft leasing, real estate development, third party asset management, and so on. We still have work to do.
In the consumer lines and personal insurance, we are focusing our efforts on doing business in 15 countries only, down from 80, while continue to provide group personal insurance in 35 countries. As we reshape the company, we are effectively sculpting it to a more focused franchise around customer segments where we think we have something distinctive to offer. Think about this as value versus volume. We are already, by many measures, one of the world's largest insurance companies, there are diminishing returns to scale. We need to focus on the parts of the value chain where we're getting properly rewarded for our capital. One of the phenomena that's going on is a shift in the capital intensity of our business model. Yes, we have a large balance sheet. Yes, we have the largest capacity for highly engineered single property in the world.
Yes, we have the backing of very strong reinsurers that trust in our underwriting skill. We think the future is not leading with our balance sheet, it's leading with our risk expertise. That's why we're investing very heavily in what it takes to become a trusted partner to our clients in helping them manage the cost of risk. We've hired over 4,500 property engineers over the last five years. That's what gives us the ability to retain over 92% of our highly engineered property underwriting because of the dependence that our clients have, not just on our balance sheet, but on our expertise. That's why we have a leading position in emerging risk categories like Cyber, where the amount of the capacity is perhaps less important than the consulting that goes with it. This is not consulting of the PowerPoint variety that we're all so familiar with.
This is consulting that is deeply informed by the skin that we have in the game. We pay approximately 1 million claims per month, about $130 million per day. Every one of those claims has hundreds, if not thousands of elements of data. Distilling insight from that data, analyzing it, and sharing it with our clients is at the heart of how we become a platform for risk expertise, constantly learning about the changing nature of perils and viewed by our clients as an indispensable partner. We win when they win, when the frequency and severity of loss declines, when they spend their limited risk budget on insuring the things they should be insuring, and cutting back on insurance where they should not be insuring. We win their trust, we also win where we add value most.
The heart of that is looking where else can we invest in the services that are natural adjacencies to our capital deployment. How do we invest in the technology and the talent to be as relevant to our clients five years from now, 10 years from now, so that we can have another 97 years to add to our previous 97 years of history? What's going to make us relevant in the future is going to build on that foundation of trusting relationships with our clients, also really embrace the innovations that are coming around us at a pace that is absolutely mind-boggling, whether it's big data, the fact that sensors and chips are going to be free, and everywhere. I say free, they're going to be virtually free. If you extrapolate the cost per chip every year, it's heading to zero pretty fast.
The implications of that in terms of how you can monitor risk is pretty dramatic. Now, I'd like to coin a term, which is insurance as a service instead of software as a service. Think about insurance in the cloud, basically as needed, because you can measure risk in a more granular and a more frequent basis than ever before. We've obviously seen it in telematics in the auto insurance space. The issue is where else does that way of thinking apply? We're investing in what it takes to do that. We created a science group, a chief science officer five years ago, recruited scientists from very diverse backgrounds, over 140 of them over the last five years. Some of them have migrated into senior decision-making roles in the businesses.
Some of them continue to work on fundamental breakthroughs that are going to make us more efficient at managing our business, underwriting risk, settling claims, and they're not in any way claiming to be a universal source of insights. They are partnering with many interesting third parties, whether they be universities, think tanks, data analytics firms, or most recently, we announced a prospective joint venture with a very interesting investment management firm called Two Sigma that is made up of hundreds of data scientists and people with experience in the digital field to go after the SME business in the U.S., a $70 billion market that's still today very fragmented. I am, as I said, very privileged to be leading a company that has a proud past, had some very big lessons to learn from its experience in the financial crisis.
I think proudly repaid the U.S. government at the end of 2012, the $182 billion of assistance with a tidy profit of $23 billion to the taxpayer. Also is very focused on what it does best, which is insurance across a broad range of risk to a broad range of customers. At last count, we have 90 million customers around the world. Again, I'd rather have fewer customers that rely on us more, hence our vision to be our clients' most valued insurer. We want to occupy the high ground with our clients, doing for them what nobody else can do quite as well. That's about focusing our energy around the future, not the past. We've had to make some tough choices, exiting lines of business that are past their sell-by date.
Those tough decisions, I think, have led to a more effective, leaner, more efficient organization focused on delivering repeatable earnings. With that, I'd like to open to questions.
Thank you. We're going to open it up to Q&A. I also want to point out here, Peter answered some questions. We've got Liz Werner also in the audience here who can also help out with any really, really difficult questions. I want to kick it off with the first one. I guess my first question to you, Peter, you've done a great job kind of reformulating AIG here. You're cutting expenses. You're getting the expense ratio in order. One of the major assets I find of any P&C, the most important asset of any P&C insurance company or insurance company is its people, right? Clearly, we've seen a lot of people exit AIG. What are you doing to retain good talent? What are you doing to recruit good talent? Give us some sense of what's going on.
AIG is extremely fortunate to have a very deep bench of very talented people with deep domain knowledge and deep relationships with both insureds as well as distribution partners like brokers. We've had to make some very tough choices as we get our expenses down. 80% of our cost base is related to people, so you can't get your cost down without making some difficult decisions. One of the examples over the last nine months in terms of difficult decisions was how to simplify our management structure to speed up decisions. I started from the top down. I had a management team of 15. I reduced it to nine. Then I looked at how we organized decision rights across the company and reduced the number of people who had two bosses, matrix reporting, by 95% over the last nine months.
That doesn't mean that one of those two bosses was not good, but you have to pick who you're going to empower and really give them accountability and incentives to deliver results. We had to make a lot of those tough choices. Of course, there are many people, that you may have read about who've gone to work for competitors over the last nine months, but this is part of our reshaping into a more nimble, agile company focused on being responsive and agile in the face of changing circumstances. I think the second quarter demonstrates how that agility works in practice. You don't achieve the degree of change in mix of business, in expense load, and mix of customers when you have committees having endless meetings deciding what to do.
You do it when you empower strong leaders to be accountable for taking decisions and to act on them. I'm very proud of the way the team has operated so far. We're very fortunate to have a lot of people looking to come and work at AIG. I don't want in any way suggest to anybody that there aren't certain individual regrettable departures, but we monitor it very carefully, and we are very pleased with the quality of talent we have at every level of the organization.
Great. Then, one other question, we'll open it up. The 600 basis points of accident year loss ratio improvement by year-end 2017, that's a number that a lot of people on the street, I would say, there's a lot of skepticism around your ability to achieve that. Now, second quarter results have probably made a lot of people turn the other way a little bit here. I guess the question is, one, how'd you come up with the 600 basis points? Number one. Number two, as a follow-on to that, if there was something that was going to prevent you from achieving that 600 basis points, what would it be, if anything?
We've, as I mentioned, set out four metrics, ROE improvement, accident year loss ratio improvement, general operating expense improvement, and return on capital. Those four metrics are not equal. There's a hierarchy. ROE improvement, and this is normalized ROE, is at the top of the hierarchy, and the other three are means to that end. We estimated that the right balance between expense reduction, loss ratio adjustment, and capital that would get a sustainable ROE improvement was 6 points. As we displayed in both the second and the first quarter earnings call, we have a dispersion of loss ratios across the commercial lines that goes from the high 90s down to the mid 40s. This is anything but a monoline company.
We have a lot of levers to play with the mix of business, whether it's complete exits from certain sublines like buffer trucking and product liability, which is about 2% of our $20 billion of net premium written, or much more micro segmentation where we are looking at client relationships where there's only one line of business sold and it's a loss maker. Getting out of that is a no-brainer. No collateral damage to other profitable business and shifting our efforts and capital around multi-line relationships. Using increasingly sophisticated data analytics to looking at even more granular segmentation of our business. The 9 points of reduction, very much a function of looking at this. Trade-off between expense reduction, loss ratio improvement, and reduced capital allocation in order to get an optimal sustainable ROE.
Over time, I think that it will be clear that we emphasize sustainable ROE rather than these subsidiary metrics because I think it's important that people see them for what they are, means to the end.
In terms of what could get in the way of the 6 points, clearly a softer pricing environment than we have would do so. We made this commitment in January knowing full well that certain segments were very soft in pricing. U.S. Gulf property pricing has continued to be extremely soft, and we've been shifting our exposure away from that, especially in the E&S space, where the only differentiator is the capital you provide and not the services to a better mix of highly engineered property and middle market property where we have more sustainable margins. In casualty in the U.S., we've had actually a 5% price increase. We've actually been able to get decent pricing. I think that clearly, the biggest headwind would be a pervasive soft market.
As companies in our sector go, we're about as diverse as they come in terms of lines and geographies. One of the important changes to our management structure over the last five years is to break down some of the barriers to capital redeployment so that we can move them away from product lines and geographies that are in soft cycles into the ones where it's harder. That's one of the big benefits that we get from our diversity.
Very good. Thank you. Let me open it up to the audience. Questions from the audience. Peter. There's got to be tons of questions. I'm going to then throw it in while everybody's still pondering and writing them down. Non-bank SIFI. I guess you briefly alluded after another company had some success in a certain lawsuit that maybe this is something you might pursue. Give us your updated thoughts about that and whether you need to break apart the company and because of the whole situation.
I've been very consistent on this going back over a year, which is that of all of the strategic issues that we face as a leadership team, this doesn't even make the top 10. The reason is that we think we have a lot of work to do to improve the fundamentals of the company, and that to engage in the complex task of re-designation under the Dodd-Frank designation as a systemically important financial institution is hugely distracting to management and is based on a flawed premise that the binding constraint holding us back from returning more capital to shareholders is the regulatory framework that we have from the Federal Reserve.
If you look at our track record where we've returned more than a third of our share count since 2010 and have plans to return $25 billion this year and next year, I think it's pretty clear that we're in a different situation than some of the banks, for instance, that have been still worried about the effect of their leverage ratios, or the pure play life companies that have almost twice as much leverage as we do. What is our binding constraint? It's largely rating agencies, as we've said. If you look at, say, S&P's capital model, that's a better way to think about how third parties think about our capital adequacy. The way we think about it is our true north, recognizing that our customers depend on us to make large long-term promises.
Capital regimes, whether it's Solvency II, whether it's state-by-state regulators, whether it's rating agencies, inevitably look at things through a broad brush stroke. We have more inside knowledge as to the risk profile of our business than anybody, and we have to hold ourselves to account to make sure that we have not only enough capital, but it's in the right place. Is it in a subsidiary or is it the holding company where it's most fungible and deployable to where it's needed? This idea of capital and liquidity fungibility is critical for us to be able to survive any eventuality, any stress that we throw on ourselves, let alone which a third party may come up with.
I think that without naming names, the most recent core challenges and events have demonstrated that staying focused on the fundamentals is perhaps the right thing to do from a point of view of priority. Can I imagine a world in the future where being a SIFI becomes a burden and that we need to explore our options? Of course. To anticipate that long before the capital rules have been actually determined, and while we've got serious work to do to improve the fundamentals of the company and return excess capital to shareholders just feels like a diversion of resources that isn't appropriate. We're going to stick to our guns on that. We'll watch very carefully what happens in the overall regulatory environment. Let's not forget, we have almost 200 regulators, every U.S. state, every foreign jurisdiction, and in some jurisdictions, two regulators.
It's not like you remove this label and suddenly you're an unregulated company. We are constantly navigating these multiple constraints in a way that keeps us focused on our true north, which is being really well capitalized to serve our customers. That is our principal stakeholder that we have to be there for, and all of these other components are really third-party validation that we are doing that, whether it's rating agencies, regulators.
Great. Thank you. Audience. Richard.
Peter, could you please discuss what you see as the best growth opportunities in the life and retirement businesses?
I think that for a long time, the life businesses around the world have enjoyed high interest rates, and therefore, too much of their activity has been around a gathering of deposits, effectively. Recycling money. As we look forward to a lower interest rate environment, perhaps for a very long time, I see a return to the roots of protection provision, where we really understand what risks, whether it's longevity risk or whether it's disability risk, that is something which an insurance company is uniquely qualified to understand and pool in efficient ways. To do so, a lot of work has to be done to reduce the friction cost of distribution. I think the overly complex product designs of the past have inhibited a streamlining of distribution.
To give you a tangible example, our term life business today is growing 5% year-on-year, and is winning in the direct market as the number one company in searches on the Internet. I think that simple products distributed efficiently that are addressing the risk factors that insurance companies are uniquely qualified to address, as opposed to being an asset-intensive savings vehicle that has a lower return on equity.
Excellent. One other question on growth. You mentioned growth earlier. When you think about commercial lines definitely retrenched as re-underwritten the book. How far are we away from thinking about growth potential in that business?
The headline number is the net premium written, the net premium earned will lag that, we need to watch both. You'll start to see a bottoming out in the middle of next year. As we've sort of laid out the dispersion in the commercial insurance lines, the sort of product set one, the 15% most profitable, have shown good growth, double-digit growth in some cases. In the personal insurance space, we're getting double-digit growth in the high net worth sector, which is good. I really want to emphasize that we're not targeting top line at all. We're targeting valuable growth, which we measure in the property casualty space with what we call production RAP or risk adjusted profit. Then on the life and retirement side, using value of new business.
It's really ensuring that we're not only getting volume, but it's volume at a comfortable ROE above our cost of capital and growing what we call a RAP, which is a spread between the two, ROE minus cost of equity times the equity employed. If the market is not there, we will not deploy the capital, if it's there, we will grow faster. We're investing in the technology that allows us to scale up and scale down. We've shed a lot of the fixed cost on the distribution side, on the life and retirement side, so that we're almost exclusively third-party distribution today. That gives us a lot of operational flexibility to dial up and dial down the volumes to be responsive to different market environments.
To illustrate that, you look at fixed annuities, we've been as high as $12 billion in a year at as low as two and a half. We're not afraid to dial up and dial down volume depending on market circumstances. We'll do the same thing in P&C.
Great. Thanks.
Pat.
Gary, you mentioned having sold nine businesses over time, and clearly had net one out there on UGC. As you look at the maybe some businesses that you now have and you look at the way the public markets put some valuation on, AIG-related valuations for different parts, have you had enough time to kind of assess far the road for the business? Do you feel like you have a strong view of that? While those two processes take time to play out, are there still larger businesses that may be candidates for-
I'm not going to comment on UGC because the S-1's filed and it speaks for itself. What I did mention in our January strategy update is that we are looking at segmentation of our business in a way that makes it easier for us and for the investor community to understand the profitability and capital allocation by subsegments. We're starting with nine publicly disclosed modular business units. For internal purposes, we break it down in a much finer way than that. Importantly, not just look at businesses, but look at parts of the value chain that we participate in. Does it make sense for us to do our own claims, use third party administrators in this country or this line of business? Make sure that we're making intelligent use of our capital to participate where we can really make a difference.
We've made some divestitures where there was a strong bid. I think that some of the sort of simpler analytical reports that I've seen on this topic make certain assumptions about multiples available that don't scale up to the scale of businesses that we have. Let me put it that way. We have a very sizable company with large blocks of capital that would overwhelm some buyers that might be more interested in smaller properties. We need to be careful not to extrapolate. We've made it very clear that we are open to offers where it makes sense, but we don't see any great benefit from advertising broadly what's for sale and at what price.
We think that we are, as I've said, sculpting the company, and therefore thoughtfully retaining the parts that make sense, adding where it makes sense with bolt-ons, creating joint ventures where that makes sense, but also shedding where that's a distraction. As you can imagine, as we narrow our focus in personal lines from 80 countries to 15 countries, there are a number of countries where we do personal lines where a third party would be a better owner than ourselves.
Anything else? Go ahead.
When you talk about investing in technology and insurance as a service, can you talk a little bit about long term, where you hopefully will see the most benefit, and is this an arbitrage to your competitors or how are others tackling that?
I think that like many industries, the changes in technology are altering the competitive landscape significantly, tilting the relative advantage of global players versus local players armed with world-class technology. Software-as-a-service empowers small companies to operate at efficiency levels that approach large companies. As a company, we need to be very aware of disruptors. Most of whom will fail, but some of whom will succeed, and we need to look at them through the lens, not of why they might fail, but why they might succeed to make sure that we are constantly revisiting some of our assumptions and never become complacent about how we've won in the past if we want to continue to win.
Yes, it's a very competitive market, but as industries go, partly because of the complexity of the regulation, there is an advantage to companies that have deep understanding of how regulation affects the way in which you can compete. For instance, personal data can allow you to do a lot of segmentation that would not be allowable in many markets. How do you get that balance right of using data intelligently to segment risk, to avoid risk in a way that is viewed by regulators and the societies that we operate in as a force for good, helping to improve safety, as opposed to using an information advantage at the expense of consumers. It's really important to think about how these innovations can be used in a sustainable way. Of course, the efficiency gains that come from artificial intelligence is transforming many industries, including ours.
As we look at how robotics and artificial intelligence is being applied at different parts of our value chain, that's really substantial.
Mike.
On the property and casualty side of the business, over the years and even more recently, AIG has exhibited a lot of reserve deficiencies. Maybe you can speak to what you're doing to change, if anything, and make investors feel comfortable that we won't be seeing those large deficiencies on a go-forward basis.
You're right. There's been substantial strengthening of reserves, most recently in the fourth quarter of last year. If you look at those over the last six years, the vast majority, over 90%, relate to accident years 2004 and prior. These are long-tail casualty, usually U.S. casualty lines, where this is what this management team inherited. I think that we have done a considerable amount over the last seven years to understand the exposures and break down the IBNR by accident year, by line of business, get third party validation from experts that are objective, and to also acknowledge this is not a precise science.
To estimate the error terms on all of those estimates, and then aggregate those error terms to get confidence around the aggregate reserve of over $60 billion to ensure that we give investors and ourselves as few surprises as possible, and to make sure that we have the right incentives in place to price longer tail lines with a proper accountability for the long-term reserve development. The investments we've made in analytics is getting us more certain. We had a recent reserve development announced in the last quarter for over $100 million for Florida workers' comp. That's because there was a court ruling in Florida in the second quarter, brand new news, and we added to reserves. The industry estimate is about $1 billion. We're the first company, I believe, to acknowledge that.
Our policy is to surface good news and bad news as it happens, not try and commingle it and smooth it. We think that that's the best we can do to build trust that our reserving is objective and timely.
Peter, on that topic. Some of the adverse development you did have most of those four prior, but there was some more recent year adverse development. I'm wondering if you had a change in the philosophy or methodology on how you're setting some of your P&C insurance company reserves at the end of last year, as we look forward.
That's a good observation. Yes. Some of the strengthening in the fourth quarter was for more recent accident years, and that was largely based on insights gained from even older accident years, where we had enough time to observe statistically significant trends that we could then map to the more recent underwriting years. It wasn't necessarily loss development on the recent years that made us change our mind.
Just by analogy, saying that what we've learned with the passage of time on similar lines, on older accident years, has better informed our reserving judgments on those recent accident years. We felt that was a prudent thing to do, and we recognized that investor confidence is going to rest on our ability to minimize the number of surprises going forward.
Great. Thanks. Audience? I've got one. Since you set your normalized ROE objective in January, interest rates are clearly a fair amount lower. I guess, two questions. One, what's your outlook for interest rates personally, and what do you think is going to happen here going forward? And number two, does that change at all your ability to achieve that normalized ROE objective, and what can you do to manage around that?
Well, despite having managed one of the largest fixed income departments in the world when I was at JP Morgan, I've always tried to resist ever predicting interest rates. I'm not going to jump on that piece of bait at the beginning. Who knows what's going to happen to interest rates. I suspect they will be lower for longer, but I also think that there's a decent risk of a surprise the other way, because anytime consensus builds around one thing, it creates the conditions to be surprised the other way around. We need to be prepared for both. In terms of our exposure to interest rates, we added some disclosure in our second quarter call, which we hope is helpful. Importantly, we break down the problem into the in-force and new business.
On the in-force, we have interest rate sensitivity in a number of areas. One is in the U.S. casualty book, where we have long tail lines, but even longer tail assets against them, so longer duration assets. There, on balance, low interest rates will help because we will get a capital gain on the bonds that exceeds the reduced embedded value on the casualty side. On the life side, you have the reverse. You have very long dated liabilities and medium-term bonds. From an earnings perspective, we estimate that the change in interest rates, roughly 100 basis point reduction since January, has reduced our expected earnings in 2017 by $250 million to $350 million of pre-tax income. From an earnings perspective, manageable, and we're not revising our targets for 2017 as a result of that because we think we've overachieved in other areas.
We also made another important disclosure in the second quarter, which is that our estimate is that the change in interest rates since January to today have reduced the intrinsic value of AIG by about 5%.
That's looking beyond just two years' worth of earnings. It's looking at the long-term asset liability mismatch of both the P&C and the life side, integrating it, then also looking at new business volumes and recognizing that lower interest rates hurt the value of some new businesses which depend on net investment income. We are very sensitive to the interest rate and spread environment in setting goals for both the life, the retirement, and the property casualty business. The volume of new business will certainly reduce in the business lines where net investment income is the main source of value. Therefore, the balance sheet will shrink, and therefore, capital will be ultimately available for return if that's where we head with persistent low interest rates.
Got you. Anything else from the audience? I've got another one. On your use of reinsurance. The use of reinsurance is going to help you achieve your 600 basis points underlying loss ratio guidance. However, is it your objective to actually achieve that 600 basis points in a gross basis? Because you can't always rely on reinsurance to be there to help your net results. We could have for a market, a big hurricane, something could happen. Right. How do you think about that as a CEO and kind of managing your expectations there?
The main change in our reinsurance strategy was a sizable quota share transaction announced in the first quarter. It's a two-year transaction that deals with the over-concentration we have in U.S. casualty business and satisfied the risk appetite of Swiss Re, our counterparty, for more exposure to a sector that they are underexposed to. It's a win-win for both firms. Think about it as a virtual partnership. They share in the gains and losses. They contribute a ceding commission that covers our expenses, and so it's accretive to our ROE, accretive to theirs. It allows us to reshape our portfolio to a better diversified mix of earnings without interrupting any client relationships. That's the beauty of the reinsurance. It's behind the scenes. We don't have to change our capacity available to the market at all in those lines.
Yes, if the market hardens in the reinsurance space, the renewal risk goes up, but so would our pricing power in the primary market. We don't see that as a huge risk at all. What we see is it's a very helpful way for us to rebalance our portfolio with minimal disruption to client relationships. It contributes about two out of the six points, or will, once the benefit has earned into the results. It's still less than two because it's going to come through the net premium earned as opposed to net premium written. You'll fully show up in the first and second quarter of next year, not right now.
Got you. One other kind of more strategic one on the P&C business is, you highlighted the relationship with Two Sigma and moving more into the SME space. Is it your desire to have a portfolio of businesses in the commercial line space that I would call more frequency, less severity, kind of as a part of the mix over the longer term? Are you happy with the mix, more severity, frequency? Do you think about it that way?
We have businesses right on the spectrum. High frequency, low severity. Our travel insurance business, lost suitcases.
Yeah.
High frequency. Canceled flights, high frequency. That today, when a flight gets canceled, no claims adjuster is involved. The check or the wire transfer goes automatically based on the computer system of the airline. Those high-frequency lines are all about expenses. How do you become super efficient at doing that? At the other extreme are satellite launch insurance. That's a severity business, and it's something we're a leader in, something that uses deep underwriting expertise, relationships. The spread of business that we have and the breadth of it is, to me, one of the strengths of the company. We have to know how to adapt our underwriting behavior across that spectrum to embrace the benefits of data and technology and know where that appropriate human versus machine boundary should be set, depending on the reliability of the tools.
The high frequency, low severity space is obviously highly statistical, and there the machine does most of the thinking. At the other end of the spectrum, there's a lot of room for human judgment.
Just as you'd expect a surgeon today to use the very best medical technology to inform how they do their surgery, likewise, we expect our best underwriters to use the very best analytical tools to supplement their judgment. That boundary is shifting as technology gets better. Even in the high-severity lines, you see that use. When you look at our highly engineered property, that's a high-severity line. We have a per property limit of $2.5 billion, which dwarfs the next biggest player in the industry. We're very much involved in using the combination of expertise and balance sheet to add value. Just those of you concerned about that number, we have reinsurance behind it as well, so that we don't have ridiculous concentration risk ourselves.
It's the reinsurer's faith in our underwriting that allows us to put that kind of capacity at the disposal of our clients.
Great, thanks. Got one other one for you here. I don't think we would've actually asked this question a year ago, but now I can ask it. If you achieve or more than achieve your profitability targets on the commercial lines loss ratio improvement and margin improvement, does that in turn translate into potentially more capital management than the $25 billion? Is there a correlation between that?
We've always said that our goal is to exceed $25 billion, and I'll restate that.
Okay.
We've also always said that's a combination of buybacks and dividends.
Yep.
We look at the relative value of those two avenues. We also recognize that the pacing of that capital return matters. We've set ourselves the end of 2017 as a deadline.
Yep.
We're ahead of schedule. We are very sensitive to various factors along the way. What might make it slower? Big nat cat occur. We might want to preserve some capital.
Especially if the market gets hard after it. What might make us accelerate it? If we are more successful in Legacy dispositions than originally planned. We are using that 25 as a target. We've set a deadline at the end of 2017 to do it. It's part of the backdrop of the broader strategic shift that I talked about, which is become a less capital-intensive business model and more expertise based. We're still going to have one of the biggest balance sheets in the industry. We just need to use it much more selectively than we have in the past.
Great. Thank you. We got time for potentially one more question from the audience. All right. If not, Peter, thank you.
Brian.
That was very enlightening. I really appreciate that.
Thank you very much. I appreciate it.