Good day. Welcome to AIG's third quarter 2017 financial results conference call. Today's conference is being recorded. At this time, I'd now like to turn the conference over to Ms. Liz Werner, Head of Investor Relations. Please go ahead, ma'am.
Thank you, Derek. Before we get started this morning, I'd like to remind you that today's presentation may contain forward-looking statements which are based on management's current expectations and are subject to uncertainty and changes in circumstances. Any forward-looking statements are not guarantees of future performance or events. Actual performance and events may differ possibly materially from such forward-looking statements. Factors that could cause this include the factors described in our first, second, and third quarter Form 10-Q, and our 2016 Form 10-K under Management's Discussion and Analysis of Financial Conditions and Results of Operations, and under Risk Factors. AIG is not under any obligation and disclaims any obligation to update any forward-looking statements, whether as a result of new information, future events, or otherwise. Today's presentation may contain non-GAAP financial measures.
The reconciliation of such measures to the most comparable GAAP figures is included in the slides for today's presentation and in our financial supplement, both of which are available on our website. This morning, you'll have the opportunity to hear from various members of our senior management team, including Brian Duperreault, Sid Sankaran, Peter Zaffino, and Kevin Hogan. The format will follow that of past calls with one question and one follow-up. With that, I'd like to turn the call over to our CEO, Brian Duperreault.
Good morning. Thank you for joining us to discuss our third quarter results. There's a lot to talk about this quarter. I'll share my perspective on how I view the results, the market, and how our business is performing. To begin, this quarter, the industry experienced catastrophic losses that surpassed historic levels and had a deep impact on all of us. AIG's response to the catastrophes highlights our strengths, our market presence, and the undeniable value that insurance provides during these times of great loss. Hurricane Harvey, in particular, hit home for AIG, as we have over 3,200 employees based in Houston. I was very impressed with the execution of our business continuity plans for our Houston-based businesses. Due to our multiple location structure, all these businesses were fully operational within a few days after Harvey hit.
While I can say that our employees are extremely dedicated and work hard for our clients through this event, I am most proud of their compassion in serving their communities. We call them our Houston Heroes, and they share their stories as well as those of others across AIG, including those in Puerto Rico. It is our people who are our greatest asset, and we believe their actions speak to the strength of AIG. To date, we've had over 11,000 claims from Harvey, Irma, and Maria, and our claims team in particular distinguished themselves in this difficult time. The severity of the recent catastrophic events forced AIG and all market participants to reassess appropriate pricing for the risks assumed. We successfully begun to raise rates in the wake of the catastrophes and will continue to do so. We have recently seen other market participants follow across the property market.
In addition to cats, this quarter's results include additional actions on reserves for the 2016 accident year in particular. I previously stated that our reserve process is reasonable and sound, and the assessment continues to hold. You will note that there was no overall change for the accident years 2015 and prior on reserves subject to our ADC reinsurance agreement. However, this was our first chance to really view the 2016 accident year, where we made many changes to our underwriting processes and tools. The 2016 accident year is still very green, but we saw greater than expected claim emergence this quarter and decided to be more cautious on the 2016 and 2017 accident years. Sid will comment further on our judgments in his remarks. The events of the quarter and my observations over the past six months are the basis for my plan going forward.
The challenges in our commercial business are ones that I've seen in the past and will be addressed by clear actions that I describe as blocking and tackling. This quarter marks a base from which I intend to grow profitably. Make no mistake, we have a great sense of urgency and have made changes. We have started with establishing an organization structure that plays to our strengths. We've announced this quarter with General Insurance and Life and Retirement. We will return to a discipline specialization in how we run our businesses. We will have more distinct business segments with clear accountability for the P&L and unit integrity. I've already taken action in creating a loss-sensitive business unit called AIG Risk Management or AIGRM. This business is distinct in how we market, underwrite, and manage risk.
You will see more actions such as this. I'm announcing today that we are reconstituting Lexington as a true standalone operation and best-in-class surplus lines writer. Finally, you heard me mention our distinct global footprint on past calls. It remains a core strength of the company, which we will build upon. We announced earlier this week the addition of Chris Townsend as CEO of International General Insurance to lead these growth efforts. You will find a change in our philosophy on underwriting risk. As you may know, it's not my style to take large limits and retentions of risk. One of our challenges has been the size of limits and where we write in the tower for various classes of business. We will also partner closely with our reinsurers as they provide another valuable set of eyes into our book.
Reinsurance is a capital management tool that allows us to better balance our capital and risk across our global businesses, as well as manage our gross limits. Peter will comment more on this in his remarks. Finally, in General Insurance, I'm declaring 2018 as the year of the underwriter. We made and continue to make great additions in underwriting talent and leadership and are taking actions to position them to perform. We will simplify our structure so that they are aligned with the market and incented to write profitable business. No one has commented more on the importance of data analytics and technology in our industry than me. However, it is the underwriter properly armed with this information that is the central control point in our business. It is important that we get the balance back.
Our use of technology and data will complement the strengths of seasoned underwriters with the skill set to evaluate business on a risk-by-risk basis. I'll take advantage of areas of specialization so that we better deliver risk solutions to the market, building on our underwriting expertise and partnering with our reinsurer. I will expect this will take a bit of time, but you should expect actions and improvements each quarter with a goal of getting to sustainable underwriting profitability. These are the steps to improve profitability and achieve our goal of making AIG better than it has ever been. The benefit of our diversification was evident again this quarter in our solid consumer results. Consumer reported over $1 billion in pre-tax operating earnings this quarter. These businesses continue to provide earnings stability to AIG and value diversity across products and geographies, which Kevin will speak to.
Finally, on the topic of our recent de-designation as a non-bank SIFI. This is an important milestone for the company, we believe it puts us on a level playing field with our competition. This positive action was the appropriate response given our size and risk exposure. Our views on growth, both organic and inorganic, remain unchanged, we will continue to pursue opportunities to grow AIG profitably. With that, I'll turn it over to Sid to discuss the numbers.
Thank you, Brian. Good morning, everyone. This morning, I'll comment on our third quarter financial results and provide an update on our capital and liquidity. Turning to slide four, we reported an after-tax operating loss per share of $1.22, reflecting catastrophe losses and reserve strengthening on recent accident years in commercial. Our consumer businesses delivered another solid quarter of results with a normalized ROE of 10.6% in the quarter and 11.6% year to date. Operating results also benefited from our ongoing expense discipline, with general operating expenses on an operating basis down 11% or over $260 million from the third quarter a year ago. Turning to slide five. The catastrophe losses of $3 billion were in line with our previously disclosed range and were comprised of $2.7 billion in commercial losses and almost $300 million in Personal Insurance losses.
Looking ahead to the fourth quarter, our early estimate of California wildfire losses is approximately $500 million pre-tax, net of reinsurance, and largely in our Personal Insurance business. Turning to slide six, I'll provide some commentary on reserves. First, we have completed our assessment on over 80% of our reserves in total during our third quarter reviews. This leaves a little less than 20% of our reserves remaining to be reviewed in the fourth quarter. As part of our studies, we accelerated approximately $6 billion of reserves into the third quarter from the fourth quarter based on lines that displayed adverse claims trends. As a result of our reviews, we strengthened prior year Commercial P&C loss reserves by $900 million pre-tax, excluding the $62 million benefit from the amortization of the deferred gain on the adverse development cover.
Based on the detailed valuation reviews, there was no overall development on the lines covered by the ADC with Berkshire. At this point, the remaining 20% of lines scheduled for the fourth quarter do not display any adverse claims trends. Our observations of 2016 underwriting and claims trends resulted in the judgments we took on 2016 and 2017 accident years. These lines are still green with a range of uncertainty, and we decided it was prudent for us to further move up in the range. Detailed analyses, claims trends, and an evaluation of our underwriting supported these judgments and varied book by book. In our European casualty business, we saw an increase in large losses driven by an increase in underwriting limits in Continental Europe, and we are taking actions to remediate this book. U.S. commercial auto, including certain program business, continued to experience higher severity beyond our prior expectations.
The reserve strengthening on the program business was due to programs we have terminated that have performed extremely poorly and should amount to only $15 million of premium in the fourth quarter. We also added to U.S. excess casualty to move up in our range of potential outcomes. Finally, we took a reserve action in our Financial Lines portfolio for accident year 2016, largely to respond to increased private company bankruptcy trends and increased claims in not-for-profit. As a result of our detailed reviews, we've also strengthened the current accident year loss picks for 2017 by roughly 4.9 points year-to-date. There is a catch-up this quarter where 3.3 points relates to the first six months of 2017. You will also note that we incurred higher than expected severe losses this quarter.
Year-to-date, our Commercial accident year loss ratio as adjusted is 68.9, as shown on slide seven. While we can't provide insight into an actual fourth quarter loss ratio, a high 60s is a reasonable expectation for a full accident year. As we look at the chart, which shows 2016 and 2017 on a restated basis, we see roughly two points of improvement year-on-year. This is being driven by our mix shift, along with improvements within our casualty and specialty business. However, the significant deterioration within our property attritional and severe loss results is largely offsetting many of these improvements. We also recognized a $284 million operating pre-tax benefit in our Life and Retirement businesses related to the annual review of actuarial assumptions.
The benefit was largely in the Individual Retirement business and reflected lower assumed lapses on fixed annuities, as well as strong equity market performance, increasing our expected gross profits for variable annuities. This was partially offset by reduced long-term return expectations. Our balance sheet and free cash flow remain strong. As shown on slide eight, parent liquidity at quarter end was $6.7 billion. During the quarter, we received approximately $500 million of distributions from our insurance companies, including tax sharing payments. Outflows during the quarter included approximately $700 million of calls and tenders of our debt. In addition, during the quarter, we also repurchased $278 million of our outstanding shares and warrants. With respect to capital, the recent tax did not require us to downstream capital to our insurance subsidiaries. Our current capital ratios at subsidiaries are expected to end the year within our target range.
We continue to have the commitment and discipline in executing our legacy strategy. In late September, we signed a purchase and sale agreement to sell the remaining life settlements contracts, which closed on November 1st. The impact to book value is a pre-tax, non-operating charge of approximately $300 million in the third quarter of 2017 to write down the contracts to their fair market value implied by the pricing on these transactions. The net proceeds to the holding company were $1.1 billion, and a smaller portion also went to our P&C insurance subsidiaries. The sale of the life settlement portfolio has allowed us to fully achieve our planned $9 billion in capital to parent slightly ahead of schedule. We will continue to make progress on our legacy strategy going forward and update you when further transactions occur.
Looking ahead to 2018, we expect $4 billion-$5 billion in dividends from our insurance subsidiaries, excluding any potentially legacy-related activity and subject to our customary approvals. Tax sharing payments from subsidiaries will reflect the impact of the losses we've seen in our P&C companies this quarter. We also continue to carefully assess the potential impacts from U.S. corporate tax reform and are in the process of studying yesterday's bill. Our previous sensitivity analysis on changes to the statutory tax rate still hold. An immediate reduction in the statutory corporate tax rate to 20% would result in a reduction in our deferred tax asset by approximately $7 billion. Similarly, we would expect our ROE to increase by approximately 150 basis points prospectively. In isolation, we view this as a slight positive from a value perspective.
As a reminder, our year-end reporting will reflect the new organizational structure, which will become effective during the fourth quarter. Commercial Insurance and Personal Insurance segments are transitioning to General Insurance, led by Peter Zaffino, and Life and Retirement, led by Kevin Hogan. General Insurance is expected to include U.S. and international results for Commercial Insurance and Personal Insurance. Life and Retirement is expected to include Individual Retirement, Group Retirement, life insurance, and Institutional Markets. In advance of our fourth quarter earnings release, we plan to release a recast of our financial supplement. To sum up, despite the volatility in the commercial business this quarter, we have a clear line of sight into how we're going to improve our results. We have demonstrated the value of maintaining a strong balance sheet and free cash flow profile, which gives us flexibility to execute on our strategic options.
I'd like to turn the call over to Peter.
Thank you, Sid, and good morning, everyone. Today, I will discuss some of my early observations at AIG, our third quarter business results for the Commercial Insurance, and our focused on building an operating structure that will drive long-term sustainable improvements in underwriting as well as our financial performance. Let me express how pleased I am to have joined AIG, and it's great to be working with Brian again. I'm also very pleased to be working with a talented leadership team and all of our colleagues from around the world. I spent the majority of my first three months here as a Global Chief Operating Officer interacting with clients, distribution partners, and colleagues. During my visits to many cities around the world, I've begun to review our operations in more detail and learn more about our ability to provide value and service to our clients and brokers.
As part of the transition to AIG's new operating structure and my expanded role, I've spent the last 30 days becoming more familiar with the underwriting process and claims process for the Commercial Insurance and Personal Insurance businesses. My early observations have affirmed many of the perspectives I had from interacting with AIG as a business partner in previous roles. Our clients and broker partners regard AIG as an insurer with deep risk expertise and very strong underwriting capabilities, along with innovative solutions to deliver on our expansive global footprint. They want AIG to be a strong and active underwriter of risk, and we have received overwhelming support of our plans to position AIG for the future. My second observation is related to the team at AIG. The spirit and collaborative nature of our colleagues, as well as their desire to rally around leadership, is truly impressive.
Nowhere has that been more strongly evident than the remarkable response to the recent natural catastrophes. While these times are devastating for those directly impacted, it is incumbent upon us and the industry to do what we do best and get clients back to business. In my opinion, this is where AIG excels. Even before the events hit, we positioned 500 claims colleagues from various worldwide locations so that they were prepared to inspect locations and adjudicate claims as soon as conditions would allow. We've responded to over 11,000 Commercial Insurance and Personal Insurance claims and have physically inspected 99% of the sites impacted by Harvey and Irma and have distributed funds to our clients. Moving on to my observations of the business. There have been many changes at AIG over the past decade, with strategic priorities shifting under numerous leadership appointments.
It's been hard to focus the organization and make the intended meaningful progress to improve overall financial results. Our new organizational structure will play to AIG's core strength of underwriting expertise in key segments, connecting our global platform and leading with innovation, making us more responsive to client needs. It will more closely align us with our clients and distribution partners, define clearer accountabilities within AIG, drive more empowerment in the field, and improve the speed of decision-making. We are enhancing our new structure by hiring very experienced underwriters and former colleagues with track records of strong financial performance to complement the talent we have here today. Let me transition to some key areas of our performance. Turning to slide 10, I will speak to Commercial's third quarter results. As Sid mentioned, we will shift to presenting results for General Insurance in the fourth quarter.
Commercial's reported underwriting results reflect the noteworthy events that Sid discussed in his opening comments. Net premiums written declined 13%, excluding FX, with nine points of the total decline driven by continued risk selection efforts, mainly in U.S. casualty and property, and the remaining four points related to the previously announced sales of Ascot and select international operations to Fairfax. On slide 11, liability and Financial Lines accident year results include the increase in current accident year loss picks for U.S. casualty. Risk selection efforts to improve performance in U.S. casualty and market competition drove the quarter's premium decline. Turning to slide 12, property is clearly a focal point this quarter given the significant natural catastrophe losses, and we also experienced an elevated level of severe losses. It's important to note that by and large, AIG did not have any recovery from our cat reinsurance programs.
Our reinsurance agreements were not triggered given the multiple loss events, each of which did not exceed our retention. Having said that, going forward, we will take a long-term strategic approach to reinsurance, working in close partnership with our major reinsurers. During 2018, at a minimum, we will look to take a lower net retention on our property cat book, take less per risk net retention in property, reduce our net limits in certain casualty lines, and look for opportunities to further reduce volatility in our results as we position the company for long-term profitable growth. In 2017, we saw incremental underwriting improvement for Commercial overall, recognized that there's been further deterioration in our property. As Brian mentioned, the third quarter events will be a catalyst for property, and we're taking actions.
We've been communicating the need for both rate and cat and non-cat risks with clients and distribution partners. While we have much to do and it is early, there are positive signs that property rates will improve. Our goal is to achieve double-digit rate increases on a risk-adjusted basis, which will deliver further improvements in the accident year loss ratio on an adjusted basis. I'll wrap up my third-quarter commentary by acknowledging we're not where we want to be in terms of underwriting performance. I came to AIG because I'm highly motivated to contribute to the success of this terrific organization. Together with our leadership team, I intend to run to problems and to develop a plan to fix them as quickly as possible.
As Brian outlined, we now have the key components of the operating structure and will continue to develop and add talent with a laser focus on underwriting excellence. These are the elements that will enable AIG to become an industry leader that meets the needs of all of its stakeholders. Looking forward, our main priority is to allow our teams to successfully prepare for January 1 renewals and not disrupt their efforts as they work with our clients. We expect to have the full General Insurance operating structure clearly defined by January 1st of 2018. We have begun making progress on three major parts of our business, which I'll just cover briefly. The first is our international business, where we just named Chris Townsend as CEO of International General Insurance earlier this week. Chris will lead AIG's international growth strategy.
Our second area of focus is AIG's risk management and loss-sensitive business, a core strength which requires infrastructure that few companies have. We named Bill Rabl as CEO of AIGRM to create best practices and consistency in how we deliver value to our large clients. Lexington is our third area of focus. As the largest excess and surplus lines underwriter in the U.S., we will invest in specialized leadership, expertise, and a distribution strategy that enables Lexington to act more opportunistically in the market. We are working towards a new era for AIG, where we will fundamentally improve risk selection to deliver a better quality book, deploy capital towards the best opportunities for profitable growth, and improve our operational process.
While we are embarking on a journey that will take some time, as the prior action years are now, I'm confident in the choices we're making to deliver a repositioning that will solidify AIG as an industry leader. I look forward to keeping you updated on our progress. With that, I will turn the call over to Kevin.
Thank you, Peter, and good morning, everyone. As you can see on slide 14, consumer produced solid results for the quarter despite higher catastrophe losses. We earned over $1 billion in pre-tax operating income, and our normalized ROE was 10.6%. Higher catastrophe losses were partially offset by lower accident year losses for Personal Insurance and the favorable actuarial assumptions update and higher fee income for Life and Retirement. Strength of the equity markets has continued to help partially mitigate the challenges and impact of the low-rate environment on our results. During the quarter, we continued to take further actions to enhance returns, strengthen our platforms, and pursue targeted growth opportunities. Turning to Individual Retirement on slide 15, regulatory uncertainties and disruption have continued to significantly affect distributors, negatively impacting industry sales, particularly of annuity products.
This disruption, along with more aggressive competition in the fixed annuity space, led to materially lower Individual Retirement sales and net flows from a year ago. In the face of industry sales challenges, we continued our disciplined approach with respect to product pricing, product features, and asset quality. We've enhanced the product design and features for our variable and indexed annuities in a disciplined manner, which has attracted new sales and continued to benefit from our broad product portfolio and diversified distribution network. Individual Retirement assets under administration were at historical highs at quarter end, driven by equity market performance and positive indexed annuity net flows, which resulted in increased fee income. As expected, we saw compression from the runoff of higher yielding assets in the low yield environment.
Commercial mortgage loan and unexpected accretion income impacted base net investment spreads, benefiting variable and indexed annuities in the current quarter and fixed annuities in the prior period. Turning to Group Retirement on page 16, our investments in VALIC to transform the plan sponsor and participant experience continued to pay off. Improved net flows reflected strong sales, including a very high level of new group acquisitions year to date, as well as improved group plan retention. Assets under administration for Group Retirement were also at historical highs at quarter end, driven by equity market performance. Despite disciplined rate management, net investment spread declined, partially caused by higher commercial mortgage loan income in the prior period, also due to the runoff of higher yielding assets and reinvestment in the low yield environment.
Looking forward across Individual and Group Retirement, absent significant changes in the overall rate environment, we continue to expect our net spreads will decline by approximately 1 to 3 basis points per quarter. Let's now move on to life insurance on slide 17. Our life insurance business continued to make progress executing on our plans to enhance ROE and return to growth. In the U.S., our new modern administrative platform, distribution simplification efforts, and narrowed product focus are supporting both top and bottom line growth. We had strong growth in both term and universal life insurance sales. Mortality experience was consistent with prior year and within pricing expectations.
Turning to slide 18, although the headline for Personal Insurance for the quarter was higher catastrophe losses, the normalized results reflected the strategic actions we have taken over the last few years to reposition this business. These actions include a focus on markets and customer segments where we have a competitive advantage and favorable growth prospects. The improvements in Personal Insurance's accident year combined ratio as adjusted reflected lower non-cat losses in the quarter. As I mentioned in our earnings call for the second quarter, we are moving our focus from margin expansion to growth at targeted margins. In Japan, we continue to make progress in our transformation while producing solid underwriting results.
We remain on track for the legal entity merger on January 1, 2018, with the completion of user acceptance testing and the resulting systems cut over in October, allowing us to achieve the milestone of soliciting new and renewal policies on the new legal entity platforms and paper effective January 1st. As this transition continues into 2018, we will have the expense of still operating multiple legal entities, which will have some variability ahead of savings being realized in 2019 and beyond. To close, I'm pleased with the progress we are making across our businesses. I would like to turn it back to Liz to open up to Q&A.
Okay. Well, listen, let me do it. Operator, let's start the Q&A.
Absolutely. Thank you. Ladies and gentlemen, if you would like to ask a question, please signal by pressing *1 on your telephone keypad. If you are using a speakerphone, please make sure your mute function is turned off to allow your signal to reach our equipment. As a reminder, we ask that you please limit yourself to one question and one follow-up question. Again, that is *1 to signal for a question. Our first question comes from Joshua Shanker with Deutsche Bank. Please go ahead.
Yeah. Good morning, everyone. Thank you for taking my question. Working through all the changes with reserve inter-quarter reserve charges and whatnot, I look at the liability and Financial Lines segment, I see about 3, 400 basis points of initial pick improvement from the first half of 2017 to third quarter '17. Given that you had these early emergent loss problems in the 2016 year, is it prudent to be picking the third quarter of '17 so much higher or so much better than the first half of the year?
Well, Josh, I think we're being realistic about our loss picks. There was improvement. I think there is some recognized starting points, but if you go accident year to accident year, there was improvements in casualty, whether it's reducing areas of high loss levels or changing reinsurance, et cetera, or basically getting rid of programs, et cetera. We have made improvements. I think they're realistic, yeah.
The third quarter book looks somewhat different from the first half book, is what you're saying?
Well, I don't know if it's the third quarter. I look at it on a year-to-date basis. That's the loss ratio you got to look at, and that's the one I concern myself with. Say, try to give you a kind of a general idea of what the accident year for all of 2017 is going to be. I think that's really, to me, our starting point, and that's the one I would use.
Your next question comes from Ryan Tunis with Credit Suisse.
Hey, thanks. Just going back to, I guess, Brian's earlier comments, originally he'd said that reserves appeared reasonable. I think he also said this morning that that conclusion still holds. If that's the case, is the level of reserve, I guess, activity we're seeing this quarter, that level of volatility, what investors should be used to or expect on a go-forward basis next?
Okay. Yeah. Thanks, Ryan. To clarify, I think the reserve process is reasonable. When I look at 2016, it's the first chance we get to look at it, really. First real view of it. I consider this more underwriting than actuarial. This is looking at our underwriting and understanding what our starting point was and whether the changes taking effect were as strong as they were, and we're setting that up. We are in a business of volatility. There's no question about it. Quarter to quarter, you're going to see movements. As I said in my remarks, and Peter repeated, we want to dampen some of this volatility through reinsurance to take some of that out and then get the right base going forward.
That's, I think, the real takeaway for you is, that I would like you to understand, is the process is reasonable. The underwriting will get better, we will get these accident year loss ratios down, we'll get the profitability.
I guess thinking about the process of whether it's establishing reserves this quarter or even just losses on the hurricanes, is there anything different about, I guess, the approach that you might have taken in establishing what the right size of those was versus what you think AIG may have done in the past?
Well, look, it's very hard for me to comment on what AIG did in the past. I don't want to do that. What I want to do is say, look, I come here and I look at what we're doing now, and I think it's reasonable. I do. To me, if you've got a problem, you recognize it, you recognize it early, you admit it, you fix it. That's our business. You can't go error free. What you can do is be relentless in the pursuit of the truth and the resulting actions that are required, and that's what we're doing. I think we're on the right track. I really do.
Your next question comes from Jay Gelb with Barclays. Please go ahead.
Thank you. One question, then a follow-up. The first, I just wanted to clarify that as the remaining reserves are reviewed for property casualty in the fourth quarter, and we realize there's around 20% left. Am I right in saying or characterizing what you're saying as don't expect anything meaningful from a development standpoint in 4Q?
I had a feeling you'd ask something like that, Jay. Let me give you a couple of data points, okay? There's about 20% left, whatever that number is, $10 billion, I think, something like that. Here's a data point. 80% of that's Basically covered by the ADC. That's made a point. The other is, as Sid pointed out, we pulled forward anything that we saw that was showing deterioration signs, the rest isn't. I think those are two indications of what the fourth quarter might produce.
That's helpful. Thank you. For Peter, I believe you mentioned in your prepared remarks, expectations for double-digit rate increases in commercial. Are you talking about just commercial property insurance? I don't know if at this point you want to give any expectations for accident year loss ratio improvement in 2018.
Thanks, Jay. Why don't I just comment on what's happening with rate? As Brian mentioned, I followed up, is that the cats have been a catalyst for pushing rate, but the accident years for property have underperformed, I think in the industry, with a lot of rate pressure for the better part of the last four years. I think the lack of cat has allowed the attritional loss ratios to kind of creep up. In the third quarter, we began to push rate, that was on cat exposed as well, non-cat exposed business, and the rates are holding. Just a reminder, our fourth quarter is the seasonally smallest quarter for property, so it'll take a little bit of time. I actually think that the reinsurance market hasn't fully worked its way out either with the retro capacity. We see capital being raised.
That's really to serve a lot of the current portfolio, how that translates into reinsurance and then into primary, we will see. We have been over-communicating with distribution partners and clients and outlining a lot of specific reasons why we need rate, and the early signs are that we're achieving it.
Your next question comes from Elyse Greenspan with Wells Fargo.
Hi. Good morning. My first question, I guess is a little bit of follow-up ties together price and reinsurance. Peter, in your comments, you pointed to purchasing more reinsurance next year with expectations that rates will be higher. Do you think you guys, is the strategy to get enough price on your own business to offset the impact of paying higher reinsurance price, and just how do you envision that all coming together in the margins in your commercial lines book?
I think she wanted you to answer that, Peter.
All right, Elyse. We had a fairly traditional cat program on a per occurrence basis. We're looking at a variety of different options in the market now. We're looking at different ways of structuring the cat. We're looking at different ways to structure our per risk, looking for perhaps some catastrophe cover in our per risk. We're looking at a variety of different structures. We'll be in a better position to talk about it in the fourth quarter. I'm not sure that on our book, we're going to see a significant amount of rate in the reinsurance structure. That's yet to be determined. That's not going to really influence what we're doing on the front end. We're pushing rate because we need to improve our accident years and property.
Can I just add to that? Elyse, let me just add to that. We're buying reinsurance because we have too much volatility, frankly. I want the partnership of a reinsurer to help me look at our portfolio. There's a lot of value in reinsurance, we don't expect rate compression. We don't think that the resulting numbers are going to be out of line with where they should be. We're going to see real value out of the reinsurance. I don't expect that to be a problem. It's going to be a help to us.
My second question. If we kind of think about the property market today and just the expectations for higher prices, just from sitting through and thinking about the softness that we've seen over the past years, how sustainable do you think a hard market could be? Brian, as you kind of think back, is there another kind of prior period you would compare the type of hardening market we're potentially looking forward to on the commercial line side?
Elyse, that's interesting. I've been around a while, seen a few hard markets, but there haven't been that many. We're not talking about 20. They're all different. They all start out different. They all end different. Obviously, it starts with the fact that the pricing, whatever level, isn't any good. We certainly have that aspect. I think a lot of times it's a withdrawal of capacity as opposed to a loss because the catastrophe itself took it away. They tend to be localized. In my experience, you can have a property hard market, and you don't have a casualty hard market. The 9/11 time was a bit different. I think the market is recognizing that the property levels are unsustainable, number one. Number two, the property market, I believe, has been living off of the cat profits. They don't exist.
Number three, I think there is a real question about whether the models for cat are hitting the mark. With global warming, et cetera, the things that we're seeing, wildfires in California. I think there's maybe a question about the whole level of exposure that we have. All those are great making for a market. How long it lasts? I guess it lasts as long as it lasts. How long is a piece of string? I don't know. I do think it has all the ingredients for a significant improvement in the property because the property has problems at the frequency level, it's got problems at the severe loss level, and it's got problems in the cat. All should be fixed. The interesting thing is the market. We'd have to do it anyway.
We started taking actions and raising rates and taking care of our exposures, and the market is doing exactly the same thing. We definitely see a sign of a market change here.
Your next question comes from Kai Pan with Morgan Stanley. Please go ahead.
Thank you, and good morning. My first question on the underlying loss ratio trends, if you look at the recast trend, improve about two points, like to 69% in the first nine months of 2017. Given the underwriting action you're taking right now, I just wonder, the pace of the improvements at two points, do you see that will be slower in 2018?
No, I don't think it'll be slower. I don't think it'll be slower. No.
Okay.
Because we're addressing non-property lines with the actions they require. Risk selection is the number 1 action taken. We've already talked about the property, no, I don't think it'll be slower.
Okay, great. My follow-up question on capital management. You purchased about $275 million of shares in the third quarter. Is that level that investor can expect going forward? How do you use the $45 billion sub-dividends going forward each year?
I'm not going to give you an expectation of what we might do in share buybacks. We bought back shares. We may buy back shares again, but I don't know. I won't tell you when or how or what, but it is a capital management tool. We've used it. We do have capital. We want to deploy that capital in our business. That's primary. I believe we have places that we can deploy to make improvements in our portfolio and long-term structure. That's the number one priority.
Next, we'll hear from Tom.
Okay, next question.
Thank you. Next, we'll hear from Thomas Gallagher with Evercore.
Good morning. Two questions, Brian. The first one just on the SIFI de-designation. Given when you read the Fed's commentary about how much you shrunk the balance sheet as the rationale for the de-designation, does this inform your view of what types of deals you might consider from an M&A perspective? Meaning, would you consider deals that added a lot of balance sheet, given what the Fed said in that process?
Well, that's interesting, Tom. When I do a deal that adds to the-- well, I think most deals do add to the balance sheet, so that's not my first thought, though, obviously. My first thought is it strategic? Is it something that makes us better? Do we get better people or better technology or better market penetration? Or are we in places we don't? Anything we do would make us larger, but I want us to be better and better balanced and have more diversification on our earnings stream. That trumps anything else.
Got you. My follow-up is, when I think about the message at AIG in 2015 and 2016 it was getting far more conservative and shrinking the casualty book. Now you have all this adverse development on the 2016 accident year for the long-tail lines. I guess what I ask myself, is it possible that there was a lot of adverse selection there as AIG was executing the pivot? Is that part of what's going on here, or do you think that's not the right observation?
Tom, look, there's probably some of that. I would say that selection trumps price. Let's start with that, okay? If one takes action and it's price-oriented, you could suffer some risk selection issues if you're not thinking selection first. I try to emphasize in my remarks that risk selection is our first priority, and we will price the products that we like appropriately. If there is some adverse selection taking place, we're going to root that out.
Your next question comes from Erik Bass with Autonomous Research. Please go ahead.
Hi. Thank you. With the issues we've seen in the 2016 and 2017 books, it's clear, I think as you alluded to, that AIG's continued to have some underwriting challenges. Based on what you've observed, what do you think is driving this, and can you provide more detail on what you're doing on the underwriting side in terms of talent, processes, data, et cetera, to improve the results?
Yeah. Look, I think there are some actions, I don't want to denigrate all the work that was done here. There is some improvement. I think back to what we said, Peter and I said about the commercial business, we attacked the structure because I think the structure was not recognizing that specialty matters. There was a little too much generalist, not specialty. We imposed, and rightfully so, better underwriting tools to help underwriters get better. We have to have the underwriters make the ultimate decision. I think the pendulum might have swung a little bit towards the science and less towards underwriting and selection. We want to restore that balance. I think, I know we're in the volatile business. We take a lot of volatility. Our gross lines, I believe, are too high. I believe our net lines are too high.
When you have issues, they're extreme. If you make a mistake, it's exacerbated. We want to cut those extremes down. I think we want a better balance. You got to pick your spots where you can make money, and where you can't, you should reduce. Some of that took place. There's a greater, maybe an enhanced sense of urgency to do that. Yeah, we have good people here, but there are places where we need good people that we're getting from the outside. It always starts with people. You put a good structure together, I think we've got a great structure. You put good people in that structure, you empower them, and you hold them accountable, and this is very doable. I've seen it before. There's nothing I haven't seen before.
This is all basic execution, it's execution that will win the day here, we are executing.
Thank you. Appreciate the comments. Just from the specialization of the business units, what does that mean from a practical basis, and how do you see that improving the results?
Well, we talk about AIG Risk Management. That's a business in our marketplace that comes in in a certain way, has to be handled by specialists. We have diffused that and spread it across the business. Now we've created the unit. Again, we used to have one, we created again. Lexington. Lexington is a surplus lines company. It requires a separate unit with people who are specialists in surplus lines work. We had taken that business and kind of spread it into our branches, people were doing both. That's not a good way to structure it, we've gone back and we're creating the Lexington. Peter, maybe you want to add something to that.
Yes, Brian, I think just to add to what you were saying is that by creating these distinct units that have very strong capabilities within their segment. If you look at AIGRM, the ability to share best practices across the loss centers of a risk management business is going to yield better results. We are doing a bottoms-up review, Bill Rabl is going to look at, there's a finite amount of risk. There's 800. We're going to do a grounds-up view of risk selection, attachment points, using all the tools, data, and capabilities that we have to price the business. We're really going to get into the details and focus on underwriting and improving the book.
Thank you. Your next question comes from Larry Greenberg with Janney. Please go ahead.
Yeah, thanks. Just following up on underlying loss ratios. I just want to be sure I'm understanding this, that usually when we hear about underwriting actions and how the accounting works in this business, it usually takes one to two for it to flow through the bottom lines. Am I understanding that with reinsurance, with cutting back limits, with some of the actions that you or your predecessors might have taken to date, that we can expect or should expect an improvement in 2018 levels?
Well, one would hope our 2018 has improved over 2017. That's the actions we're taking. Yeah, reinsurance can speed it up a little bit, I suppose. Things still have to earn out. 2017 is going to earn out into 2018. That's just life. Look, there's nobody that wants to get this done any faster than me. I know you're impatient. I'm even more impatient than you are. There is a sense of urgency here. We'll get it done as fast as humanly possible.
Great. Then the volatility in the larger severe losses in the property line. Is there any way to kind of give us a tolerance level for severe losses or how we should be thinking about that prospectively as you cut back limits?
Peter's going to answer that.
Yes. I think, Larry, back to what we had talked about on the risk appetite for property per risk. When we were re-underwriting the book, I think we took some very good actions to try to improve the overall portfolio. I believe, focusing on areas of key segmentation and continuing that evolution. I think it'll be a better chance for us to update in the next quarter when we have an understanding of our per risk. I think our tolerance is going to decrease, some of those spikes that perhaps we've seen in the past will be leveled out. I think that's what Brian was referring to in lessening the volatility that we're seeing on some of the losses, even in casualty, where we just have too big of net.
I don't think you can draw to one specific reason as to why it's occurring other than we want to protect the balance sheet and protect our results through reinsurance purchasing above severe.
Thank you. Your next question comes from Brian Meredith with UBS.
Hey, good morning. Two quick questions here for you. First, if I look at the general operating expenses down 20% year-over-year in the commercial line segment. Brian, Peter, I'm wondering when are we going to start seeing the investment in the business that you talked about start to kick in on the expense side? Could that have some offset with respect to the margins in the business as the loss ratio maybe goes down, you'll start to see the expense ratio, at least in the interim, kind of move up here?
Yeah. Investment is we're investing in people. I said there's good GOE and bad GOE. That's a good GOE.
Right.
We're investing in people. I said this the last call. You could see some offsets to reductions over here with increases over here, but we're adding to the strength of the company. Ultimately, that's a good thing. We have to be expense conscious in this business. We, as a company, still have an expense level that's in excess of our peers, and therefore, in excess of what the market will bear. We've got to address that. There's no question about it. There's different ways to address it, but getting better people in has to be a priority, period. Sid, you want to say something?
Yeah, Brian, I'd just remind you, one of the areas that we've been most aggressive on and that has contributed to our expense discipline has been areas such as professional services fees, overhead IT infrastructure, real estate, and we're going to continue to attack those areas, I think with a tremendous amount of discipline, both at AIG level and in commercial.
Right. We're going through this, we mentioned the structure. The structure calls for what I call unit integrity, more specialized units, more of them. Technically speaking, that probably adds a little bit to expense levels. My experience is when you put someone in charge, give them a unit, they squeeze the living daylights out of expenses. I think ultimately, that's another great way to get our expenses under control. Give people responsibility for a P&L, and you'll see expense discipline.
Great. This next question quickly, just on the share buyback that we saw in the quarter, kind of take it a little bit different way here, approach this, Brian. $250 million, not a whole lot, not that meaningful. Why buy back anything? Why not just keep the cash in your balance sheet for these investment opportunities?
I got to say, what a great question, Brian. I didn't expect to say, why are you buying any stock back? This is interesting. It's a capital management tool, and we deployed it, and I'll continue to deploy it. Look, I don't have anything against stock buyback as a management tool. If I can find something better to do with the capital, you know I better be doing that. That's what we get paid to do. No, we make the decisions as the moment arises, and we'll continue to do that. Look, I think I'm being told that we should wrap up. Operator, I'm going to just do some closing remarks now. First of all, thank you investing public to listen to the story. I think just be confident in our efforts. We will right this ship.
It's started already, and I am a very optimistic man about this, I really am. I really want to finish by thanking the employees. Look, we've gone through a very difficult quarter, catastrophes that have affected us as a business and affected us personally, and I can't be prouder. Thanks, everybody.
Once again, that does conclude today's conference call. We thank you for your participation. You may now disconnect.