Good day and welcome to AIG’s Fourth Quarter 2018 Results Conference Call. Today’s conference is being recorded. At this time, I would like to turn the conference over to Ms. Liz Werner, Head of Investor Relations. Please go ahead, ma’am.

Thank you and good morning, everyone. Before we get started, 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 2018 Form 10-Q and our 2017 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 expressly 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, which are both available on our website. This morning, you will have the opportunity to hear from members of senior management, including CEO, Brian Duperreault; CFO, Mark Lyons; the CEO of General Insurance, Peter Zaffino; and the CEO of Life and Retirement, Kevin Hogan. Joining us on the room are others in senior management, including our Chief Investment Officer, Doug Dachille. This morning, we’ll allow a few extra minutes at the end of the prepared remarks for Q&A. We’ll continue to follow the same format of asking one question and one follow-up and ask that you then get back into queue for additional questions. Thank you. And with that, I’d like to turn it over to Brian.

Good morning and thank you for joining us today. In addition to reviewing our 2018 fourth quarter and full year financial results, I’ve planned to spend a good part of my remarks, reflecting on my first full year as CEO of AIG. Throughout 2018, we uncovered many issues and challenges and took actions on a number of fronts to lay the foundation for long term sustainable and profitable growth. We reorganized and repositioned our businesses, significantly reducing risk and volatility, recruited industry leading talent and fostered a culture of underwriting excellence and accountability. The problems at AIG were deeper and more pervasive than I originally anticipated, but we have put the right people and the right strategies in place that will allow us to accelerate our progress in 2019. As Mark will discuss in more detail, our fourth quarter results were significantly impacted by a decrease in net investment income, due to volatility in both equity and credit markets, particularly in the latter part of December, which impacted the operating results across our businesses. Our results were also impacted by cats, however, these losses came within our previously disclosed range. In addition, we reported 365 million of unfavorable prior year loss reserve development, driven mostly by underwriting decisions from 2016 and prior years. We are pleased that the fourth quarter also showed positive improvements in general insurance, reflecting actions we took in 2018 that demonstrate that we are on the right path to restoring this business to profitability. For example, the accident year combined ratio ex-cats for the fourth quarter was 98.8%, a 140 basis point improvement over the fourth quarter of 2017, largely driven by an improved loss ratio of 63.9%. The expense ratio of 34.9% was 90 basis points better than the third quarter of 2018 and we…

Thank you, Brian and good morning, all. This morning, I’ve planned to go over the key financial impacts for the quarter as well as an actuarial discussion around prior year loss reserve development. I'll begin with some of the largest or in some cases, most notable financial impacts occurred, namely investments in legacy, then turn to general insurance and life and retirement. Despite volatile financial markets and their impact on investment returns and by consequence business segment results, there are clear green shoots that bode well for the future and I will highlight them as well. Turning to slide 4, on December 5, AIG provide some guidance at the Goldman Sachs Conference in New York and I'll now provide some color on the actual variances from that guidance. Consensus fourth quarter adjusted after tax earnings per share was $0.42, post the Goldman Conference whereas AIG reported a $0.63 per share adjusted after-tax loss, representing a $1.05 variance. All per share variances comments that I will make will be utilizing the consensus tax rate of 25.4%. This permits each variance discussed to be a pure impact without any tax interaction. Instead, I will isolate the pure tax impact for consensus separately later in these comments. So, the major items comprising $1.05 per share variance are as follows. First, a $0.38 per share difference associated with capital markets volatility and by this we mean, net investment income differences for general insurance and legacy and differences above credit and interest for life and retirement, along with policyholder and advisor fee reductions, net of associated offsets like DAC and advisory fee expenses. Included within this is an $86 million pretax impact or $0.07 per share after tax impact associated with hedge funds and Japanese equity marks, usually recorded on a one month lag that…

Thank you and good morning. Since Mark provided a detailed financial information for general insurance, I will expand on Brian's prepared remarks and highlight some of our more significant accomplishments in 2018. This will include inside on the progress we have made to overhaul our core underwriting capabilities and reinsurance program. I will then give a brief overview of market conditions and summarize our view, entering 2019. As Brian noted, a significant amount of foundational work was completed in 2018, as we repositioned general insurance and worked towards creating a culture of underwriting excellence. Our principal focus was to outline a new underwriting risk appetite, improve underwriting capabilities and create business units that can positively distinguish themselves in the market. Key components of this work included a rigorous review of the entire portfolio, an aggressive limit reduction effort and embedding an enhanced governance and control framework. Detailed portfolio reviews uncovered significant complexity and exposures, resulting from the prior large limit deployment strategy that made AIG an outlier in the industry. I'd like to provide you with more detail on our key areas of focus. First, with respect to the general insurance organizational structure, we established or reestablished numerous business units, focused on the strategic positioning for each business within the portfolio and empowered our leaders with end to end accountability to drive results. We also filled critical positions, including Chief Underwriting Officer, Chief Actuary, Head of International, Head of Claims and the CEO of Lexington. Our new leadership team is working to lead the company closer together and we're building a strong bench to execute on improving our underwriting performance. The new organizational structure was designed to better position us to serve our brokers and clients and allowed us to outline a more coherent risk appetite and underwriting strategy. Throughout…

Thank you, Peter and good morning, everyone. As you can see on slide 12, life and retirement recorded full year adjusted ROE of 12.6%, consistent with our long term expectations. The year-over-year change from the annual actuarial assumption update unfavorably impacted ROE by 156 basis points, which was more than offset by the favorable impact of tax reform and reductions in attributed equity consistent with our ongoing de-risking. Adjusted pretax income of 3.19 billion represents a $641 million reduction from the prior year. The primary drivers of the year-over-year difference were the annual actuarial assumption update, which accounted for 382 million of the decrease, lower returns on fair value option securities of 114 million, higher deferred acquisition cost amortization of 31 million, higher new business expenses of 33 million, 17 million for modernization investments and 39 million for miscellaneous other items that lowered 2017 expenses. Looking ahead to earnings in 2019, having largely completed our portfolio repositioning, we expect adjusted pretax income to be essentially flat with full year 2018 with a consistent adjusted ROE. These expectations assume a modest improvement in equity markets from year end. In addition, absent significant changes in the overall rate environment, our current expectation is that base net spreads will decline by approximately 0 to 2 basis points per quarter at least through 2019. We also expect lower yield enhancements as interest rates rise. Our 2018 results reflect strong growth from our ongoing strategy to leverage our broad product portfolio and diversified distribution network to satisfy customer needs. As pricing conditions improve in the second half of the year, we significantly increased sales at indexed and fixed annuities. With growth in individual retirement, group retirement and life insurance, we increased total premiums and deposits for the year. We maintained our disciplined opportunistic approach in…

Thanks, Kevin. We gave you a lot of content, which used up a lot of the first hour. So we will, as Liz said, we will go a little long on the Q&A. So operator, let’s go to Q&A.

[Operator Instructions] Our first question comes from Kai Pan with Morgan Stanley.

You guys have reaffirmed the guidance for 2019. I just want to make sure, like including the underwriting probability in the first quarter of 2019, from the 98.8% underlying combined ratio, you need more than 3 point improvements, assuming a 4.5% of AAL to get to that level in the first quarter. I just wondered, how do you bridge the gap, the 3 point improvement in one quarter? Is that mostly from expense evenings or from the underlying loss ratio?

Well, I think it's going to come from a variety of sources. We thought you might ask this question. So Mark is going to answer and Peter just probably on the supplement.

Yeah. We can tag team that. Hi, Kai. So let me address that in kind of two ways, a conceptual approach so you kind of -- you get it and a partial numerical approach. So, first off, if you look at how the book has changed and you look at some of the issues, all areas of the book have gotten a lot of focus. Now, North American commercial is a book that, as I said before, a lot of recurring pain points, it’s probably got a little more focus on it. But good kind of improvements. So when you look at fourth quarter of last year to fourth quarter of this year, North American commercial was 31% of the earned premium, now it's 36% of the earned premium where all the focus has been. And with Validus and Glatfelter coming into play, that's going to be pushing a little bit more and they've got very good, Glatfelter in particular, very good program results and the program unit terminated about 20 programs. So you can see loss ratio improvement, because of what’s left and new business coming in with Glatfelter that’s going to help that [skew] [ph]. Secondly, we've talked about rate changes that we've gotten on the book that overall average like 4%, 4% really belies mixtures by unit. And when you get into North American commercial, first, rate changes are on a gross basis and what's going to happen in 2019 is going to be a little different [skew] [ph] to where the rate increases are a little bit higher and therefore there's more distance between loss trend and achieved rate change, so keep that in mind. Next is think about that with all the reinsurance that Peter described, you're going to have a changing mix more so…

Peter, do you want to just say anything about the expenses.

Yeah. I think as Mark said, I mean, when we looked at it in terms of looking at how we're going to be below 100, it was the mix of the business, it was going to be how we were reducing the volatility in the portfolio and we ended up having a forecasted better accident year loss ratio, the reinsurance then reduced volatility, but also shifted that balance as Mark said on the overall loss ratio. So we'll see casualty business in the US having a bigger [session] [ph], so therefore a better balance. And then I don't want to lose sight again of Validus and what Glatfelter will contribute. And as Brian said, on the expense side, we saw a lot of ramping up of the acquisition expenses this year. That was from a portfolio shift in addition, when you looked at the comparisons in ’17, there were some onetime anomalies, so it looked like it was growing a little bit more and we've been addressing the general operating expenses in the back half of the year, so while you didn't see perhaps a lot of material impact in the first half of ’18, you saw it in the back half of ‘18 and we believe that that will continue to not only sustain, but we're going to be focused on it throughout the year. So I think when you look at all those different components, you get to below 100.

And I'm just -- let me just add two things. One, [indiscernible] AAL, the AALs, we haven't given you one, but one can assume they're not going up, they're going down either through a combination of efforts, reducing our gross and net risk, reducing concentrations, and of course reinsurance structures that we put in to place, which are all reducing the AAL, I'd say that. So the only other thing I'd say, a little caveat, so give me a break here, is and we talked about volatility. So volatility in this book is diminishing, but it's not -- we haven't gotten it all out of the system yet. We've got risk attaching covers and things like that. So there's still a few of the multi-year policies left and the [unearned] [ph] bleeding in, is we still have some stuff with larger per risk limits. So the volatility in the first quarter should be greater than the volatility by the end of the year. So, you can't -- you got to just recognize that volatility thing is there, but I think Mark and Peter outlined why we've said entering 2019.

My follow-up is on net investment income, the $13 billion annual run rate you guided to, what's the base assumption for market return in that $13 billion. I'm wondering what's the sensitivity around that, for example, this quarter, market up 10%, how much of – so better than the run rate you’re going to realize in net investment income.

So, we're making the assumptions consistent with what we made in the financial supplement last year. I mean to give you a sense of the volatility from our perspective of return on assets, I mean, we're managing 313 billion of assets. The volatility of net investment income, which is a function of both the economic risk and the accounting mechanisms by which we account for these things is roughly 25 basis points of return volatility on assets.

And we'll take our next question from Josh Shanker with Deutsche Bank.

I fear all these questions are going to be about how to get to 100%, but I'm going to also try and dovetail on Kai. In the quarter, we saw a big increase in the acquisition expense ratio, offsetting the improvements you guys have made on the GOE expenses. Going forward, where should we see the acquisition, you’re buying more reinsurance and what not that's taking up, is that going to offset the improvements you're making on the operations in terms of people and proper allocation of expenses?

Well, I think, let me start by saying that increase in the acquisition expense was really related to personal insurance in particular. That book of business runs at a high combined ratio, but one under 100. So it actually is additive to the portfolio, particularly when you think about the volatility of it. What was masked was the PCG business deteriorating, so you didn't see the loss ratio improvement. So, the acquisition will stabilize, we don't see that growing. I think there's reason to believe we would get that down. And we will be working on that -- the fact that the offset didn't happen because of the PCG loss ratio deterioration. Peter, do you have anything you want to add to that?

No. I think it's fair and we don't see acquisition expenses increasing in 2019. We just had, like you, said the ramp up of shift in portfolio and then again having some tough comparables year-over-year and believe that we will continue to focus on the GOE and don't expect to see expenses go up in acquisition.

And one related question, I guess maybe unrelated, you are not responsible for the decisions made by the previous management team, but trying to understand what this Go Big strategy meant. There was also a strategy called rapport where they were trying to minimize I guess the capital consumption of various underwriting decisions. Were the two things in conflict with each other, did one of them not exist, was the company trying to minimize its underwriting action order return capital, how did those two things dovetailed with each other and were there conflicting messages, I guess, given the underwriters, what -- explain what was going on before a little better so we can understand how it can improve?

Sometimes, you can’t explain things. I'm not sure I can, but let me try and answer those things. So, I think there was a belief in the balance sheet, starts with a belief in the balance sheet and that's okay, we need a balance sheet, we need a balance sheet to take risk. And it's got -- your risk has got to be matching that balance sheet, but if you tell underwriter, it's okay to write 2.5 billion and by the way, who else is going to do it, so you think you'd corner the market, right. So if you don't get paid for it, you've got a problem. So you put out an extra $1 billion of limits with almost no payment, because it's impossible to get value for that kind of additional limit. So when you add that to, let's say, underwriting where the underwriters were not selecting properly or they were adversely selecting against, being selected against and you go large, you go large on risk, you shouldn't be writing, it just exacerbates everything and then you try to -- then as an actuary, you try to reserve this stuff where the volatility is impossible, how much risk can you put into your reserve levels to try to recognize the extreme volatility that might occur. So look, I don't -- all I know is, I wouldn’t have done it and I know we're not doing it. That's about all I could say. Hope that helps.

We’ll take our next question from Elyse Greenspan with Wells Fargo.

My first question, so in response to Kai’s question, Brian, I think you said that sometimes there are some volatility, given some multi-year policies potentially in the first quarter. I just want to reaffirm that based on how you see your book running that you do expect to come in below 100 in general insurance in the first quarter, and then a couple of Q4 numbers that might help us think through that, was there any kind of current accident year adjustment that would have impacted the fourth quarter, chewing up earlier quarters and can you let us know what valid, if it was additive to margins in the fourth quarter?

Those are three different questions. [Technical Difficulty] Yes. I did -- I just put a little copy out that there's volatility, but yes, I'm reaffirming that we’re entering 2019 expecting to make an underwriting profit, including AAL. So that's answer number one. Then there was a question about whether or not there were things that occurred in the fourth quarter. Mark, you want to take that?

Yeah. I also applaud you, Elyse. I think that’s like three question, part A through H. So the question about accident year 18, were there any movements, there were some ups and downs, but the net was 30 basis points on basically $7 billion for the whole year, so it’s $40 million. So it’s nothing to get excited about. And there was one other one, right. Validus

This is Peter. Is it with or without cat, because if it’s with cat, it was not accretive because the one that if you look at ’17, they're very good at how they buy reinsurance, but ’17, if you look at the comparables, was more active in the third quarter than the fourth quarter. So if you look at both quarters together, it was roughly the same in terms of what contributes to net cat, if you take out cat, it was in the range of where we were as general insurance. So it wasn't terribly far off, but.

Do you have a follow up series of questions, Elyse?

Yeah. I mean the one other thing that I think just stood out to me, Mark, maybe you can give a little bit more color, Peter, the international margins, I know you made the year over year compare that they got better relative to the fourth quarter, they didn't get better relative to the third quarter, did you guys see any one-off kind, were you guys on any large property losses or something that might have impacted the numbers or was it seasonality that you're not necessarily comparing international Q4 to Q3.

Yeah. It's a good question. If you look at history, you're going to see it’s bounced all over the place, it’s a lot of different lines of business, it’s 80 countries, reflecting kind of like what my prepared comments on FL, internationally get a poppier out of India, then you get one out of APAC, and then let’s break the next quarter, but if you look back to the full accident year of ‘18 and the full accident year of ’17, they're basically mirror images of each other. One was 65.2%, one was 65% even. So, you're going to get pop. So sequentially, I don't think gives you any information content.

We will take our next question from Yaron Kinar with Goldman Sachs.

Mark, I thought your walk to mid-90s accident year combined ratio was very helpful. And I think I understand the math around the impact of rate improvement, but ultimately there must be some assumption around loss trends there as well, right. So I just want to make sure that the way I'm thinking about, the way you guys are thinking about are roughly the same namely that to get to this mid-90s, you're assuming that the rate improvements are in excess of loss trends, even in line that where you're getting both the rates clearly, those are lines that are also facing steeper loss trends.

I think, yes, the answer is yes, but let me -- I just want to emphasize something here and that is rate alone would not have improved this portfolio. Okay? Even if it was above the loss trend, because if the portfolio is a collection of risk, you shouldn't be writing, you can't get enough money for that stuff. So there is, the untold story is the improvement of the portfolio, which really affects the loss ratios and combined ratios and that's an important thing to keep in mind. Do you have anything else you want to ask, Yaron?

Yes. I do. Actually maybe shifting gears to the investment portfolio, so you guys have clearly spent a lot of time and effort to de-risk the liability portfolio, especially in general insurance, what are your thoughts around the investment portfolio today broadly and in general insurance specifically?

Well actually, we've had three years to do the de-risking of the investment portfolio. So if you look out over the past three years, we've dramatically reduced all the fair value option equities, we eliminated a whole host of legacy investments that we took over in 2015, things like life settlements, which were both illiquid and had modest returns for the risks that we were getting, we reduced our hedge fund exposures by over $7 billion. We reduced our private equity exposures, we changed where the investment -- risk investments were held. We reduced the amount of those exposures in the life and retirement business because there's so much more capital markets exposures to the underlying core business that it didn't make sense to include those types of asset classes in those books of business. If you look at the exposures that we have to equity markets, let's understand when you look at our overall investment portfolio, it's still an ALM based book of business designed to match our liabilities. So we -- all the alternatives are really allocated to the surplus of the general insurance books and it's very modest and what we try to do is manage liquidity, economic and accounting volatility and also try to get excess risk adjusted returns. So in looking at that portfolio, we're very comfortable with what we currently have in our alternatives, which is a mix of private equity and hedge funds, but we're always tweaking it, but the absolute holdings that we have in scale, we're very comfortable with, we'll continue to make ongoing rebalancing of how we allocate that risk bucket, but I think the size of it will continue on going forward.

We'll take our next question from Erik Bass with Autonomous Research.

Two questions on life and retirement. First, can you talk about how much impact the markets had on DAC amortization in the fourth quarter? And then secondly, you've talked about expecting a couple of basis points of core spread compression per quarter. I think the decline was much greater this quarter due to higher crediting rates and you mentioned some unusual items, but how should we think about the rate base level of this for 2019.

Thanks, Eric. First of all, in terms of the fourth quarter impact, the impact on DAC from the market was -- versus the prior year just over 80 million. And of course, it's a combination of where the equity markets levels are, an impact from the credit spread widening, et cetera. And so, if you look at the overall fourth quarter, the sharp decline in the AUM in the last four weeks impacted not just the DAC, but also fee income, which is a reflection generally of where the equity markets are and also the impact of the spread widening on mostly the fair value options and the impact of the markets on yield enhancements and accretion income is also something to keep in mind. So, as we look at the whole question of the cost of funds and base spreads, I think what's important is there were some anomalous items, mostly in group retirements in the fourth quarter, but year over year, cost of funds was 2.73, which is just 3 basis points from the 2.76 in the prior year. So it's not actually about the prior year, fixed annuities year-over-year was flat at 2.65 and variable and index almost flat at 1.24 versus 1.26. So really the cost of funds is not something other than that one-off item in the fourth quarter, significant concerns. And as we look at the base spreads, obviously, in particular some of the market effects were primarily what you saw in the effect in the fourth quarter. And assuming where the markets were at the end of the year and where we believe interest rates are going to be, our interest rates were, we are still confident of that maximum negative spread compression of the two basis points in 2019.

Just want to clarify further -- your guidance, you talked about life and retirement earnings being flat in 2019. Just to be sure, that's off of the adjusted 2018 number you show in the supplement and then what are you assuming for the equity market return?

Yes. That is correct. We are assuming that comparable also on the ROE, ‘19 versus ’18. In terms of equity markets, we're assuming a modest improvement in the equity markets from where they were as of the end of the year.

We’ll take our next question from Andrew Kligerman with Credit Suisse.

First question goes on Mark. Last quarter on the call, late November, you talked about how you had reviewed much of the book and saw no material red flags upon that material review. So Mark, I'm wondering, I think you mentioned a little earlier something about amortization. I'd like a little clarity on what you saw subsequent to that November call and what your confidence is with regard to prior year development goings into 2019.

So with respect to November, so when we looked in for the fourth quarter effectively backwards, there's really no adjustments of note on anything prior reviewed. So, as far as I'm concerned, those numbers are stable, like the reallocation between couple of lines, but in the average, nothing moved. So if you're -- by end of November, you're really talking about Goldman's December 5th conference, which I assume you're referring to December 5th already.

They didn't invite me.

Okay. Well, I'm sure you might have read about it. But what's our responsibility as an insurer, it’s setting reserves for all occurrences that have happened 12/31 and prior and this was December 5. So 3 weeks of losses to emerge, but so at the time, everything was still in flight, but I'm not making excuses. The point is that when you get into the depth, you didn't hear us talk about personal lines problems, you really didn't hear us talk about rest of world casualty in any material way and you didn't really hear us talk about the buffer or comp excess on it. So, it was financial lines, that was the center issue, mostly US issue, some scattered pop. So back to the reasonable range we're comfortable with that. To your question about, well, I think that really was the net of your question is how do we feel about things, how we've been through a whole cycle. I feel pretty good about it. [indiscernible] Let me say something the confidence in reserves starts with the confidence in the underwriting this -- producing those reserves. So the more and more comfort we have about what we're doing as a company and the work around taking the volatility out to make the results more predictable just adds to our confidence. You want to add something else Andrew?

So, I understand the business mix and the pricing increases, but where I'm not getting the math is, if I make an estimate of Validus and Glatfelter premium, I still see 2019 over ‘18 getting a premium increase and even with the rate increases, which you say are about 4%, if I subtract that out, maybe premium were about flat. And so I guess, how are you going to get this loss ratio improvement, what business are you departing that would get this improvement, if premium is so flat, it's just hard to see what the moving parts are that are going to get you this this loss ratio improvement, so maybe you could give me a little math on what's going away and what's coming in and still being able to maintain flat premium.

Well, thank you for the question and it's a complicated answer, because the portfolio has been shifting across the world, as we outlined in terms of going from a large limit strategy to using limits a little bit more discretely, but also Brian keeps emphasizing, it's also the risk selection and rebalancing the portfolio. So within every business and every geography, we have been looking very hard at getting the appropriate returns once we make good risk selections, we have the right coverage and that has happened in a dramatic way. So the one, I will highlight because it's happening across many of the portfolios would be the Lexington, the Lexington was doing admitted and non-admitted, it was doing business from wholesale and retail, it was doing large limits on property and casualty as well as having a very large program business and so it has been recalibrating its risk appetite to substantially play in a role where it has very strong expertise. We have brought in industry leaders, focusing on the E&S space, focusing through wholesale, making sure we're using tighter limits when we do the risk selection, going into having a better balance with not only large accounts, but in the middle market and so the portfolio has been starting to move in a direction where we're recalibrating and reconstituting what the portfolio looks like, at the same time, there are some businesses that are further along some that are not and with that, you're going to have the general insurance excluding Validus and Glatfelter, will decrease year-over-year on a premium basis, then the complexity of the reinsurance, so I know it's hard to do the math, but we believe that we are repositioning the portfolio with a better risk selection and that will be reflected in the accident year loss ratios as we earn into 2019. Mark, do you want to add anything to that?

Yeah. Just the fact that, I think what we said on last quarter's call that we don't need a hard market to improve this book. It comes back to the composition and what we’re letting go if we can't get rate and structure that we want and what's coming in and everybody focuses on price, I get it, because it's measurable and it’s something you can compare, but a whole flow of business, a whole understanding of the cat exposures, the whole structure, price is last and that's what we measure.

Andrew, let me just – you have a risk that comes in the door, right and we've been writing it for some time, we've been attaching too low, so it's either deductibles, it’s an access placement, we've been attaching too low and we've been giving way too much one that it's at the top. Okay. So you get rid of the limited copy that we're getting paid for, it just improved your portfolio immediately. You raise up from where you were to a higher limit where the risks diminish, right. Now, that'll cause you to get less premium, you've got a much better risk on your hands and that doesn't even count getting [indiscernible] So, you've got to trust this, I’m telling you, the premium is flat and we don't need a hard market to make this thing better, as Mark said.

Yeah. Just very broadly then, I mean, because what your strategy makes impeccable sense bluntly, it would sound to me with these reinsurance issues and lower limits, that at least 25% of your portfolio is turning over the good stuff in, the bad stuff out, is that a good assessment.

Getting back to, some stuff is going clearly, but we're not like getting wholesale out of a lot of business, but what Brian said is key. You can stay -- we could – if we have 100 risks, we could still stay on the 100 risk, but on a different place on the structure where the risk reward trade-off is better, 100% renewal retention on the account and the volume is down 33%, but it's a better loss ratio expectation. That's what you've got to internalize.

The next question comes from Tom Gallagher with Evercore.

Brian, so your description of the Go Large strategy, large limits and de-riskiness, can you talk about where we are on this de-risking process, would you say it's largely done or is 2019 going to continue to be a risk reduction year? And the point of my question is usually when you materially reduce risk, normalized earnings go down, absolute ROE goes down. I understand risk adjusted earnings should be getting better and cost of capital should be lower, but I guess broadly speaking where are we on that risk reduction journey, what does that really mean for ROE about 2019. I know you've really given out like 3 year forward ROE, does it mean near term ROE is going to be more depressed, you're going to get more of a back end loaded improvement, anyways, pretty broad question there, but what are your thoughts?

Yes. So, is the process done? Well, as we described, Peter and I, Tom Bolt’s work with all our underwriters has been distributed, so we've set the tone of what we want to do. Okay. So the strategy, yeah, I think the strategy is done. The implementation is being done and risk by risk, as they come up for renewal, you make adjustments and things bleed off. So that's what I was saying, the first quarter is going to be riskier in terms of volatility and the fourth quarter just as the portfolio earns in at the strategy that we've been deploying. So strategy is done, implementation is weaving its way through the book this year. This ROE, de-risking ROE thing, well, that just presupposes that you're getting paid for the risk, but if you don't get paid for the risk and you stop taking it, you are pretty close up. And that is not even risk adjusted, so when I said 8 travels to double digit, a lot of that has to do with a continuing effort on our expenses, which is a significant task for us over the next several years and that's why it's that journey is a little longer, hope that helps.

That does. And when I think about, to your point, to me, the clearest expression of what risk reduction is going to mean is AAL, but I know previously, it was 4.5 points and I guess you're not giving specific guidance on that, why not give specific guidance on that because that's the one tangible thing we can point to to describe the actual earnings benefits from the risk reduction program?

Well first of all, the AAL is one measure, right, but it's not the only measure of risk, right, so you've got a lot of risk kinds of measurements that you should deploy. Well, let's see, this AAL number is a moving, it's moving, right. Obviously, it's moving for a lot of reasons, we're reducing concentrations as I said earlier, we're changing risk profile in terms of per risk size, the per risk limits that are being deployed out there and the reinsurance is being put into effect. So you have to understand that the AAL is a moving number right now, getting better. So when we feel we've got a good number, I would not have any problem, tell me what it is, so I'm not hung up about it, but I can tell you it's moving south. And it's not the only measure though.

Tom, just one thing on your preface on that AAL. AAL is explosive, the underlying risk associated with AAL, right. So it's just that that's in your face. Cat occurs, earthquake occurs, front page of The Journal where everybody can focus on it, that’s why with cat bonds, everyone thinks they understand the model risk and all that. Now look at the history of AIG, look at the history of any large company. It's the deferred risk, not immediate risk associated with casualty business. So primary casually, excess comp, I mean look at it at the past of the industry. That risk is simply deferred, but it's as big and real and more capital exposed as property, but everybody gives a short shrift, but that's what you have to focus on and that's what Peter and his team has done.

We'll take our next question from Meyer Shields with KBW.

So looking forward to 2020 if I can, is it fair to assume that the impact of multi-year P&C contracts and the expenses associated with 2019 life and retirement growth that both of those will be sources of improvement in 2020.

Well, certainly the multi-year should have bled off by that, no question about that. I’ll let Kevin talk about.

Yeah. So Meyer, on the expenses, I think I mean if you look year-over-year right, ‘18 versus ’17, we had a reported increase in GOE of about 110 million, 40 million of which were one-time items in ‘17 that helped ‘17 and don't repeat. So, the expense increase base is about 70 million. Around half of that is actually directly associated with new business growth. Obviously, our premiums and deposits are way up, serving those policies, et cetera, has its additional burden. So relative to the projects, we're certainly getting an immediate benefit from those and we'll continue to invest what we need to in order to keep up with the increasing regulatory demands and customer expectations, but a part of that increase in GOE is related to new business and so to the extent that our new business continues to grow, then, we're going to continue to see new business expenses associated with that growth.

Meyer, do you have a follow-up if you have one, and then I'm going to cut this off. So, do you have anything else?

I do. Just looking within casualty, is the mix of business between the four lines of business you report, property, special risks, liability, financial lines, is that going to shift dramatically when we take into account all of the things that have been done on gross and net writings.

Yeah. Peter, you want to do that?

Thanks, Meyer. The portfolio will continue to shrink on a net premium earned basis because we just placed a 50% quota share of the first 25 million and life property, we have been reducing the gross limits substantially, so that portfolio will continue to be recalibrated and again you'll have the impact of the reinsurance in 2019 and 2020.

Okay. So let me just close here. Thanks, Meyer. Let me just close here and I – first of all, I want to thank you for participating on our extended call today and bearing with us. We had a lot that we wanted to talk about. I hope we gave you a good sense of the hard work we've been doing and continue to do it here at AIG. You know maybe it's a cliché, but it's absolutely true, our greatest strength is our colleagues and I want to thank them all for all they're doing, not just for the company but for our clients and our stakeholders. Thanks, guys and have a good day.

And that does conclude today's conference. We thank you for your participation. You may now disconnect.