Robert Block Thomas J. Wilson - Chairman, Chief Executive Officer, President, Chairman of Executive Committee, Member of Equity Award Committee, Chairman of The Allstate Insurance Company, Chief Executive Officer of The Allstate Insurance Company and President of The Allstate Insurance Company Steven E. Shebik - Chief Financial Officer and Executive Vice President Matthew E. Winter - President of Allstate Auto, Home and Agencies Don Civgin - President of Allstate Financial, Chief Financial Officer of Allstate Insurance Company and Executive Vice President of Allstate Insurance Company

Josh Stirling - Sanford C. Bernstein & Co., LLC., Research Division Ian Gutterman - Adage Capital Management, L.P. Robert Glasspiegel - Langen McAlenney Brian Meredith - UBS Investment Bank, Research Division Michael Zaremski - Crédit Suisse AG, Research Division Matthew G. Heimermann - JP Morgan Chase & Co, Research Division Adam Klauber - William Blair & Company L.L.C., Research Division Michael Nannizzi - Goldman Sachs Group Inc., Research Division Randy Binner - FBR Capital Markets & Co., Research Division

Good day, ladies and gentlemen, and welcome to The Allstate Corporation Third Quarter 2012 Earnings Conference Call. [Operator Instructions] As a reminder, this conference call is being recorded. I would now like to introduce your host for today's conference, Mr. Robert Block, Senior Vice President, Investor Relations. Sir, you may begin.

Thanks, Matt, and good morning, everyone. Thanks for joining us today for Allstate's third quarter earnings conference call. First, Tom Wilson, Steve Shebik and I will provide some color on our results for the quarter. And then we'll go into a question-and-answer period. Also on the call are Don Civgin, Head of Allstate Financial and Insurance; Judy Greffin, our Chief Investment Officer; Sam Pilch, our Controller; and Matt Winter, Head of Auto, Home and Agencies. Don Bailey, who leads the Emerging Businesses, will not be able to join us today since he's on the East Coast a result of Superstorm Sandy. Last night, we issued our press release and investor supplement and filed our 10-Q for the third quarter. We also posted a slide presentation to be used in conjunction with our prepared remarks. These are all available on our website. Beginning with Slide 1, this discussion may contain forward-looking statements regarding Allstate's operations. The actual results may differ materially from these statements, so please refer to our 10-K for 2011, our 10-Q for the third quarter 2012 and our most recent press release for information on potential risks. This discussion will contain some non-GAAP measures for which there are reconciliations in our press release and on our website. And I will be available after this call to answer any follow-up questions you may have. Now I'll turn it over to Tom for his perspective on our performance.

Well, good morning. We appreciate you spending time with us today. Before we discuss results, let me say a few words about Hurricane Sandy, which I know personally affected many of you. First, I hope you all are getting your lives back together. As you know, this was a massive storm affected most of the Eastern half of the country. If you look on Slide 2, you can see the breadth of that. Putting it in comparison to other events, this is at least twice the breadth of Hurricane Katrina. The good news is that the wind speeds were substantially lower than Katrina and more like last year's Hurricane Irene. In fact, the winds speeds were about 2/3 of Katrina. And as you know, the damage is exponential as that speed goes up. And while it's, of course, too early to call ultimate damages, several modelers have put out insured loss estimates, Eqecat's at $5 billion to $10 billion, AIR's at $7 billion to $15 billion, with economic losses at twice that amount. If the storm's damage is over $8 billion in insured losses, that will make it the fifth-largest hurricane in U.S. history. Of course, a lot of the damage at this point would appear to be caused by flooding. That is covered by your auto insurance. As you know, the federal government provides flood insurance, and it's typically not covered by people's homeowners insurance. And of course, in these times of trouble -- that's when our company, Allstaters, do great work on behalf of our customers. We have 1,100 catastrophe claim adjusters out in the field as we speak; another 500 at our service centers. We have 24 mobile claim centers. We're reaching out to customers. We're doing advertising. We're walking the neighborhoods. We are very proud of…

Thanks, Don. Let's take a closer look at the details for Property-Liability and Allstate Financial for the third quarter. Beginning with Property-Liability on Slide 5, we show both the top line and combined ratio results. Overall written premium grew 5% from the third quarter 2011 to $7.06 billion, primarily due to the acquisition of Esurance. The growth rate increased from the second quarter 2012 as both the Allstate and Encompass brands also contributed to the positive result. Looking at the results by line, Allstate brand standard auto declined by 0.2% from the third quarter 2011, and it was similar to the second quarter 2012. Excluding New York and Florida, the 2 states where we have implemented significant profit improvements over the last 18 to 24 months, Allstate brand standard auto grew by 1.4%. New issued applications of 460,000 remained slightly below the prior year, but we're at the average level produced over the last 6 quarters, which ranged from 451,000 to 472,000. The retention ratio remained at 89.0%, down 0.10% from the third quarter 2011. Policies in force declined sequentially as the level of new business was not sufficient to offset the renewal losses Allstate brand homeowners' net written premium of $1.69 billion increased 3.2% compared to the third quarter of 2011. The favorable impact of rate changes on average premium more than offset the decline in units. We continued to see great changes in order to achieve our stated return goals. In addition, we are rolling out our new product, House & Home, now in 17 states, including 5 states that rolled out in the month of October. Encompass continued to show positive results in both premiums and units, with increases over prior year of 5.3% and 3.8%, respectively. Esurance contributed $282 million in premiums written during the third…

Thanks, Bob. We continued to proactively balance portfolio yield and return objectives in this challenging low interest rate environment, delivering strong investment results with a total portfolio return on a GAAP accounting basis of 2.4% for the quarter and 6.3% to the first 9 months of the year. On Slide 11, you can see longer term trends in the composition of our portfolio. We ended the third quarter at $93 billion in amortized cost and a fair value of $99 billion, reflecting improved valuations, which along with portfolio income and improved underwriting cash flows from our Property-Liability business, more than offset the ongoing reduction from Allstate Financial's spread-based business. The chart on the left-hand side chart of this slide helps to illustrate the results of our proactive steps to enhance the profile of our portfolio from a risk and return perspective. Note that the drop in the portfolio's amortized cost resulted primarily from our planned reductions in Allstate Financial's liabilities. As we headed into the financial crisis, we proactively reduced financial sector and real estate related exposures while increasing the liquidity of the overall portfolio. Coming out of the crisis, we have favored the strong fundamentals and stability of credit, including high yield, over public equities. During the most recent quarter, we opportunistically reduced risk by selling $722 million of structured securities, realizing a loss of $119 million. The securities increased in value over the course of the year, justifying the sale into a strong market. Collectivity, these actions have resulted in a strong, well-positioned portfolio. We have been proactively managing interest rate risks by optimizing our fixed income portfolio's position on the yield curve, focusing on intermediate-term securities. This will reduce reinvestment risk and lessen the sensitivity of our portfolio's value to increases in interest rates. On the right-hand…

Okay, Matt, if you can start the Q&A, please.

[Operator Instructions] Our first question is from Josh Stirling of Sanford Bernstein. Josh Stirling - Sanford C. Bernstein & Co., LLC., Research Division: The question I would ask would -- recognize since you can't comment on an actual estimate, of course, it's way too premature. But you are getting reports of first notice of loss. You're out there in the field talking to customers. You have some sense of the volume of the storm. I'd love to get a sense from these operating metrics sort of where you are, you've -- say, in perspective in relation to Hurricane Irene last year. And then similarly, you probably are starting to get a sense of the mix and composition of the claims that you're actually going to have to adjust. And I'd love to get a sense especially when we think about severity going forward. Is this going to be driven by lots of large total unit losses or are you going to see more sort of modest -- far many, but may be more than previously, but sort of many more sort of smaller losses that are sort of damage to ancillary structures and things like that?

Josh, this is Tom. You're correct in saying it's way too early. Even on loss counts, it's a little difficult because sometimes people have cell phones; they can get access to it. Sometimes they don't. I can tell you, we're very busy. We also do, as you know, flood adjusting for the federal government where right through on the flood, we're one of the bigger handlers to the program as well. So we take all of those claims in. Anybody calls this customer, we go out and see him. So it's really too early to tell on either the volume, the mix or the severity. What I can tell you is when we do our estimate, we'll try to give you as much information as we can at that point in time. I can tell you that the -- from a risk management standpoint, we are down in policy counts in all of the affected areas by a reasonably substantial amount. And if you're interested in that, Matt can give you some are details on it. But we are -- the actions we've taken over the last 5 years in places like New York, New Jersey, all up along the East Coast, have produced the policy counts we have there, in the 10% to 30% range. And that will obviously impact what our losses are relative to what they would have been. But what you really want to know is what are they going to be, and I can't give you any more clarity than that. Josh Stirling - Sanford C. Bernstein & Co., LLC., Research Division: I certainly understand. The investors, I think, at this point are sort of obviously dealing with modeling firm estimates. I think you guys are sort of sophisticated users of the models, and you've got a lot of your own data and your own perspective over time. I'm wondering what sort of -- to use the phrase, sort of, "known unknowns" model errors and sort of historic sort of biases that you think that historically perhaps we've seen that you guys might be watching for as you handicap the estimates themselves? And one topic of conversation, I think that people are curious about is whether this has a possibility of being something like Ike, which ended up going further inland and ended up driving sort of surprising levels of losses with what would have been perceived to be modest wind further inland than typically models would have suggested at the time.

Josh, maybe I can give you a comment on the models and then a suggestion maybe on how I can help you sort out what you're trying to get as you look to impact on us. If -- we do obviously have our own models. We buy all the models from everybody else and we have our own policy data actually loaded into that so we have a wide range. Matt and his team have done a number of scenarios. It's a really wide range. So that number could be low; it could be higher. What I think we could do to give yourself some comfort as to how big it could be would be look at -- go to the website and put in gross losses into the reinsurance section. It's got the interactive model, which will tell you what the net ought to be. And you can see that at certain levels our reinsurance programs kick in, which caps our exposure. Oh, Bob tells me it's not interactive. Sorry. I guess, the interactive part will be you'll have to sign in. I thought we're a little more sophisticated than that, Josh. Sorry. But Bob can help you.

We can help you out.

Our next question is from Ian Gutterman of Adage Capital.

I guess, first time, when I look at the auto results, obviously very good. I think you said it was a 93.7, I think, accident year. But if I add on it to the normal caps, a couple of points, that's high 95s, which again, is a solid result. But I'm wondering can that get better? Can we get that -- in the past, you've obviously seen low 90s, but even in 94, 95, is that possible given rates are starting to improve? Or is 96 kind of the best you can do in the current rate environment?

First, I just want to give one point of clarification. The information we provided is calendar year, not accident year. So...

The underlying to that.

The underlying, sorry.

What the underlying says...

But the underlying to those [ph] cats, I took the underlying and added 2 points of normal cats to it.

Okay. The way we look at the auto business is it generates really good returns for us, so return on capital at that level -- we've shown some slides before -- depending what you want to assume in terms of capital levels, is in the 20% range, plus or minus. That's a really attractive return on capital. We'd like to grow that business at those kind of returns. So I would rather see us grow it and maintain auto profitability than squeeze a few extra dollars out of margin.

That actually went into my follow-up, which I was going to ask. How is the outlook for that? I believe there was an internal memo that hit one of the newswires that suggested you're broadening your targets and maybe -- but trying not to seem to be a little bit more optimistic about growth going forward, and certainly with competitors, some of your big competitors having to chase rates to improve their profitability back to what it used to be. It would seem maybe you're in a little bit of a sweet spot here?

Let me maybe provide you an overview on growth in total, and then Matt and Don can give you some more detailed perspectives on both the Allstate brand, Esurance and Allstate Financial. I think you have to think about it in total as you think about our strategy. So first, our strategy focused on these 4 segments have competitively differentiated products in those segments targeted to those customers. The largest segment, as you point out, is the advise and brand, focus one, that's served by the Allstate Agencies. The overall policies in force are down there, which is largely as a result of the homeowners' actions we've taken, which was intentional on our part to improve returns in that business. Along with that, standard auto is down a little bit partly due to the homeowners, partly due to New York and Florida. Matt can talk about what his opportunities are there to grow. I think the biggest opportunity is to improve our customer retention, because these new business levels are kind of running about where you expect them to be. We have expanded the disclosure. As I've mentioned earlier, you can see Good Hands Roadside, which serves that segment largely, is up substantially. But we don't count it as a policy since they don't pay us until they actually use it. The advice and brand neutral segment that's served by Encompass, that independent agency is up slightly. We need to recapture some lost market share there. Esurance is that self-served branded segment. We're growing it a rate as fast as we're comfortable with. And then Allstate Financial is growing the life policies sold to that lower left-hand segment and our Benefits business. Matt, do you want to make some comments more specifically to the -- I guess, you are up to standard auto so we could zoom in there.

Sure. It's Matt. I'll comment a little bit first on the some of the drivers. So Tom mentioned our 2 primary drivers that have impacted growth have been our homeowners' return improvement actions and some of the auto profitability issues in New York and Florida. And I'll take it up one higher level and I'll say that I think our fundamental approach to the marketplace is to be very proactive and very disciplined. And sometimes, that creates a first mover disadvantage on us, where we're first to take action, where we see need to do risk mitigation, we're first to take rate where we see emerging environmental and economic trends, and we're first to take proactive action that actually can sometimes impact our growth prospects. So I think you know that we have about 16 states where we're either closed for homeowners' new business or have significant new business restrictions in them. I think Sandy is a good indication of why we did that, and it will prove to have been a good move for us to reduce our exposure in some of these coastal areas. But all of those actions, whether they're profitability actions in New York and Florida on auto or these homeowner actions have an impact on our ability to grow. And as Tom said, you can't take all of these profitability actions. You can't do the non-renewals. You can't do reinspections. You can't tighten underwriting without some impact on our retention, and that is really what we see as our prime opportunity for growth at this point because we think we were early in most of these actions, we were proactive. And as a result, I think we're further along on the timeline, and we are at the point of stabilization, and we are at the point where we are beginning to see pockets of growth. You mentioned broadening the target. We have several growth initiatives out there, including broadening the targets and price optimization work, some additional work to analyze our ability to grow even faster in selected areas. And so, yes, we have those pockets of growth emerging. We have about 20 states right now that had positive year-over-year growth in standard auto. And our hope is over 2013 and 2014, we will dramatically improve that number without losing any of our discipline on the profitability side. So our first mover disadvantage, I think at some point, will work into a first mover advantage, where we will emerge out of this cycle faster than some of our peers and we'll be able to take some of the growth opportunities that are presented to us, which I think are many at this point.

Great. So it sounds like you'll be disappointed if we don't see growth in standard auto for 2013?

I don't remember saying that exactly, but you draw your own conclusion.

Ian, it's Don. Let me just give a quick update on Esurance. As Tom said, of course, we acquired Esurance so that we could grow in that self-directed segment in a profitable way. That's been the goal; it's to grow our market share profitability. Since the acquisition, we've done a number of things. We've linked the brands together, so it's Esurance and Allstate company, which has played well. We've clarified their position in the market through their features and their technology. We've developed a new ad campaign, and we've thrown a lot of weight behind the new ad campaign intentionally. And here's how I would characterize where we are today. I think we're very happy with the results we're seeing. The response rate with new quotes is up dramatically. The conversion rate is up; the retention rate is up, I mean, you can see it; over 22% improvement in policies since just the beginning of this year. That's a big number. What you're also seeing, though, is a high combined ratio. Part of that is intentional. The advertising model, the accounting model is such that we have to expense that advertising in the quarter we take it, even though the value of the policy continues to linger on through retention. And so, we are watching the GAAP combined ratio. But I'll be honest, we're more interested in the economic combined ratio. If we're convinced that we're writing good business that will be profitable over its life, and we are, then we're comfortable tolerating the GAAP loss ratio being over 100 for some period of time. We also have a little bit of an elevated loss ratio in the third quarter. Gary and his team are watching it closely. We're working together to make sure that we have all our resources between Allstate and Esurance on-the-job. I'm confident that number will get back in line. But all in all, I'll tell you that the growth prospects for Esurance are going up so far very well, and I think we're very optimistic about the future.

Our next question is from Bob Glasspiegel of Langen McAlenney.

I need some little bit of help on your reinsurance slideshow, given that I can't plug this stuff into a spreadsheet. You gave an example of a hurricane going through New Jersey, which seems to have its own cover. And there's 2 different covers, which I don't quite understand how they interact. But I think you talked about a $700 million type loss working its way down to $200 million net, but maybe if you could just give -- if the number was $10 billion in the 3 states, simply what the gross and the net numbers would play out, that would be helpful.

Bob, I can give you a call back later and we can walk through the steps. The 3 contracts -- the contracts in New Jersey all work together. So it's -- I'll give you rough numbers, but it's about $150 million retention, and then, it's 95% reinsured up to the top of the contract. So we can walk you through that.

That's New Jersey.

That's New Jersey property losses only. That does not include auto.

Right. And then of course -- and as I said, auto flood is covered, and it's at 21 states. So it's a big storm. But we can help you sort out here.

So if you had a $10 billion sort of loss in the other 2 states, what would the net number be?

First, I assume you're talking about $10 billion as the industry level.

Right. Right. Let's say, you had 10% to 12% of that, so $1 billion gross, what would the net be?

I think it's highly dependent on the mix of which states and what type of losses. But if you want and you come up with 2 or 3 different scenarios, Bob can show you how those would work through the reinsurance program and you can come up with an estimate on it.

Okay. I think others might be interested in this as well. But on Encompass, it looks like a mini-breakout from what's been a troubled line for a long, long time. I mean, you're growing in premiums and the underlying underwriting is improving. Am I overreacting to this quarter or are you -- do you think you got the horse running the right way around the track?

Well, I guess -- I would agree. I think, first, this business is a good business for us to be in, serving that customer segment. Second, as you know, that business is probably down $700 million or so in premium over the last 3 or 4 years. And so it's got some room to come back. Some of that, of course, was intentional because we're in places we didn't want to be from a cap management standpoint but there is room to grow that business. Three, we do have new leadership there. Don Bailey has got a new leader in place there. We like the results that he and the team are now driving, so I would hope to be able to see it continue to grow. I'm not sure what breakout is. We all have different assessments of breakout. But I think if you look at what's going on in that customer segment and the companies who serve that segment largely through independent agencies, I think we're well positioned to grow, yes.

I was referring to the underlying profitability, underlying combined ratio.

Oh, breakout in the profitability. Sorry, I thought you meant breakout and growth. Our profitability has gotten better. I like the trends. I don't -- I'd like to see it -- I'd like to see our combined ratios from down low, but I'd like to see the business grow as well. So I don't think we're going to -- this business, as you know -- we've had that business -- shortly after we bought it we took the combined ratio down from 117 down into the low 90s. I don't expect it to get into the low 90s at this point in this cycle, given where our competitors are. But I think it can be better than it is today, and I think the business can grow at the same time.

Our next question is from Brian Meredith of UBS.

Just 2 quick questions here for you. The first one, looking at the slide on the underlying margin trends on the standard auto business, looking at your underlying lost trends and what's going on there. I'm just curious, maybe you can talk about kind of the difference between the paid loss trend in which you're showing here, and the incurred loss trend. Because the paid loss trend appears like in the last couple of quarters has been above where your average earned premium per policy has been.

Yes. Brian, it's Matt. So you have all the issues that come up with paid loss trend, and I guess, BI paid severity is a good example of that. So you look at BI paid severity and it looks fairly dramatic. But as Bob said in his opening comments, about 1/2 of that increase can be accounted for as a change in the mix of claims settled by report year and a shift in the state mix, and the other 1/2 is consistent with the medical CPI. So we do have, I guess, some distortion that's always present when you look at the paid versus incurred. But I think if you look -- and I'm assuming you're referring to the bottom chart on Page 7 of the presentation, the underlying margin trend. I think if you look at that consistent pattern over the last several quarters with the nice trend line between earned premium and the losses, I think our goal is to manage that and our goal is to look at the loss trends in aggregate. I think it's somewhat dangerous to focus on the attribution analysis of each component. We have a term Tom and I throw around a lot here. It's "false precision." And we can get trapped in our own map, and we look at it -- we do the attribution analysis and we try to understand each of the drivers, whether they're medical inflation or whether they're a report year or a shift in state mix. But then we look at the aggregate loss trends and the aggregate premium trends, and we look at our rates that we're taking. And our attempt is to manage some of those levers so that each of those -- on one quarter, one is going to go up, the other is going to go down. It's like pushing on a balloon. And our job is to manage the overall size of the balloon and how much air is in there and the pressure being exerted on it. And I think history will show that our team is exceptionally good at managing that. They manage each of the components. They manage loss trends very carefully while maintaining a good customer experience and good customer treatment. And they're very proactive in looking at emerging economic and environmental trends and taking rate where it's appropriate and warranted and mandated and justified. And so I won't get too caught up in the paid volatility. I get a little more focused on the underlying trends, and that's why we tried to show them to you in the slide presentation.

Brian, it's Bob. Just a couple of things to keep in mind. Remember while you're looking at the paid severity trends, you also have to look at the frequency trends. Frequencies are down. And the fact that bodily injury and property damage together make up about 45% of the standard auto losses. So there are other coverages that are in there -- collision comp, uninsured motorist, and everything -- where the trends are good as well. So you're looking at 2 of the big covers, but there are still other pieces to that puzzle, and also the fact that I think you can tell from our reporting on reserves. Our reserves are reasonable and they're accurate.

Yes, you bet. Absolutely. And just one quick one on the life insurance business, if possible. Your investment spreads really kind of improved this quarter. I wonder if there's anything unusual going on there.

Our investment spreads, Brian, did get a lot better. I think you see several things going on. Some of it is what you'd expect, so we continue to work on crediting rates, and that's being offset somewhat but the lower yields in the portfolio and the continuing lower asset levels. But the investments spread, the biggest impact is the option valuation from the equity index annuities. I think if you look at apples to apples, you're going to have to take that out, and then you get back to what I said, which is the crediting rates offset by yields and the asset levels.

Our next question is from Michael Zaremski of Crédit Suisse. Michael Zaremski - Crédit Suisse AG, Research Division: In regards to Hurricane Sandy, it looks like insurance regulators in at least a few impacted states are saying the storm won't technically be classified as a hurricane; therefore, our hurricane deductibles won't apply. So I was curious how the difference between a hurricane versus a tropical storm classification impact Allstate.

First, if you're interested in the science of it, a hurricane is -- it gets its energy from the water. The winter storm, or Northeaster, gets its energy from the temperature differential in the air. And so that's why it's been classified that way. The tropical cyclone deductible, it probably will not be triggered up in the Northeast, and we're prepared for that, and that's what's fair and accurate for our customers. It may or may not -- I'm not sure where we are down in the far Southern states. There maybe a few states that would -- but the damage is much less severe down there than where the hurricane turned in. Michael Zaremski - Crédit Suisse AG, Research Division: So would you say the kind of estimates that a lot of the reputable firms are putting out there take into consideration that those -- the higher deductibles won't be triggered?

I don't know for that. I don't know, Mike. Not everybody has a tropical cyclone deductible. I don't know how good their models are on that. It's, of course, very complicated because even when you have it in the marketplace, not every policy has it on it. And so I would say, to the extent it is not triggered, industry losses would obviously be higher than when it is triggered because that's a higher deductible than you would have on a normal all peril policy. But I don't think it would substantially alter the numbers, which are such big wide slots anyway. Now you know that the ranges are billions of dollars, so I don't think it would impact it much. Michael Zaremski - Crédit Suisse AG, Research Division: Okay. And as -- in regards to, Tom, the 13% ROE target on a normalized catastrophe load basis obviously. So where will most of the increase come from now given that you guys seem to be getting close to the goal within homeowners not looking to increase auto profitability? I guess I do recall at the Investor Day you talked about unlocking capital within the life insurance segment. Maybe you can update us on that.

Okay. First, my entire team is raising their hands saying they're going to deliver it. They each want to [indiscernible] First, I think -- I would say we're above that target for the last 12 months, and I think that is reflective of where we would hope it would come from which is maintaining auto margins, improving returns in the homeowner business and getting returns up in the life business. But of course, there were 2 things that happened this quarter, which we wouldn't expect to happen every quarter. One is low path [ph] and the other is the change in the investment spreads that we just talked about. So we still have work to do to fill up the profits ahead of that. So homeowners is -- we feel very good about where it's at. Matt and his team are working hard on it. We've been at this for a while, as you know. We're definitely at a much better position. And we're doing a good job on auto profitability. We do have a little bit of a headwind in the life business with low interest rates, which starts to drift over into the property casualty business. We're going to work to manage our way through that and still achieve our goal, but there's a little bit of a headwind there. To the extent we can reduce the size of the annuity business faster, we will seek to try to do that. It was down about almost $4 billion over the last year. I think it's like $3.8 billion. But it was only down about $700 million in the last quarter. So you can see that rate of decline is down a little bit, and because of low interest rates and people don't want to surrender their policies. So Don and his team are hard at work trying to figure out other ways we can reduce the size of the annuity business, not the life business. The life business is getting good returns for us, the Life Insurance business. So to the extent we can reduce that business, that, of course, frees up capital and we get a higher return out of that as well. So we have a little more work to do on all fronts, but we're feeling better about where we're at. Michael Zaremski - Crédit Suisse AG, Research Division: Sorry. I guess just as a final follow-up then, so within the Allstate Financial, should we be looking at the statutory capital? Because that amount of capital has kind of been staying the same over the last year. The annuity portfolio has been running off, attrition in life has been growing. So we should we expect -- should that number be expected to drift downwards if the current run rates continue?

I would -- as the size of -- as Steve pointed out, the size of that portfolio is down about $20 billion or so in the last 3 or 4 years. As that size of that portfolio comes down, I would hope we could have dividends out of that business which would be the statutory capital, yes.

Our next question is from Matthew Heimermann of JPMorgan. Matthew G. Heimermann - JP Morgan Chase & Co, Research Division: A couple of questions. One, just on homeowners, I was just trying to get a better sense of what the underlying improvement was. On my math, kind of adjusting for the non-cat weather that you've benefited -- you've quantified in previous quarters. It looks like year-to-date, we've got about 470 basis points of improvement, which is a little faster than I think the run rate you all have talked about in the past, and I'm just curious if you could remind us -- I think the difference is there was some adverse weather in 2011. So just remind me if I'm thinking about kind of the puts and takes right.

Matt, it's Matt. I'm trying to understand your question a little bit. Are you asking about what component is sustainable and what is -- what component is based upon the lower-than-normal cat? Matthew G. Heimermann - JP Morgan Chase & Co, Research Division: Well, you've quantified the lower-than-normal cat, so if I look at just loss ratio, I'm getting year-to-date improvement, 9 months over 9 months of about 470 basis points. And I think, historically, when you've talked about -- so that seems high to me relative to what you've talked about kind of your expectations of core margin improvement. So one, I guess comment on whether or not that perception is correct, about how 470 might compare to your run rate expectation. And then secondly, if it is in fact, a little bit faster, is that attributable to the fact about 2011, I think in 1Q in particular, had some unfavorable weather that might be making that look little bit better. I'm just asking because I want to make sure that as I think about next year, I'm not botching anything up.

Well, that is true. So we had some unfavorable weather that exacerbates the year-over-year view. But also, I think the other way of looking at this is that this is a fairly volatile business and trying to do it over year-over-year while we're declining, and if in selected areas on the coast, is a little distorted because we're changing the mix of the housing, we're changing the risk profile of the homeowners business as an overall block. And so trying to compare year-over-year as if it's a stagnant steady mix is a little distorted. I think we're changing the risk profile of the business a little. House & Home, as you know, it's gone in. It was 12 states in the third quarter, as I think Tom mentioned, another 5 in October, so 17 states. That has a different risk profile. And we underwrite roofs differently. We assess the risks differently using some tools that we're used to using on the auto business. And so while -- we are seeing normal year-over-year improvements, where you could back out some of the things, we're also seeing a changing homeowners business, which I think is going to continue to improve. And so I would say that the quality of what is there is higher than it was even a year ago. And I think a lot of what you're seeing that you can't get to mathematically is a result of that change in quality of the business. Matthew G. Heimermann - JP Morgan Chase & Co, Research Division: Okay. And then as we play that forward in the next year, do you feel like -- is that something that we should expect to continue to help a little bit on the margin as we go into next year, or now we kind of -- is the impact going to be a less dramatic?

Well, it will be less dramatic. As we put in House & Home, we try to put it in the highest leverage areas first. And so we clearly wanted to maximize the early impact of the shift. And a lot of our efforts on nonrenewal and reduction of homeowners in the coastal areas and high-risk areas is well into the latter part of the timeline. So I would not straight-line it. I think you'll see a continued improvement as a result of those 2 efforts, but not at the same pace as we've currently experienced it. Matthew G. Heimermann - JP Morgan Chase & Co, Research Division: Okay. That's helpful. Sorry that it took so long for me to ask the question in a way that made sense.

No, no, no. I'm sorry I was struggling with where you were going with it, so I wanted to make sure I was answering what you were truly asking. Matthew G. Heimermann - JP Morgan Chase & Co, Research Division: Yes, and that was helpful. Then I guess just one on Sandy. And this just goes to -- one of the -- I'm going to bring up Katrina not because I think that should be the baseline when we think about the loss but because there were a lot of adjusting issues because surge is really a big potential loss driver here. Whether it's insured or uninsured or government insured is a different question. But Katrina, right, there were some regulatory push to cover flood losses. But wind -- and there was a lot of gray as to what caught -- and in some cases because there was a lot of gray and there weren't houses left to actually adjust. I'm curious when you think about Sandy based on -- and I know it's early, but do you think -- how is the gray area in this loss wind versus flood on the personal line side contrast with kind of your experience with Katrina? Is that something that gives you a lot of pause, recognizing that normally when you get big events, the tides usually go to policy holders. But I'm just curious more whether or not there will be as much gray.

Let me give you a quick answer because I want to make sure we get a couple of other questions in here. Of course, it's too hard to tell, I haven't been on the ground out there so it's hard for me to see. I will tell you that as it relates to Katrina, we did very well on wind versus flood. We were proactive about our approach to that. And of course, the issue there was a house on the coast, big hurricane comes, house is gone. Was it the wave and then the wind came? Or was it the wind and then the wave came? And trying to sort that out. That doesn't appear to be the situation here, but it's way too early to tell what the issues will be. I will tell you that we have a history, though, of doing what's right for our customers and giving them what they paid for.

Our next question is from Adam Klauber of William Blair. Adam Klauber - William Blair & Company L.L.C., Research Division: How aggressively do you plan on rolling out Drive Wise next year? And is your intent to use that to augment growth, augment profitability or both?

Well, we don't disclose which states we're going to do it in, Adam. But I will tell you, we like what we see. The customer uptake is a little higher than we thought it would be initially. We think we have a slightly different offering than other people because the device stays in the car as opposed to coming out of the car, which gives us the ability to have a different customer relationship, which says that it should work on both of the elements you've talked about. Adam Klauber - William Blair & Company L.L.C., Research Division: Okay. And one follow-up question. It appears that you make more money this year than you have in probably 4 or 5 years. How is that going to reflect when you think about the buyback for next year?

Well, obviously, I can't comment of what we'll make this year because we're still trying to sort out with Sandy. But I'd like that to be the case, obviously. And let me maybe just go up to capital. We are obviously well capitalized. Today, we have plenty of capital, any which measure you want to put it at. We also generate substantial capital. And we have access to capital in the marketplace. So we have 3 sources of funds from which we could buy back shares or use our capital to grow the business. Obviously, our preferred option is to grow the business particularly in the auto business where we're getting very high returns. Occasionally, we make what I would call, modest acquisitions, whether it be Esurance or Encompass or the Benefits business we bought. And we've gotten good returns on those, but those tended not -- they tend not to be a big portion of our utilization of capital. Of course, as you know, about 80% of the capital we returned to shareholders through dividends or share repurchases. And that has been our history for a long time. When you look at dividends, we, of course, make the decision -- the board makes the decision every quarter. Right now, when you look at our dividend rates, whether that be payout or yield, they look a little low relative to where our stock price is and earnings are. So we'll consider that in the beginning of the year as to what we should do with the dividend. And then on share repurchases, we're not obviously done with the current program. When we are, we'll look at all the sources of capital we have, either the capital on our balance sheet today, our capital generation potential or ability to access capital because as you know, there are some pretty attractive markets today with our equity life securities to use that to restructure and lower our cost of capital. So that's the way we think about it. When we get done with this, which we're pretty close now, we'll be thinking our way through them.

Our next question is from Michael Nannizzi of Goldman Sachs.

Most of my questions have been answered. One question I just have, when you look at insured losses that will come out of Sandy, do you -- how much of that end up getting absorbed or do you think will end up getting absorbed by the National Flood Insurance Program?

Really hard to tell, Mike. We don't really -- it's almost impossible to tell right now. I would tell you but you want to go way up. The National Flood Insurance Program ends up being a net drain on the U.S. government. It really is -- if it was your business, you would seek to restructure that business charge of the accurate prices relative to where people's houses were. You'd try to expand the coverages so you weren't only ensuring those houses that get flooded 2 and 3 times. So there is -- I think -- and Congress has struggled with that over time, but their losses are -- the net losses will probably end up being bigger than they would need to be if you brand it as a private business rather than as a political entity. I can't give any estimates for the exact number. I have no idea.

Great. And then just maybe one on Allstate Financial. I mean, it seems like, I mean, you're clearly moving towards a more underwritten profile as opposed to a spread business. I mean, what is your kind of expectation for that business longer-term? I mean, it doesn't seem like other than the life products, now it doesn't seems like there's a lot of operational synergy or overlap with your kind of captive agent platform. I'm just trying to understand kind of where do you see that business moving as you execute your transitional strategy there?

Okay. Well, we look at it a couple of ways, Mike. First, we look at it strategically as it relates to our customers, and then we'd look at it to say where do we get a decent returns. So strategically as it relates to those personal touch point [ph] down in the lower left-hand corner, want advice and a relationship and want branded products. They want to buy all things -- more things from one person rather than multiple people because they don't just feel like having 15 different relationships. Life insurance, some retirement products obviously fit into that category, and we do well there. The life sales, I think are up 6% or 7% this year so far. And we like what we have there in terms of bundling that together. That's good for the customers. That said, it doesn't mean we need to make everything we sell. So we sold the VA business in 2006. Our fixed annuity business today, we use a lot of outsourced products. We have some proprietary products still in that channel, but we use outsourced products. We could -- on the Annuity business, if you don't like the return, now we'll just sell an outsourced product with someone who has a different return objective on it. On the Life Insurance business itself, just life insurance policies, we're getting pretty good returns on our capital well above our cost of capital there. And so we've done that. But we try to separate it into 2 things: what do we need to do for our customers, and then what do we need to do for our shareholders. And we manage our way through that. And that's why we're taking the size of that annuity business down is we don't need to have shareholder capital employed in that to execute our strategy.

And our final question today is from Randy Binner of FBR. Randy Binner - FBR Capital Markets & Co., Research Division: This is another Allstate Financial question. But I appreciate all the commentary on trying to kind of redeploy capital out of particularly the annuity business into what you've perceived to be higher kind of better risk-adjusted returns in really the other protection businesses. And so, I guess just as I think about modeling to the extent that you're successful in kind of getting folks to give up those annuities and then you're able to free up capital from the annuity business, that takes time, and then reallocate it, I'm just trying to think if there is going to be kind of a donut hole or a lag in the earnings that you generate as you kind of transfer that capital around. And so, you know, I guess I'd be kind of interested in your thoughts on that approach, kind of how that affects 2013 earnings. And then I'm not sure if I've heard this in the commentary, but are you thinking about kind of more aggressively trying to go after annuity holders and maybe giving them lump sum payouts or other incentives to surrender?

On the last part of your question, the answer would be yes, we are actively looking at that. On the 2013 earnings capital generated from accomplishing that, I would say it's not likely to have a meaningful impact on overall earnings. So if it -- if you just -- you could do the math. For every $100 million of earnings of Allstate Financial, even if you redeploy it into some of that business, we could redeploy it get a higher return than what's it's getting today, even in investing. But it would not be meaningful to force. Randy Binner - FBR Capital Markets & Co., Research Division: So I guess it happens gradually enough over time that there isn't a big block of capital that go -- it needs to get upstream then reallocated and then earned against. It's more of a gradual process because -- I guess in our model, it seems like as you move capital out, there's a little bit of a lag, but it sounds like it's more of a gradual process from your perspective.

I would say -- yes, I'm sure there's a lag, but it's not in terms of the overall scope and projections of the company. If you would change auto frequency by 0.5 point, it would have a bigger impact than this would. So you're right. But I would just say as it relates to Allstate Financial, we're approaching it as a "do it smart, do it when you can." And so don't look for sort of a simple, easy solution, where it all goes -- you just change it all. And it's got a kind of do it by line. So Don got the annuity business broken into a whole bunch of different segments, some he cares more about than others. And those that are the biggest in the hole he's working harder on and those that are earning a positive or above cost of capital return. So it's highly segmented, this strategy. Let me close and just say, first, thank you. We had a good quarter. We're on track for all of our core priorities. If those of you who are still on the queue, Bob is available today, working all day. So feel free to reach out or he'll reach out to you. And as it relates to Sandy, we're working hard to do what our customers pay us for. We'll provide additional perspective on loss estimates when we can give you a more informed perspective. So thank you very much, and we'll talk to you next quarter.

Ladies and gentlemen, thank you for participating in today's conference. This concludes the program. You may now disconnect. Good day.