Welcome to the Progressive Corporation's Investor Relations Conference Call. This conference call is also available via an audio webcast. [Operator Instructions] In addition, this conference is being recorded at the request of Progressive. If you have any objections, you may disconnect at this time. The company will not make detailed comments in addition to those provided in its quarterly report on Form 10-Q and Letter to Shareholders, which have been posted to the company's website and will use this conference call to respond to questions. Acting as moderator for the call will be Matt Downing. At this time, I will turn the call over to Mr. Downing.

Good morning, and welcome to Progressive's conference call. Participating in today's call are Glenn Renwick, our CEO and Brian Domeck, our CFO. Also on the line is Bill Cody, our Chief Investment Officer. The call is scheduled to last about an hour. As always, our discussions on this call may include forward-looking statements. These forward-looking statements are based on management's current expectations and are subject to many risks and uncertainties that could cause actual events and results to differ materially from those discussed during this call. Additional information concerning those risks and uncertainties is available in our 2011 annual report on Form 10-K and our quarterly report on Form 10-Q issued during 2012, where you will find discussions of the risk factors affecting our businesses, Safe Harbor statements relating to forward-looking statements and other discussions of the risks, uncertainties and other challenges we face. Each of these documents can be found via the Investors page of our website, progressive.com. Lenny, we are now ready to take our first question.

[Operator Instructions] Our first question is from Michael Zaremski with Credit Suisse. Michael Zaremski - Crédit Suisse AG, Research Division: So I was curious. So with rate increases matching the rise in severity, but seemingly above the combined loss cost trend of frequency plus severity, would your targets be to drive margin significantly below 96 to provide a cushion for future loss cost spikes? Or is it more to get back to 96 in order to begin growing PIF again?

More the latter, 96. We're never so perfect on all of that targeting, nor do we know exactly what future trends will be. But if we can just quickly recap through the year, let's just focus on auto, we've taken about 6 points of rate, give or take. A little bit differently, agency and direct, but that's reasonably close. Second quarter, I think we explained very carefully that we got a little bit behind on timing so we've taken some additional rate during the second and third quarter. And as I said on my letter, I think we've got a nice matching now. We will continue, that doesn't -- that's not a onetime issue, so we'll continue to take trend and do our normal rate revision processes. So rate revision processes are completely dynamic, happen all the time. What really happened in the second and third quarter was just a little bit of a catch-up for, perhaps, some trend that we may have gotten just little behind on, but our targets remain at 96. And as I said in the letter, while we're thankful for the inflationary premium growth, which is certainly apparent, we'd love to get unit growth and that's what we seek. We seek the combination of growing and good margins. So continue to try to think of us as a company that's going to grow where the opportunity exists and hit our target margins, but target margins come first. Michael Zaremski - Crédit Suisse AG, Research Division: Okay. And as a final follow-up, the expense ratio seems to be cracking recently below historical trend. Can you talk about the drivers there and how long these levels could persist if at all?

Yes, Brian.

You're right that the expense ratio in the third quarter, in particular, in the last several months, has been a little bit lower. A big component piece is in the third quarter, we consciously decided to reduce a little bit of our advertising spend relative -- certainly to what we had [indiscernible] coming into the year and actually, or spending on a little lower rate than last year at this time. So some of that is contributing to the lower expense ratio and you could see that more obviously in the direct channel expense ratio versus the agency. And then the other component piece is just with the earned premium growth being at higher levels than we've experienced for a number of years, we're gaining some leverage on the higher earned premium relative to salary costs and other fixed expenses in the organization. But it's the combination of the lower advertising spend plus higher earned premium growth rate that's really driving it. Michael Zaremski - Crédit Suisse AG, Research Division: So that one, that lower ad spend, I believe, at the Investor Day, you might have alluded to you not being as happy with kind of what you're receiving the -- from the ad spend, or could you comment on that?

It was less about being comfortable with the yield on the ad spend. It was our decisions here in the third quarter were more trying to make certain that we hit the profit objective, that 96 combined ratio that we talk about quite frequently and want to meet. So it was a conscious decision more there than the yield or the acquisition cost for cost -- the cost for sale versus targeted cost for sale was slightly higher than our targeted, but that wasn't the real driver. The real driver is trying to ensure that we meet 96 combined ratio for the year.

Our next question is from Vinay Misquith with Evercore Partners.

Just to follow up on the expense ratio on the advertising expenses once again. So the expense ratio within the direct platform has averaged roughly 20% to 21% over the last few years, but that spiked for the last couple of years to around 22%. What do you think is the more normalized rate? I mean, does your -- the expense ratio within that channel has come down. So would you think the roughly 21% rate is the more normal ratio, or is the 22%?

Well it would be -- it's tough to say normal. Here's what I'll tell you would be the drivers. When we talk about this, a fair amount in terms of our ad spend, assuming we are meeting our overall profit targets, we would base it upon the yield that we get in terms of our acquisition costs, cost for sale versus what we call our targeted acquisition cost. And while we're comfortable with that, we'll keep spending. If we go above it by a lot, we would then revisit those spend levels. And so -- and we do that on a regular basis. It can be monthly, quarterly or et cetera that we're evaluating that. So that would be one component piece, and so it's pretty dynamic. We believe, as we are much more comfortable with our aggregate rate accuracy, that we will actually go back to higher levels of ad spend. But we'll see how that plays out. But that is the thought at least right now, that although third quarter was down and fourth quarter is likely to be a little bit lower, we will return to a higher ad level spend as we get more and more comfortable with our rate accuracy. A second component piece is, what I mentioned in terms of gaining some leverage on fixed costs and other, call it, employee compensation cost, et cetera. We've got, over time, we definitely aspire to grow earned premium growth at faster levels than those costs, so that should have a long-term decrease in terms of our cost structure, both in LAE and expense ratio. Then the final thing, which is very, very important in a direct channel is the new renewal mix because remember, we have very different targets for new business versus renewal business and we allocate all of our advertising costs in new business. If we don't get the yield as much in new business, yes, we would reduce or adjust our advertising spend. If we are getting the growth, that Glenn so referred to earlier, we will continue to spend as long as we're meeting that new business target and renewal business target. So I'd say it's a combination of rate accuracy, we'd spend more. Other expenses, long-term, we actually hope to gain some leverage on. So what will the run rate for the expense ratio be for direct going forward, I think it's a combination of those effects. But, that's ...

Sure. Yes, just on the advertising expenses, once again. Next year, do you expect to increase the spend more than the growth in the top line, or sort of in line with the growth in the top line?

I'd say probably pretty closely in line with expected earned premium growth.

Okay. Fair enough. So the expense ratio impact on that would be about flat year-over-year, but higher sort of on the absolute dollar spend?

Yes. But it's not going to match one for one, but closer to earned premium growth than in certainly in when we are increasing advertising spend 20% or 25% a year, plus perhaps some earned premium growth that was driving the direct expense ratio higher in past years. Now, it's more closely -- more closer in line to earned premium growth.

Fair enough. Then the other question was really on the margin of safety. So I believe management has always said you have a 96% target on the combined ratio, but you -- but historically, you've had a margin of safety that resulted in a combined ratio being closer to 92%, 93%. With the introduction of Snapshot, do you think that you can now reduce the margin of safety because you can now model your business better? Or do you think you still have to keep the historical margin on safety?

Vinay, I don't think you should sort of take into the margin of safety issue you're calling at the 92%. When the marketplace is such that -- and I don't want to give a long history of why the marketplace has acted the way it has for the last decade or so, but certainly when it's available to us, we will continue to grow, and if we can make a 92 or a 93, then that's a combination we would accept if the market conditions were favorable to do so. I think what I've been saying for quite some time now is the market conditions are much more in the mode of we will be closer to our 96 to be able to continue to grow. So think of us not so much as trying to build in a safety margin, but trying to be as aggressive in the marketplace as possible, but put a very high premium on attaining our 96 or better. Snapshot really doesn't change that. We're pricing Snapshot product or the discounts that we embed from Snapshot so that we hit exactly the same type of targets. Giving discounts is a very, very tricky thing. You're clearly reducing the margin on that particular, or the price on that particular customer, but we expect to make the same margin. So don't -- I would just encourage you, don't think about sort of a 92, 93 as some act of built-in safety margin. Think of us for what we say we're going to try to do, is grow as fast as possible at a 96% and everything we do, whether it's Snapshot or a Name your Price or anything else, we design to meet the same targets.

Our next question is from Josh Shanker with Deutsche Bank.

Looking in the Q, I see a combination of things, one is that your new policy applications after being up for the first half of the year have shifted down. I'm wondering if you can comment on whether it's accelerating or whether negative 4%, given the new rates, is kind of where it stands? And two, it's difficult for me to turn renewal application growth into a retention statistic. You guys are a bigger company than you were a year ago. Are retention statistics more or less the same than they were a year ago, or has something changed?

Josh, on new applications, you're right. Clearly, the rate pressure, I mean, pretty simple physics here. Took rates up, we see the effect in new apps. The 4% that we've put there, not going to forecast it, but until we see some competitive actions that are even stronger than the ones we've seen, and clearly, we are seeing competitive actions that are, in the vast majority of cases, positive, but I expect that we will see, for some time, pressure on new applications. I'm not going to give an estimate of what that will be, but that's not going to turn around any time soon. That will take a little bit of market force. But given that we're going to continue to focus on our target margins, we are very comfortable with that trade-off. Renewal apps, yes, you see. You know there's sort of a reason that will be. We obviously put on new business, it's coming through as renewals. We've not -- we don't give specific PLEs by different classes of business. But we've given some indication that, at least for our Agency book of business, policy life expectancy is actually increasing. And for our Direct book, which had actually been experiencing some negative, it's now much more of a flat PLE. So while there will be a delayed effect on that as well, and I'm not going to forecast exactly what they will be, as of right now, we're very comfortable with our retention statistics for the customers that have actually experienced rate increases as well. New business probably more concerning and I expect that, that will turn around at some point, but I'm not going to forecast that point.

And you've mentioned that, I think the word you used was integral, that Progressive Home Advantage is an integral part of the product offering. I think that's the word you used, I may be mistaken there. But can we go into a little bit -- is there any metrics you can give us about the success in implementing Progressive Home Advantage? I know it's been slower on the uptake than you guys would have wanted starting back 3 years ago. Do you have some success stories you can share with us?

Let's break it into 2 pieces, our Direct piece and our Agency piece. And I don't want to sort of necessarily recount any statements that might have been made, but actually, we're quite comfortable with what's happening on, certainly, on the Direct side. Recognize that one of the key objectives there, we'd like to bring in new business to PHA, but one of the things we've always set out to do is make sure we don't lose current business that we have when there is a life change that requires someone to get a home policy. And about 50% of our business in PHA are our own customers maturing into greater insurance needs and us being able to accommodate them. So frankly, we're actually very happy with PHA on the Direct side. Second point would be on the Agency side where we very clearly said, we did not get out of the box, a good solution for agents. The company was fine, it just wasn't suited for agents. As you know and announced in June that we've entered into a relationship, a significant relationship, with ASI, a relatively small company, but very much in tune with, culturally, where we are and the sorts of things we want to achieve for our agents. We have been taking, over the last several months, the notion of Progressive and ASI as a little bit -- a lot more of an integrated offering for agents. Clearly, not the same as if it came from one company. But in this case, 2 companies that are very, very good at what they do, coming together, providing some, not only technology that integrates the product, but actually some product features. And that is being very well received by agents. I have personally been out and talking to…

Is there anything about the product where ASI has become your preferred vendor compared to your other partners?

For agency, it is our preferred vendor for agency. For direct, we'll have multiple carriers that we plan to use. But on the agency channel, this is one where agents have a workflow that needs to be respected and the technology that we've worked on together to present our products together, to present sort of data sharing, so on and so forth, it is absolutely our preferred option in the agency channel.

Our next question is from Rob Ryan with UBS.

What are your observations on Hurricane Sandy? And what would you like investors to know as it relates to auto insurance in general? And specifically, Progressive?

Sure. I think we all want to know sort of where Sandy settles out and I'm sure every insurance company right now is doing the best to give what they can. Let me tell you what I know and then try to avoid saying things that we simply don't know. Overnight reporting, I expect has probably bumped us up to about a 6,000 claims. So we've got about 6,000 claims, that's a little higher than the number I had last night, but I suspect that, that's well on its way to that number. Think of that as 2/3 flood, 1/3 wind. About just short of 80% of it is New York, New Jersey, the rest would be another 10% Pennsylvania, Connecticut, and then obviously, other states that you all know are affected. That's sort of what we know right now. Spud [ph] claims for us are probably going to be, more often than not, total losses. So order is not good for cars. So we have some estimates. We don't know exactly where and we have great models, but all models are exactly that. Don't know where the reporting will end. It's the advantage for you and for us. We will release results in about 12 days or whatever that is, 12 to 15 days, something like that, for October. And literally, as every hour goes by, the development of claim reporting is helping us. So we'll actually have a pretty good estimate of our exposure by the time we will produce results in about 12 or 13 days. So far, 6,000 claims. I don't want to even speculate as to where that is. If I gave you sort of a ballpark-ish thing, you would sort of maybe take it the wrong way, but I suspect we're seeing half of the claims we're getting or in that neighborhood. So, but there are people, clearly, as we all know, without power and so on and so forth. So these reporting patterns might be just a little bit different than the experience. As you would expect from Progressive and hopefully, all other insurance companies at this point, the job #1 is getting everybody back to the conditions that they can move on with their lives. So we have deployed and we are active, as you would expect, in the marketplace with resources and doing a great job for the people that have reported their claims. And hopefully, we can get as many of the unreported claims in within the next few days so we can get on those as quickly as possible. Hopefully, for any of you listening that were affected by the storm, hopefully, you're able to get your lives back to a situation that's comfortable for you or your families or whatever. So that event, insurance companies, I think in general, are stepping up and I know that Progressive will step up really well.

Great. Completely unrelated, the investment portfolio. What have been your actions considering the low interest rate environment? And what should we think of in terms of strategy going forward?

That's fair. We have Bill on the line so I'll let him take us through that.

Sure. Our actions have been consistent with what they've been historically, which is to try to look for value where we see it. We are not trying to hit the yield bogey, which is something nice about the way we run things here is there's no pressure to try to stretch to reach some artificial or hoped for number. We take what's available. And for us, that's been trying to find spread product primarily in, say, corporates or munis or CMBS or sectors like that, that offers some reasonable, relative value and some decent absolute value as well. That's become a little bit harder to find as spreads have come in from all sectors. But still in all, we have found a few pockets to invest in, and when we see those opportunities, we'll take them, and when we don't, we don't feel pressure to put money to work and take risks that we don't think are appropriate to take.

Our next question is from Meyer Shields from Stifel, Nicolaus. Meyer Shields - Stifel, Nicolaus & Co., Inc., Research Division: Two questions, if I can. One, with all the uncertainty surrounding health care reform and I guess the election, how are you, I guess, booking reserves? What inflation rate for medical costs are you assuming?

Look, that's perhaps, create a bigger disconnect from any other health care bills and so on and so forth -- the reform -- because that's just probably stretching it to believe that we could be that precise. The easy answer for you for our outlook right now on bodily injury, which is one of our clearly most important trend estimates, both for pricing and for reserving, 4 to 6, 3 to 6. Don't want to be vague there, but that's sort of a range that feels about right to us. So and frankly, if I go back in my career, sort of 3 to 5 is being a fairly good estimate of bodily injury trend which is largely reflective of medical costs as one of the underlying drivers of it. So that's our assumptions going forward. So what we've talked about, and I discussed in the letter, they would sort of taken a shot at rates, we've gotten back to margins that we think are about right, recognize trend is always ongoing, so we're going to have to price that in. And our current estimate of trend going forward is sort of that the 3 to 5, 4 to 6 kind of range on bodily injury. So far, frequency, if I had to sum up frequency, it's mostly noise, plus or minus 1% on most coverages. So not a lot of frequency as the big driver of things. That can get awkward if that's the driver. So it's mostly a severity issue, and I would say if you're using your models or your estimation or you're thinking about how Progressive might act, we're thinking around 3 to 5 in bodily injury trend. Meyer Shields - Stifel, Nicolaus & Co., Inc., Research Division: Okay. No, that's helpful. I don't know if it's possible to disentangle the ramp up of the Snapshot free trial from the issues you've had with raising rates and the impact on new applications. But can you talk a little bit about how the performance of this particular strategy has played out in 2012 and maybe 2013 expectations?

Yes. Probably mixed bag, to be honest, our volume of people actually taking the Test Drive has underperformed our internal estimates as we formulate estimates on everything even if we haven't done them before. And frankly, I would say this, volume-wise, is under our estimations. So just call it the way it is. Certainly, we got an awful lot of interest or clicks, but I'm not sure that, that necessarily is the operative measure. Of those that ultimately take the trial, the funnel that we had expected is a lot of the estimates we had made that are panning out. So actually, we are taking people all the way through the funnel, and the good news is that of those that are self-identifying and taking the test, a very large number of them are actually getting the opportunity to get a discount off of what our rates otherwise would have been. Who's taking the test? That actually is quite a nice positive for us. If I just focus on the 3 companies, the 3 auto companies that are, by market share, larger than us, we have more than 50% of the people who are taking the Test Drive are actually distributed amongst those 3 companies. So it is actually very encouraging for us that while Test Drive is not likely, nor was it ever designed, to try to get at the shopper who was going to be shopping in the next day, next week, next month. This was to try and get people to sort of think about shopping who would not otherwise have shopped, I would expect to get market share from companies that, perhaps, maybe had more entrenched customers. And that seems to be happening. So I'm very happy with the mix of customers that are taking us…

Our next question is from Matthew Heimermann with JPMorgan Chase. Matthew G. Heimermann - JP Morgan Chase & Co, Research Division: I guess one question I had for you was around advertising. If you all chose, depending on the effectiveness of the rate increases and where margins settle out on a state-by-state basis. If you choose to take advertising up, what -- how should we think about the normal, kind of lag, between when you spend and when you actually see the growth show up in terms of policy counts and then ultimately, in earnings? And I guess, the reason I'm asking is because the data that I've seen, which is actually a little bit more focused on life insurance than PNC, is that it's, for Direct response that it's often -- it often takes 6 to 8 months to actually get scale in the response rate. So just trying to figure out that dynamic.

Sure, Matthew. For us, when we change our advertising spend, we actually see a fair amount of effects pretty quickly. I wouldn't say it's a 6- or an 8-month lag that life might have. And it's not like one-for-one. But when we reduce our spend, we generally see a decline in prospects. And when prospects -- prospects being people who quote with us --and likewise when we increase it, we see a return, but I would also say it probably does differ a little bit based upon media type. What I mean by that, for example, if we were to spend more or less on paid search, where people who are looking for car insurance, you know there are shoppers in the marketplace, there, you clearly see a very fast reaction to increases in spend and bid prices and ranking and people quoting with you. TV, radio and other media types might be a little bit of a different immediate response versus lag function. But when we -- generally speaking, when we change our ad spend, we will see a corresponding activity in terms of the number of people quoting with us fairly fast. Matthew G. Heimermann - JP Morgan Chase & Co, Research Division: Okay, that's helpful. And then, I guess the other thing is right, there's obviously a lot of questions around frequency and severity, and there is, for a whole host of reasons. But just curious, when we think about -- new car sales have picked up, it looks like people are -- the car fleet's actually growing in addition to seeing replacement. My presumption would be, and I might be stretching with this, but my presumption would be that most new car sales are either leased or financed, and that the lenders generally acquire, I think, lower deductibles than one might choose if they potentially owned the car. So curious if, one, if that deductible difference really exists? And if so, is that something that potentially, on a kind of -- on a reported basis, potentially makes the severity look worse than it really is, just on a mix basis?

I'm not sure I have a really meaningful answer for you there, Matthew. I understand, but I don't think our deductible mix has changed very much for a good number of years. So I think to see that flow through into severity numbers, you'd really be stretching it. Matthew G. Heimermann - JP Morgan Chase & Co, Research Division: Okay. That's fair. I do -- it was just the only thing I could think of to ask about that probably was independent of frequency.

Our next question is from Josh Stirling with Sanford Bernstein. Josh Stirling - Sanford C. Bernstein & Co., LLC., Research Division: I thought I'd start and ask you, you were just talking about sort of ad spends slowing. I'm wondering if you can give a sense of the -- what seasonality has to do with that. And whether the election cycle and sort of add to the cost of ad spend sort of influenced your thinking? And whether we should be thinking about sort of first quarter as likely a bigger quarter for you guys us to spend advertising, and presumably, as consequence get more impressions and prospects?

Those are all issues. But I would tell you, think much more in terms of what we've already said that where we didn't have rate adequacy, we pulled back on our advertising. Certainly, there is no great reason to advertise to bring in customers if you don't think they're making a target margin. That is the biggest issue -- just think that issue. To the extent that we all know that certainly, if you live in Ohio, political advertising is a big part of what's going on, and that will come to an end here reasonably soon. Not a major factor for us. That's much more of a local buy. A good chunk of at least our television is national buys. So much more driven by our own view of do we have the rate? And if we have the rate, we'll advertise. If we don't have the rate, we'll pull back. First quarter is a big buying period for auto insurance. Always has been, always will be. So expect that the fact that we are signaling that we feel much more comfortable about our ongoing rate levels, that we are very consciously being prepared for the first quarter. Josh Stirling - Sanford C. Bernstein & Co., LLC., Research Division: That's great. The thing I'd love to get your color on related to Snapshot, is it sounds like we're making progress on sort of using it as an underwriting tool, getting than average drivers. It sounds like it's either margin neutral if not expensive. The question is, is around getting customers to pick up the phone. And it sounds like we've worked through a few different iterations. You've got something on deck. But you've got 2015, 2016, there's sort of the patent cliff as it were around sort of the window of time that you guys have to take advantage of this. And I'm just sort of wondering, is there something that you think they're working on that will ultimately be -- make this the killer app?

Totally understand your question. And if you were inside Progressive, you would see the same sort of energy on everything. The fact is, we'll keep working at this with the kind of diligence that we work with everything else. Every time we get, we've got now more than 1 million customers have taken Snapshot. So it's not some sort of small data set. We learn things every time. We are learning things about the underwriting, the discount, the measurement period. So expect to see this be very dynamic in change and reflect. You mentioned this sort of cliff, that's a self-imposed cliff. We have said very clearly that we had a lot of intellectual property. In fact, our sixth patent, we've been told that our sixth patent will be awarded here very shortly. And we've been very clear to the marketplace that we're willing to license this starting the second quarter of 2015. And between that time, we'll be very clear that we don't think people should be infringing on the things that have been a significant amount of energy and work for us and have that patent protection. So we'll protect ourselves during that period of time, but we invite them to show interest in licensing if they so wish. By that time, we'll just continue to push, push, push. But this is clearly something that is very exciting. There is no question. There can be no question. This is an underwriting variable that gives us insight to driving behavior and segmentation that no other variable has given us to date, at least in the same way. How we get that through to people and how we market it is something that we're going to keep pushing at. And I wish I could get you an answer and say, "Yes, we have got sort of the killer app in our back pocket. " It's not likely to be that way. The more people try it, as I did note the awareness level of it, that's a pretty -- maybe I'm kidding myself, but I think that's a pretty cool thing to have more than 50% of the consuming public aware of something like this. Now, we just need to turn that awareness into action. We're going to keep plugging away at it.

. Our next question is from Adam Klauber with William Blair. Adam Klauber - William Blair & Company L.L.C., Research Division: With applications falling down, what's the time lag until we see PIF slowdown?

Well, I think you've already started to see PIF growth slowdown. So and it's a combination of declines in new business and then how many of your existing policyholders retained. And as Glenn mentioned, so far, our retention levels have stayed pretty strong, but I'm certain we will see some decline just based upon rate change because this is a price-sensitive product. But in terms of rate of growth, in terms of policies in force, you've already started to see it start. Adam Klauber - William Blair & Company L.L.C., Research Division: Okay. And then you also, I think in the 10-Q, look like quoting and direct was down. Was that more of a function of increased rate? Or lower advertising? Or a combination of both?

As it relates to the quoting, that would be more relative to ad spend and response rate to ad spend. When we talk about conversion, that is more relative to how many convert into sales and that change, in terms of conversion being down, would be related to the rate changes. And it's pretty clear where we have seen most of our decrease in conversion is -- are in states where we have raised rates. So it's pretty clear. Where we have raised rates, we have seen a decline in conversion, and states where we have not had to change rates much, we haven't seen a decline in conversion at all. Adam Klauber - William Blair & Company L.L.C., Research Division: Okay. And then as far as Florida PIP, it looks like frequency, I think you mentioned, was down 7%, severity was down 2%. Is that due to rate and underwriting actions taken hold in the quarter, or is that just some good trend during the quarter?

The frequency and severity you referenced actually was for all PIP states. But Florida, as you may or may not know, has a lot of things going off on it which was a great a number of, what I'll call, reopened PIP claims based on some district court rulings and they came at a level that was very hard for us to predict. That seems to be -- and I'm just going to say, that's more under control now. So, and we're getting a better handle on what those true costs will be. So that's less concerning. Secondarily in Florida, there have been a requirement for state -- for carriers to file a PIP adjusted rate effective -- what was it? November 1 -- or October 1. October 1, even though the rules and regulations that actually it applies to don't start until January 1. So we've taken those rate adjustments and we feel reasonably comfortable with it. But in general, we're seeing much more controlled severity in Florida PIP, but if you know anything about PIP, you don't want to make too many assessments that are too long term. So far, it seems to be a little bit quieter in the last 60 days.

Another state, I'd say, was a bigger PIP state and has seen a decrease in frequency is New Jersey and we've seen that for a couple of quarters now. Adam Klauber - William Blair & Company L.L.C., Research Division: Okay. And one final follow-up. I think in the Q, you highlighted that favorable reserve development was mainly from 2009 and prior. What years, and I guess, what's changing those years right now?

Al Nees [ph],our corporate actuary's here, he can respond to that.

A large portion of the older accident years open reserves are set at policy limits, that people have a 25,000 limit and of course, those as they settle, you're not going to have any settle above that amount, so you have favorable development. So we've been adjusting our structure to compensate for that in the next coming through as favorable development.

[Operator Instructions] Our next question is from Mike Nannizzi with Goldman Sachs.

Just one question, Glenn, you mentioned just sort of claim counts -- certainly not holding you to anything. But just I'm just curious like if you were to kind of look back at Irene, what sort of claim counts did that end up -- did you end up seeing from Irene?

Ooh, some [indiscernible]. Yes, it sounds about right. Know that you're catching us here without specific data, but I think we have 3 people here would know, and let's call it 7,000 or 8,000. [indiscernible] bigger than that.

Right. And I'm sure from a tracking perspective, it's certainly higher. That makes sense. And that is -- that's commercial and personal, I assume, together?

It is. And think of the vast majority of being, what I'd say, 6,000, say, at least 2/3 are auto. We're going to have boat claims that are in there. And all of our other products have claims. But the vast majority are auto and commercial just in terms of unit counts because it's not a very large unit count. It's just not going to be quite as significant in terms of actual individual claims.

Okay, great. And then I guess one question kind of emanating from your letter. Glenn, I mean you kind of talked about margin improvement. I know you look at margin, and we kind of focus on earnings and there is certainly some obvious relationship there. But as I look at cats and kind of loss trend we've seen and the rate that you're pushing now, in order to get either -- to get to ROEs that -- or even just earnings growth that makes sense kind of given your relative evaluation, I'm just trying to understand, how do you think about the levers to pull to give you either exposure growth that will lead to earnings growth down the road? Or the benefit of margin improvement from rate gains even though that doesn't come with exposure improvement? And how do you think about those and prioritize those as you manage your book?

Yes, clearly, there's a lot of moving parts if you wanted to use them. But I don't want to be too dry in my answer. But we're going to try to grow our business at or better 96. And that's what drives everything for us. I'd love to think that we could count on higher investment returns, so on and so forth. You know that story as well as I do. We're going to stay the course, grow our underlying book of business and make our underwriting margin. And then under different circumstances, which right now, none of us can foresee, we may have a different proportion of earnings coming from investment income. But think of us as a "grow as fast as possible at a 96". And that, for me, is the common denominator that will keep us on track and we'll weather storms, whether or not the earnings from other parts of the business come in or not, that will be the one constant.

And the other thing I would say is just we remain committed to how we manage our capital and in terms of always needing and wanting to have sufficient capital to grow the business. And, but when we think we have more than we need for that plus contingent purposes, such as hurricanes, we will return it. And so that was a major driver for why we decided to clear the special dividend of $1. Because we want to efficiently manage our capital, which might not be your earnings question, but certainly, when you think about return on equity, it addresses that.

[Operator Instructions] Our next question is from Ian Gutterman with Adage Capital.

I just wanted to follow up on the earlier PIP discussion. I just wanted to get a little bit more as to what the puts and takes are as far as pricing and advertising versus what should be positive from things like Snapshot and the home expansion? And when I look at, I guess we do Direct first. Direct, this is the first time, at least in my numbers, that I think you've ever had 3 straight months of sequential PIP decline. And it's not the first time you ever raised rates. So I mean, why is the response been so much more dramatic even in the face of this sort of secular growth from Snapshot we've been expecting?

Well, probably just size of the Direct book now, I mean, we're just not going to see the same sort of PIP growth that we may have seen when we were smaller. I understand your question, but the competition is well understood. We are fighting for every app we can get and some of our marketing initiatives. We are positive, those would be your -- the puts. But right now, the take is definitely rate. And as Brian said earlier, it's a very rate-sensitive business. So while we look forward to turning that around and having some of the initiatives, Name your Price continues to work very well for us. Snapshot, we continue to be very excited about. We like Test Drive to, volume-wise, to drive some more business. But I don't know that I can be any more specific on balancing out the puts and takes in that. Right now, the take is absolutely rate.

Okay. And then on the Agency side, it's somewhat similar. Again, I was just looking at the last time you saw a sequential quarterly decline in PIF of this magnitude was the fourth quarter of '08 when you were showing negative overall PIF growth in agency for a couple year period. I mean, does this imply we're going back to shrinking PIF in Agency until people catch up to you on price?

It may. It may. But also understand that the Agency business is very elastic and with comparative raters, any changes across the board by other competitors go through very quickly for the agents. So we expect that we will get the benefit of favorable pricing as others change their prices to keep up. But impossible to know exactly when that's going to happen. In addition to that, the comments I made with regard to our penetration or our planned penetrations and more preferred book of agents with homeowners, that will be relatively small, but it will be significant for us over time.

Okay. It feels that a little bit from your comments, maybe I'm reading too much into it, but that elasticity has picked up versus prior periods when you've led the pack on rates?

I wouldn't say picked up in the recent time, but picked up over a 5, 10-year period. Comparative rating and the number of apps that we get that come from comparative raters, I think in the Q, we actually tell you, we get -- we got more looks from just the mere fact that we're exposed on comparative raters. This is an issue of conversion and just rate.

And I'm currently showing no further questions. This concludes the Progressive Corporation's Investor Relations Conference Call. An instant replay of the call will be available through Friday, November 16, by calling 1 (866) 467-2403 or can be accessed via the Investor Relations section of Progressive's website for the next year. Thank you for joining. You may disconnect at this time.