Welcome to the Progressive Corp.'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, quarterly report to shareholders and letter to shareholders, which have been posted to the company's website, and will be -- 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.

Thank you, Caroline, and good morning, everyone. Welcome to Progressive's conference call. Participating on today's call are Glenn Renwick, our CEO; Brian Domeck, our CFO; and Bill Cody, our Chief Investment Officer. The call is scheduled to last about 1 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 they 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 2013 annual report on Form 10-K, 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. That document can be found via the Investors page of our website, progressive.com. Caroline, we are now ready to take our first question.

[Operator Instructions] And our first question comes from Michael Nannizzi from Goldman Sachs.

Just one question, Glenn, to start out. You kind of talked about the personal book outside of auto and that was running at a comfortable margin -- it implied it was running at a comfortable margin, can you go into a little bit more detail in terms of what you're seeing out of the non-personal auto book and what we should expect from that as you continue to likely grow that business?

Michael, you're referring to the special lines business?

Correct.

Yes, okay. There's 2 outside of auto.

[indiscernible] not commercial, yes.

The special lines business, as you know, is just a great business for us and this year -- every year has always got a story with it. This year, it was just a little bit surprising in the sense that the loss ratios ran a little softer, more favorably to margins, pretty much through the entire seasonality of the line. So it's really hard to sort of put a single issue, sometimes weather makes a really easy explanation. But I would just say there we actually had some nice favorable margin results. We don't necessarily think those will be reproduced. We feel good about our pricing, but it certainly also gives us the opportunity to be more stable with those rates going into next season. And as I indicated in my letter, while we're very, very happy with the business, we still think we can actually grow it a little faster than we've been growing the last couple of years and we've got the right rates to do that. So overall, it's just a really good story there. I'm not sure if I'm addressing your question. Commercial is a different story. And I think this is one where perhaps a little bit more what I'll call a classic play, at least we try to make it a classic play from Progressive. We get ahead of rate at least as early as anyone in the marketplace, we think. And commercial clearly started to see some trends, and John Barbagallo gave you some insight to that at the investor relations meeting some of which, frankly, we did some misses on. So we got a lot of that cleaned up in late '12, mostly in '13, with some rates, some better understanding of some underwriting requirements. And it was for that period in…

Great. And then, I guess talking about the agency channel a bit. I know you made a few comments in your letter. What about Snapshot there? I mean, I know you've got 1 stated lease where you making some headway, but is it possible to offer like a reward or something to the agent specifically for integrating Snapshot? I would think that from their perspective that price is going to be very competitive for the drivers that select that product. I'm just curious, what efforts have you made or what changes have you made this year? Or do you anticipate to try and address either via Snapshot or outside that you haven't already undertaken or you hadn't undertaken in the past?

Good. I appreciate that question. Yes, I did make some comments so there shouldn't be any ambiguities to how I feel about the reductions that we have seen in new business applications in the agency field. And while words are just words, the most important one is we should never be underestimated there. We have a lot of changes that we're planning to make, and I did state they won't change overnight. Let me focus on the Snapshot. I want to make sure I answer your question. And I'm going to tell you, ahead of time, I'm going to be a little more obtuse than I like to be, simply because we're hitting at a time where we're just starting to come to market and explain to others. So this is not the format to do that about the change. But I've indicated I think in at least maybe the last call and in my letter that we took some changes into the Massachusetts marketplace with Snapshot, and I'm referring specifically to agencies here, and the impacts were very favorable to us. We will release a new version of Snapshot that will be released into the marketplace late this year. December 5 is the first entry and then into next year. Now let me just spend a little time on that because I think it's important. With the amount of data that we've been able to get, now approaching 12 billion miles, we're able to do something that I consider very, very important. We're able to clearly confirm that it is an explanatory variable that can't have a proxy from 3 or 4 other variables. However, we can start to create a reasonable estimate of certain groups that we would expect to get a discount. And to the extent that…

Our next question or comment is from Cliff Gallant from Nomura.

I wanted to ask about the jump you saw in PIP. In the 2013 10-Q I think was a 17% jump. Do you have any insight as to what might be driving that?

Yes, let's -- maybe, Gary and Brian will provide a more detailed insight into that. It's absolutely correct to see that jump off the page in the Q. It was put there to give you, certainly, a first indication of the volatility of this type of a coverage. I think we've discussed that on many occasions. I will say that this one is going to be just a little tricky to totally explain because as you know, Florida, only a year or so ago was implementing House Bill 119. There are some changes there. As the market evolves, as our procedures evolve relative to what's in emergent medical care. Obviously, the real obvious ones are broken bones, lacerations, emergency room and so on and so forth, we're fine. But there are some others where it's not as clear. So those things are settling down in the marketplace and it may be that the denominator last year was a little different than what the real run rate would be. So again, percentages can be misleading. Michigan had a rough third quarter from a weather perspective and probably had a slight difference in trend in that quarter than it's had for the year-to-date. And all I'm doing here is just meant to -- we'll always tell you what we see. But in this case, it's a convoluted picture and it very much commands our attention. But maybe Gary and Brian collectively can you give you a little bit more insight into the actual trends.

Yes, this is Brian. I'll try to give a little bit more color and then Gary can fill in if necessary. I think there's -- I will try to break it down into various component pieces. Certainly, as Glenn mentioned, on a quarter-over-quarter basis last year, we actually had recorded that PIP that was variable was down 4% which is sort of not necessarily the norm for the PIP coverage to see some areas going down, but some of that was driven certainly by the regulatory changes in Florida. But last third quarter was a lower-than-norm severity. Another component piece is some of the aggregate PIPs of areas explained by geographic mix differences and the PIP coverage varies greatly by individual state. The limits are not the same, et cetera. So geographic mix composition makes up some of it. And for example, we are growing more in -- and faster in Downstate New York area, Southern Florida, and also in Michigan, all of which those geographies tend to have higher severities within their -- Michigan, just in general, and then those geographies within those states. So some of it is explained by geographic mix changes and where we're growing and then what you may call sort of the underlying severity trends. And it varies actually, again, a little bit by jurisdictions. In some jurisdictions, we are seeing more billings per feature. And in a few of those jurisdictions, we're actually seeing the charge per billing going up. We think the aggregate of those sort of underlying severity changes is more in the 7% range, so higher single-digit ranges. And that certainly has our attention and is higher than recent trends, but that is how I would sort of describe the various components. Some of it is low denominator, some of it is geographic mix, and then some is the base underlying what we would call base underlying severity trend. It differs a little bit by jurisdiction, but certainly, anything going up of that magnitude has our attention and we continue to dissect it. Product managers for the individual states as well as claims organizations in those states are paying great attention to it, and we're ready to react if we were to see these trends continue. But I wouldn't say 17% is the underlying severity cost. I would say the underlying severity costs are more in that 7% range, high single digits, and even that is high. So it certainly has our attention. And I think the 7% more closely probably approximates more our pace of severity trends that we're seeing in PIP. That's how I'd characterize it. Gary, if there's anything you want to add or?

This is Gary Traicoff, Chief Actuary. I just think Brian gave a nice summary. So nothing to add to that other than we think probably that 7% range is kind of what our true underlying cost is.

So those long-term followers of Progressive, the one thing that certainly has been a consistent statement we've always made is that PIP can truly be volatile and while the run rate might be considerably less than 17% and it's all a matter of your starting point, all that sort of stuff. This certainly was a, not so much a wake-up call because we're always watching it, but it certainly one that we'll be watching a great deal because PIP can get volatile quickly.

Yes. And just since we're on the topic of PIP, I would say, the other thing to note is that while PIP frequency and aggregate PIP is down, there are some jurisdictions, principally Florida, where we are starting to see an increase in PIP frequency and that covers all the geographies throughout the state for the most part. So PIP is very much a state-by-state view and different sort of outcomes in each of those states. But Florida, from both the frequency going up as well as seeing severity going up as well as some of the changes that are in the marketplace as a result, House Bill 119 in Florida very much has our attention these days.

Our next question or comment comes from Josh Stirling from Sanford Bernstein. Josh Stirling - Sanford C. Bernstein & Co., LLC., Research Division: I think you did a great job of being accurate but obtuse on the first question about your Massachusetts product, but I wanted to ask you a couple of follow-ups, if I may, on that. So it sounds like that we're going to get a new version of Snapshot starts going in December 5. When do you think we should -- when should we expect it to be rolled out nationally? Is it -- are these going to be agency-only changes? Or do they go into direct? And I'm wondering, I mean, I recognize you're going to be circumspect about this, but what can you tell us about what you'll be looking for, say, for this, your quote volume and close rates for making a switch to the product?

Yes. Let me see if I can hit most of those. Both channels. So the changes would reflect both channels. Some of the changes are the sorts of things, and you certainly have had at least as much interest in this as anyone so you'll appreciate it. The algorithms are always, always maturing. In fact, I will be obtuse here in terms of what we're looking at. But I sort of get -- it just feels really cool to me that we're now able to look at things that sort of would have seemed just totally ridiculous to think about few years ago. The kinds of things that we can see in the data, the regularity of certain behaviors that we wouldn't necessarily have even dreamed about a few years ago, we now actually can see and test those hypothesis. So the algorithm is just getting richer. That will be consistent in both channels. To the extent that we have already experienced a very different take rate in direct and agency, what I'm hoping for is that moving the -- or some part of the discount further forward will have a more dramatic effect in agency than it will in direct. I may be wrong on that, it may have equal effect in both channels, but for me the biggest thing is that it does have some effect in the agency channel. That's what we'll be looking for. So from your continuous study of this, you should be thinking about this as how are we maturing the product based on data. Because some of the things we could have done a long time ago but you wouldn't have nearly the certainty that you're headed in the right direction based on the data that we have today and then just the continuous evolution of our algorithm which will only continue to get richer as we put in even more features, including GPS into the Snapshot device which we have -- they had previous experience with but will actually use that more aggressively in the next several years. Josh Stirling - Sanford C. Bernstein & Co., LLC., Research Division: That's helpful, Glenn. Separately, I wonder if you can give us an update how you guys are thinking about your many... I'm sorry, go ahead.

I'm sorry, I missed your first -- the rollout will be starting in December, but first quarter will pick up momentum for many states and then we'll just keep the rollout going. Josh Stirling - Sanford C. Bernstein & Co., LLC., Research Division: Oh, great. I thought I was going to have to ask that one offline. So if -- just a quick one. Give us your thoughts on your mobile strategy as it relates to Snapshot. Obviously, there's a major competitor of yours who's now making a bit more noise about their 16 states. They have a mobile product UBI app launched. We all know you guys have had a beta test of the Snapshot iPhone app, but I don't think we've seen anything -- seen you guys launch anything widespread. And I guess, I'm curious, if you could help me understand, is this a sign of you guys making a different technology bet? Are others getting ahead in the sort of the arms race here? Or should we just stay tuned?

No, I can comment a little bit on that. Obviously, the chip is a great way to be sure that it's that vehicle and so on and so forth and we're much more comfortable with that as a reliable data set. I think I've said on numerous occasions, I would be just as happy to get out of the business of plugging things into cars. That's not really an integral part of our business. And as cars become moving IP addresses, that will happen. And we are very much, as you would -- could reasonably assume, very much in sync with the manufacturers that are making those kind of changes so that we will be there when that's available. But being out of the dongle business, that's okay by me. The interim of a iPhone or Android device, yes, you're right, we've had that. We obviously, therefore, match it against other data that we feel is reliable and we continually try to find something that we would be as comfortable that the proxy data from a device that may or may not be in the car at the time. It's the same device that's in your pocket when you're landing an aircraft. You're going to get data, but it won't relate to the driving experience. So it's not quite as simple as it would seem. And now we also have other vendors that we've invited to actually participate in a, I would say, a bake-off a little bit to see who can actually do even better than we may have done. So we're very interested in mobile devices. We are not sort of folding the tent on that one, but we're also not going to go with something where we think the approximation to data isn't as good as what we can get barring another source. When that threshold is crossed at some reasonable point, that would be another arrow in our quiver, so to speak.

Our next question or comment comes from James Engle from John W. Bristol & Co. James Allyn Engle - John W. Bristol & Co., Inc.: My question is on policy life expectancy. It looks like for 12 months, agent was flat and direct was up 6%. And for 3 months, agent was down 6% whereas direct was up 1%. Could you give us some color on that? And also, could you weave in what's going on with Progressive Home Advantage?

Yes, I'll first give some color on the PLEs and then I'll turn it over to Glenn to describe what's going on in the PHA program. You're right in terms of the numbers and the changes. We're seeing PLE growth continued in the direct channel and starting to see some degradation in the agency channel. It used to be that direct and agency PLEs were roughly about the same and now the direct channel has a few months higher average PLE than the agency channel is. And some of the degradation we'd see in the agency channel relates to a decrease in some of our renewal rates and some of that has been more recent, so that's why it's showing up more in the trailing 3 months rather than the trailing 12-month measure I'd indicated in May, that the 3-month uses more recent data. And I would say a lot of the change or a fair amount of change in the agency PLE is due to actually a little bit decrease in our renewal retention rates whereas we've seen some in the direct channel but not nearly as much. And part of why -- one of the things that is helping with the direct channel PLE is we continue to get more of our policies to have more than one product, and in fact that's a lot of strategy of the PHA program and the bundling there, and we're continuing to grow the percentage of policies in both channels that have that. But obviously, we're growing more in the direct channel. Glenn might want to share some more details about what's going on in the PHA?

Sure. I think there's maybe 5 points to touch on since most of you know the broad details there. It's all about getting multi-policy households and that is effectively continuing to go up. These are not massive increases on a quarter-by-quarter basis, but it's a great trend and we're very happy with that. We did add our 12th carrier to our PHA program in the last quarter and probably more important than any of those things is ultimately the issue of are we really bundling. So it's one thing to sell a product, but the real question is we were trying to sell bundles of home and auto, and the best measure of that probably we get the cleanest measure from our direct channel. And you could just think about 3/4 of the home and auto are bundled together of those customers. So while we will sell a home policy or a renter's policy, and it may not be with the auto at same time or vice versa, about 3/4 of those policy holders are households ultimately end up bundling. Now the development that will discontinue to take on more and more strength would be our in-house agency where we have relationships with some of the carriers that we offer a homeowner's product to our customers. I'll touch renters in a second, because that's a little bit different. We have, historically, been able to do that in an online mode or perhaps we will actually refer in some cases a customer back to our partner and their call center. We are now adopting more and more holding the quoting relationship inside Progressive and being able to make sure that we use the breadth of companies that we represent. If it's not possible on online mode, then we'll have an in-house…

Correct. Yes, with agents it gets considerably trickier to have more than one. But more importantly than that, we have just a terrific relationship with ASI, and are very happy with them and will continue to work with them. And there are probably things that we, again, we can talk about this time or our IR meeting next year, but there are opportunities for us to do some things with ASI that I think would be quite exciting.

Our next question or comment comes from Paul Newsome from Sandler O'Neill. J. Paul Newsome - Sandler O'Neill + Partners, L.P., Research Division: A year ago, we were talking a lot about medical cost inflation and the effects of the reforms there. Now it's a ways out, are there any lessons learned? Did we see anything that affected -- in your medical cost trends, that were affected by those reforms or did it become just sort of a nonissue?

Paul, I'm going to refer this to -- there's a couple of folks in the room. So I don't see anybody with a hand in the air with something important to say. So I think that -- and I mean that in a light way, I think that there's nothing that we can sort of pull out that says, "Hey, here is the one thing you can take away from the accountable care act and how its trickled through to our costs. And if you look at the very macro level, while I'm sure on a couple of calls maybe earlier this year I've talked about BI trend which is one that I tend to watch a lot more, maybe 3 to 5. I think you might strengthen that 4 to 6, but I don't know that we can relate that to the accountable care act or any form of reform. The bigger issue that Brian went into in some detail in PIP, I don't think you can sort of relate that to the medical card. That's much more specific to the individual state issues. So I would say nothing dramatic and any small effects, they're never going to be noticeable or detectable.

Our next question or comment comes from Vinay Misquith from Evercore.

Just want to follow up on the homeowners on the agency side. Glenn, why, in your opinion, do you think that the traction is not as strong bundling products on the agency side? And so is it that the branding of the homeowners needs to be better? Do customers prefer, I mean, a joint highly-branded product for the homeowners and the auto together, or is it something else?

No. I think we've just got some opportunities to really make our bundle work even better than it does today. So the acceptance with some agents is actually very strong. There's no question. This is a place where, unlike most things, we're the ones that are coming into the market a little bit later than some established carriers. But we've got to break in and we've got to make our presence meaningful. I think we've got every ability to do that, but that's not going to be one that we turn around instantaneously. But I would say actually we're gaining with agents on a regular basis and I sort of alluded, and that's all I'm going to do at this point, I think there's a lot more that we can do with ASI to really make our products very attractive as an even more integrated bundle. I like to think of this as not so much 2 products, but starting to sort of present at least to the consumer that this is just the insurance you need. It's sort of our problem to figure out which part of it falls into which line of insurance. And the more we can do that and make this a more viable consumer proposition, I think that's going to be our strength to market and strength with the agents. So being a little bit circuitous on your answer, I don't see any major roadblock there. In fact, agents have accepted this, consumers have accepted it. We need agents to accept it even more willingly. And I think the biggest issue there is they have options. And many of those options are good options. So let's just be realistic about that. We need to come to market with continuously better ones.

Sure. And do you think the commission rate for the agents is maybe a stumbling block? And would you consider raising the commission rate for the agents?

Yes, it's a little bit of a stumbling block only in the sense that others will pay more, and not so much that we think we're wrong and there are 2 parties to pay the commission. While I'm not going to get out over our skis too much, what we can consider other options, everything can be on the table as we think about a bundle that comes to market where we really, really want to play in this marketplace, but assume that, that will be part of something more comprehensive as opposed to just a tactical reaction to market commissions.

And the last thing, just a follow-up. On the order side, you're doing pretty well in terms of margins. Curious as to whether you see the PIP growth issue as a margin issue whereby you can maybe lower price a little bit? Or do you think it's more to do outside of margins and therefore you would rather just keep your margins where they are?

We have our target margins. We're very clear about that. You know what those are and when we do better than that, that's good. We acknowledge it. The thing that is always tricky and very easy to see from a macro level is why don't you turn the switch and put it a little bit here, increase growth is not quite that simple. I would just tell you, you should assume that, I for one, am not willing to give away margin unless there is a more than proportionate growth to be had. But in general, as we've said all along, we want to grow as fast as possible at or better than our target margins. So while we're reporting sort of 92 and change year-to-date and for the quarter, we've also given you some indication that there are some trends and severity that are starting to outstrip the frequency declines. So for me, my general outlook and this is very general because states are specific, we have a rate level that can stay in place for a while longer without dramatic change. And that's always a good thing for a consumer versus up and down. So I think we're well positioned as we go into the remaining part of this year, and most importantly, well positioned for the bigger season in the first quarter next year.

Our next question or comment is from Kai Pan from Morgan Stanley.

You mentioned that your biggest business opportunity is in the Robinsons consumer segment with bundled products. I just wonder, given the development in the reinsurance market was alternative capital coming in, have you considered that actually putting homeowners business on your own balance sheet that you can actually transfer some of the volatility while sort of basically growing your customer base?

Kai, not directly. And while if we would ever do something like that, you could reasonably assume there would be fair amount of reinsurance, I also don't overly pick a strategy based on markets that are available today but may or may not be at a different point in time. We will always consider our long-term strategy of being a destination insurer there will almost unquestionably be things we will do over the next several of years. But I'd much rather tell you what those are when we do them. But I wouldn't be jumping to any conclusions about writing homeowners in our paper.

Okay, that's great. Then on capital management, if you consider you have operating earnings about $900 billion -- $900 million each year and then how much of that would you say you need to retain for your future business growth versus than what's priority for the variable annual dividends or special dividends this year?

I'll let Brian take that one.

In terms of how much do we need for growth, we try to target insurance underwriting companies at a 3:1 premium-to-surplus ratio. In most jurisdictions we have that ability. We also have some additional capital at some noninsurance companies. But for the premium growth that we're talking about, and let's call it $1 billion a year or slightly more than $1 billion dollar a year, you might need $400 million of capital for the growth, $350 million or $400 million for the growth at that 3:1 sort of ratio. Then obviously, in terms of our capital management structure, we have the variable annual dividend which is certainly formulaic and it's 1/3 of the aftertax underwriting profit times our gain share. So that amount for the variable dividend obviously can change each and every year based upon both the underwriting margin as well as the gain share factor. I can just let people know as of September month end, if you were trying to calculate what the variable dividend through 3 quarters is, based upon our underwriting profit and our gain share factor, it would be about $0.49 per share based upon 2 or 3 quarters of activity. Obviously, we have the fourth quarter underwriting profit to go and the gain share factor will be finalized in December. But you can calculate the magnitude of the variable dividend based upon that. But those 2 components, the growth, $350 million to $400 million plus the variable dividend, which can change every year. And then, after that, then we have the opportunities for both share repurchases. And so far this year, we've repurchased around $235 million worth of shares. And then, we still believe our current capital position is strong but we feel good about that. It gives us opportunities that we choose to in terms of continuing the share repurchases and then consider other alternatives.

Just finally, what's your cash -- holding company cash utilization at quarter end?

Bill, the picky assets at the end of the third quarter were $1.3 billion?

Yes.

And that can act just like cash but the assets held out at picky are $1.3 billion.

Our next question or comment is from Mike Zaremski from Balyasny.

My first question is on the expense ratio. If I look, it's been slowly but surely turning down over the years and I'm aware if we go back many years the ratio has been at 20% before, but can you elaborate on what's been driving this reduction so we can better understand if it's -- it can continue or potentially reverse?

Sure. Brian, jump in on this one, as well. I think, as we talk about the expense ratio, let's break at least for direct, or well for both I think is fine, into 2 pieces: acquisition and non-acquisition expense ratio. I think that's an important distinction to make. Acquisition, as you know, for agency is primarily commissions and we've been stable there for some period of time. Don't expect any major changes there. So that's a function really of earned premium. And on the direct side, we will spend, and we've gone through this discussion several times, we will spend what makes most sense relative to the yield that we see. So yes, that could go up faster than our premium under certain circumstances, but that would be based on good business judgment. The non-acquisition expense ratio you should reasonably assume has a great deal of management oversight that says we would need a really good reason to grow any expense faster than earn premium. So we should be able to continue to keep our expense ratio very much in the levels that you're seeing, at least, on the non-acquisition expense ratio. Something like benefits, not quite as controllable and may not necessarily be a direct function of earned premium, but we can control things like salaries, headcounts, certainly premium taxes, those sorts of things. So while you've seen that go down over time, that's been a very deliberate action for us to get to something closer to the run rate expense ratios that you're now seeing on a pretty consistent basis and that the trend being a favorable trend. The biggest change that we will likely be able to bring about in our expense ratio will be continuing to increase policy life expectancy and having more and more of our customers with multiple products. That denominator effect, if you like, or scale effect, is the one that will give us the next quantum. But the active management, which has been going on for some time, to control non-acquisition expense ratio has frankly been extremely positive and got us to a level that we feel very comfortable is sustainable. But for some really exogenous event like benefits costs or something like that which is not likely to move the needle significantly anyway. Brian, anything in that...?

No. I think just a conscious focus internally to try to keep what we call the non-acquisition costs growing at less than earned premium growth. And I don't think we're going to change that view or philosophy any time soon. I think it's going to continue to be a focus. We know that has an influence on the competitiveness of our rates in the marketplace, which ultimately is how we win. So I think we've done a nice job in recent years. And I don't think there's going to be any letup in there. As Glenn mentioned, the advertising and how much you spend on advertising is a function of how much we think we can spend relative to the yield. And on that front, if we see greater and increased opportunities, we'd be more than willing to send more money on the advertising side and that requires us to keep a focus on all those other costs which we are and we'll continue to do that.

On the ad spend, are there any separate trends there in terms of ad spend getting more or less expensive kind of an absolute basis when you occupy the Internet or Internet ad spend with Facebook and Google versus the TV networks. Is there anything at play there?

Well, certainly in terms of our media mix, it certainly is changing in terms of more of the increase in advertising spend, would be more web-based, social media, et cetera, versus maybe TV per se, although TV still is a very, very large component of our media mix. And I would just say that, in general, we continue to see cost per impressions, CPMs, that media outlets charges continually going up a little bit. And I don't see that necessarily going down, but that's more of a function of supply and demand in the marketplace, but currently certainly CPMs are going up -- still going up a little bit.

Okay, great. And lastly as a follow-up, Glenn, to you commentary about, if I understood it correctly, you're offering certain agency customers a discount based on data you've learned from Snapshot, what percentage of the population of the drivers could the discount you're speaking of pertain to? I know you want to be reserved, but just hoping to get a sense of whether that could impact a large or small population of your agency quotes.

Actually, the discount will apply to both direct and agency. I just indicated that I think it'll have a little bit more of a marketable effect in the agency environment. So just to clarify that. I'd call it large because it just won't be the same discount for every customer. So there's that kind of notion I went through earlier of expected value. So there will be a large percentage of customers be able to get some form of discount advanced through the point of sale. And in some cases, it might be a larger discount. In some cases, it's smaller. But the number that will likely get some part of the expected discount advanced will actually be significant. And we see numbers far in excess -- and you've sort of seen this, we're not giving out the exact numbers but our indications are that far in excess of 50%, 60% sort of individuals are deserving of the discount that only Snapshot and driving behavior can identify. So those haven't changed.

Our next question or comment comes from Bob Glasspiegel from Janney Capital.

How are you dealing in a low interest rate world from an investment perspective? If I look at your 10-Q, this may be totally unrelated, but it looks like you sold governments and munis in the quarter and rerisked in corporate RMBSs, in particular AMBSs, which is consistent with the sort of the opposite direction of how they moved in the second quarter, third quarter, so it makes some sense. But is the strategy to rerisk to get more yield, do you think rates are going to stay low for longer than maybe you would articulate it at the last couple of meetings in public? Where are you on the investment portfolio overall?

Sure. Bill, why don't you take that?

Yes, sure. The strategy stays the same as far as trying to protect our capital and focus on total return. And as you've pointed out, during the quarter, we do make moves in the portfolio that are based on value. So we added some ABS, some short ABS which used some cash and some short treasuries for us because they offered us a little bit better yield and a lot of safety. So not high-yielding assets but better than cash and short treasuries. Munis, we sold a little bit because they had runs really far and for us, as a corporate taxpayer in some instances, we're able to sell them at relatively flat yield to the Treasury, so if you can sell a credit product at almost Treasury yields, you should do that. And the corporate market where we sold earlier in the year is that same sort of situation where spreads have tightened pretty sharply to the point where, in some cases, they didn't represent good value to us. We reversed that a little bit in the third quarter as spreads widened particularly in the high-yield market where we're able to add some paper at, we thought were attractive spreads. So unfortunately, we have to operate in the market that we're operating in. What we're trying to do is keep it in the fairway and maintain flexibility. Where we find some opportunities, we'll put some money to work. And where we find some opportunities to sell, because pricing has gotten, in our point of view, way too aggressive, we'll sell. On the duration front, yes, we do think that rates will rise. I think the U.S. economy is doing well. And the Fed just announced yesterday that they will end their 2-week program now and -- or at the end of the month and there's a lot of speculation as to whether or not they raise rates in the third quarter or fourth quarter next year. And that, whether it's one quarter or another is less relevant to us than the overall direction. And our view is that rates will go up as the economy does better and we'll have a better opportunity to spend duration and take some duration risk and get paid a little bit more for doing that.

How long have you had that view that rates are going to go higher, Bill?

It's been a long time. Now it didn't pay off for us last year but, yes, it's been a while.

Now it seems like your tactical moves have been really good particularly in the equity side, but the view, I mean, you're with the consensus that rates are going higher and that hasn't seemed to work, but you're going to stick with that position from here. And the last question is to what extent does your view on interest rate layer into the actuary's view on less cost severity. I mean, is there an integrated strategy on the asset liability side of the balance sheet on what inflation could it be? Or do you operate more independently, which is the case for most insurance companies?

This is Gary, Chief Actuary. We operate more independently. We're truly looking at the medical and the inflation cost would come in. Obviously, everybody knows what interest rates are, but we're really kind of independent of that.

So on the liability side, you could have a view that inflation won't be that bad, isn't consistent with the investment side if possible. I mean, I know you're looking at a more micro level than Bill is, but they do interact to some extent.

Yes, what we're going to look at countrywide, obviously, medical costs, CPI costs, things like that. So that does come into play somewhat, but not necessarily what the interest rates are doing.

[Operator Instructions]

This is Brian. I just want to clarify, at least, maybe some of my earlier statements relating to capital and how much capital is required. I referenced $1 billion of premium growth, that's approximately what our premium growth has been in the last year. So use that as an example as opposed to saying that's actually what our premium growth will always be. I used that as an example, so I just want to clarify that one.

Our next question or comment comes from Arash Soleimani from KBW. Arash Soleimani - Keefe, Bruyette, & Woods, Inc., Research Division: Just had a follow-up question on the Florida PIP comments earlier. Just curious if you have any thoughts on how the outcome of the upcoming governor's election could impact your PIP severity, if at all ongoing forward?

I don't, and I'm sorry. I'd like to be as helpful, but I simply don't. I don't think that's going to change most of what House Bill 119 was all about and how that's being interpreted and reinterpreted through the quote systems. That would be -- that's beyond my paygrade. Arash Soleimani - Keefe, Bruyette, & Woods, Inc., Research Division: Okay, that's fair, just thought I would ask.

I would echo Glenn's comments.

And our last question at this time comes from David Small from JPMorgan.

Yes. Could you just give us some more color about the growth challenges that you're seeing in the agency channel? Maybe just talk about competitive behavior and why you think you've been struggling to grow.

Yes. We do a lot of decomposition and I'll try to give some color here to be responsive to the question without absolute specifics, but we decomp that. We know some part of it is rate, some part of it is actually some underwriting actions that we have taken, and that will stabilize over time and frankly we're okay with that. But that's a small piece of it. There is clearly some changes in receptivity to some of the more aggregators that we talked about last time on the call. That's a piece of it. But let's not suggest that it's not partly and significantly the competitive environment. We've seen some results from competitors. I'm sure you've seen them. But I'd suggest that they are getting some significant pickup in new business in some cases after a period of not doing quite as well. So it's a competitive environment, one that we've played in for lots of years and we will respond. We will respond, however, without breaking our discipline around what we've always said, we'll go as fast as possible at our target margins. Frankly, I sit here. I was comfortable making the statements that I make. This is not what I like. We've got so much of the business working really well. We've got some new things that we are not ripe to talk about yet, but our agency production in new business is probably the one place that prevented this quarter being an even better quarter because it was pretty good. But we're on it, and there's no question. This is, in the trenches, competition in the agency channel and we've got to come at that with our best tools, new products, new Snapshot design, some multiproduct news with our homeowners partner. All of those things are absolutely happening and as I did indicate in my letter, don't -- those are not things to sort of Tuesday next week it will be turned around, but as we go into '16 -- or '15, excuse me, I think we'll be even better positioned.

Just a follow-up. Do you think that your competitive position with any average agency has changed given the moves that your competitors have made just in terms of the where you're priced relative to peers or other metrics that you look at?

Yes, I think -- yes, that would be the conclusion I drew because, frankly, we've changed less and then the very early part of the year we were clucking along at some very different rates. So the only way that really happens is if someone comes and introduces some variable, i.e. a price point, that is lower. And we may or may not be prepared to match that price point, but it's not quite that simple. It's more complex. But yes, I'd say the competitive environment has absolutely tried to change our position, our conversion rate in individual agents.

I'd like to thank everyone for joining the call. We look forward to speaking to you again in 2015. I'll turn it back over to you, Caroline.

That concludes the Progressive Corp.'s Investor Relations Conference Call. An instant replay of the call will be available through Friday, November 14 by calling 1 (888) 562-2794 or can be accessed via the Investor Relations section of Progressive's website for the next year. That concludes today's conference call. Thank you for your participation. You may disconnect at this time.