Matt Downing Glenn M. Renwick - Chief Executive Officer, President, Director and Member of Executive Committee Brian C.

Domeck - Chief Financial Officer and Vice President William M. Cody - Chief Investment Officer

Vinay Misquith - Evercore Partners Inc., Research Division Michael Nannizzi - Goldman Sachs Group Inc., Research Division Adam Klauber - William Blair & Company L.L.C., Research Division J. Paul Newsome - Sandler O'Neill + Partners, L.P., Research Division Josh Stirling - Sanford C.

Bernstein & Co., LLC., Research Division Gregory Locraft - Morgan Stanley, Research Division Brian Meredith - UBS Investment Bank, Research Division James Allyn Engle - John W. Bristol & Co., Inc.

Robert Glasspiegel - Janney Montgomery Scott LLC, Research Division Meyer Shields - Stifel, Nicolaus & Co., Inc., Research Division Ian Gutterman - Adage Capital Management, L.P. Keith F.

Walsh - Citigroup Inc, Research Division Joshua D. Shanker - Deutsche Bank AG, Research Division

Thank you, Wendy, and good morning. Welcome to Progressive's conference call.

Participating on today's call are Glenn Renwick, our CEO; and Brian Domeck, our CFO; and Bill Cody, our Chief Investment Officer. This 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 the actual events and results to differ materially from those discussed during this call.

Additional information concerning those risks and uncertainties is available in our 2010 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.

Wendy, we are now ready to take our first question.

In our own test drive of Snapshot, we had a few issues with foreign-made cars, where we had to get a second device. And the customer service representative that helped us out was able to confirm that there was some communication issues with nondomestic vehicles.

So I was just curious how prevalent of an issue is that really in reality.

I am sorry. I'm not going to be able to give you a great deal of information on that.

If it was a significant issue, I'd know about it. I know there are some contingencies, but nothing that's sort of -- it's what I would call a significant level.

So I can't give you more specifics than that, but I’m happy to follow up on it. But really, given the numbers that we're dealing with, that's a very, very small number.

Okay. That's good news.

And then just for the second, a follow-up. I'm kind of asking you to predict the future here, but should we expect maybe some modest retention slippage in the back half of 2012 from the rate increases on auto?

Or at this point, do you feel that enough of your competitors are taking rate that your own rate increases, even though they're warranted, shouldn't inhibit your competitive positioning at all?

You're on the right track. And frankly, I think we just have to let the numbers do the talking.

And I don't mean to sidestep it in any way, shape or form. The fact is if you take rates come up, conversion likely comes down to some degree.

Retention gets some pressure on it. I think the key thing to remember, however, when we discuss rates -- and hopefully, I've discussed that in a fairly open and forthright way, is rate increases -- it always depends from when you're coming.

So we have a very competitive position in the marketplace. Even a rate increase may not necessarily take us out of that competitive position in certain circumstances.

And we have other tools that allow people to actually manage their rate, even at renewal, if they choose to. So it's a really a tough thing to sort of speculate on.

Clearly, we've worked so hard, as you all know, on retention, and we don't want to lose that. But if you had to put down a bet when you take rates up, generally, you're going to be fighting against conversion and retention.

We hope rate, even though I've expressed disappointment -- and that's the way I am. I'm just -- we have a target, and we don't hit it, we call it that.

We have caught rate change from what I would call sort of lower single-digit severity changes to mid single-digit reasonably quickly, reasonably quickly, not as quickly as I would have liked. And hopefully, those rate changes are not the kind of rate changes that force consumers to say, "What really is going on here?"

And while this may not be able to jog consumers' memories for a long time, it actually had some really inert rate levels for quite some time now. So maybe 2%, 3%, 4% changes, whatever they might see, are something that they are perfectly wanting to tolerate and recognize that's cost of doing business.

Okay. First question is on the base of margin improvement.

Severity trends, I believe, as a company is about 6% right now. Frequency is about flat.

And I believe you're taking rates up maybe -- I mean in the 5% to 6% range. So just curious how or when we should see margin expansion.

Yes. As I tried to point out in the accompanying letter, we have 6-month policies.

And I'd say that's enough to get it across the board right now, six-month policies for almost all of our order business, 12-month for special lines, 12-month for commercial. So the answer has changed a little bit based on what we're talking about.

Your assessments of rate need and rate changes is fairly accurate. We've got, in the last month, about 18 rate changes put through just in July.

We had another several in June even since we spoke to you. But every day matters, and every day, you get that opportunity to renew new customers into that rate level.

But you can take an expected value of 3 months from the level of the day of rate change to get sort of half of your renewal book at the new level. In 6 months, we get it fully at that level.

We obviously are pricing to a future trend date, so we want to get on a trajectory where our rates and our cost of business are in sync. What you should probably look for rather than significant explanations, one advantage we have here, we should get to see the monthly results.

So while we have hit a -- last couple of months, points that none of us are happy with, those actions that we have taken, as I've said, these are not ones that are causing a great distress. I don't like it, but I know how -- we know how to fix it as a company and start looking to see that decline.

When we talk on these calls, it's always in a very aggregate level. Frankly, it's all a state-by-state issue, but the aggregate probably is a fair representation of the overall state.

So I would say start looking for it as the months come through, and monthly reporting should be able to give you some indication of that. And hopefully, we get back to a point where our rate level and future increases are in direct line with our cost as best as we can do it.

Brian, do you have any points that I must have missed there?

No, I think that's fair representation. Obviously, John Sauerland, in June, had mentioned we had taken up our 2% rate change up through June and July's activity, and other rate changes here soon we'll put more closely to the 5% range, with likely a couple percent more to go by the end of the year.

And some of our historical loss cost, particularly severity, were higher than what we had anticipated, so we're playing a little catch-up there, and then some of it will be a function of whether our future loss is constant. We have dialed up some of our expectations as to what future severity changes will be.

Sure. So are you expecting future severity, that's for the next 6 months, also to remain on the 6% range?

Because it seems that with the 5% or 6% rate change and 6% severity, there will be no margin expansion at all. So are you taking rates up slightly higher?

Or do you expect severity to come down in the next 6 months?

Also, if you're doing that mental math, recognize losses are not 100% of the cost, but -- so loss adjustment expense and our expenses shouldn't necessarily go up at the same rate. No, clearly, we price to include our margins.

So when we're doing what's called an indication, we have our target loss ratio, and we'll always take rate both at the time. And while I'm not going to go into it, we have other ways of continuing to take rate.

So my preference is to continue to take small bites, and very seldom do you ever get rate perfectly correct on one adjustment. What is, in my mind, a strength of Progressive is our ability to continue to take small adjustments.

So I don't want anyone to think that this is just a onetime we're taking rate. We have been taking rate for a long time.

We now realize we just need to take more rate, and we've made that adjustment. And we're making a good number of adjustments at one time, but this will not be making adjustment and sit back.

We'll continue to monitor. We'll continue to track every cost that we have.

And while we can't always explain why severities are going up or changing, we certainly observe it. And we have enough data to observe it in a way that we believe is credible.

Some things that we clearly watch -- I mean, I get the question a lot of what's driving the severities. I would tell you that at least, we observe increasing utilization of emergency room facilities.

We see increased utilization of radiological procedures. We actually see increasing unit costing of both of those.

Always hard to say exactly how much they contribute, and almost every area of partial contribution is -- probably has some weight. I think you know the story on the used car marketplace, the used car market prices relative to the physical damage side of the business.

But we expect to stay ahead of trend and get our margins back to exactly where you've always said, 96 or better.

So fair enough. Just one follow-up on the severities.

There has been some discussion around that, and you guys are showing 6% severities. Other than showing rising severities but maybe not as high, do you have a sense for what's sort of different, or what you're seeing?

Or do you think you're -- that you're ahead of the curve versus peers?

I'm not sure that we can really make a claim about ahead of the curve. Every company that is reporting is a very credible company.

They have, clearly, the ability to assess their own data, and so I would assume that everybody, in some sense, is right. It is frustrating not to be able to reconcile.

Something I might just point out to you though, there is a very different dynamic in books, books that have been growing a lot or perhaps not growing for some time. They may show different dynamics relative to the maturity of the book and the percentage of new business.

The one other thing I'd actually say relative to bodily injury severity in particular, our actual paid severity is less than the 6% on the incurred basis. Some of the incurred basis increase in severity also relates to strengthening of BI reserves from previous accident years.

But our paid severity is actually a couple of points less than that 6% increased severity. I think some other companies talk about paid severities, whereas we're reflecting some of our incurred severities, which includes reserve changes.

Just going on the severity topic. So I guess one question is given your ability to segment the data so well and a lack of specificity around the drivers of severity when it inflects, when you saw some change over the year, why not take a bigger bite of the apple just given that you weren't exactly sure what was causing the underlying change?

That's clear. And as I said, the mea culpa, in my own letter, saying we may have -- now we know we were later than we should have been.

Let me try to get to that but also give you a sense of -- I'm talking frequency for a second, and I'll get to your severity issue. 2011 was probably a lower frequency run rate than what we believe is sustainable.

We come into the first quarter of 2012, we had a significant dip down in physical damage frequency. We probably really should be closer to a trend line that reflects more 2010 frequency.

So here we have a little bit of a depressed frequency environment. It's real, but it may not be sort of a good indicator of the future.

We're also starting to observe severities, as I put it in the first quarter letter, certainly going up, not quite what we're seeing now. So there was no need right at that time to overreact.

And overreactions, obviously, we lose some momentum in the marketplace. So the key is to try to time without ever sort of going over.

Now we know that severity probably had gone a little faster on us, a little stronger, and that's what we're reacting to. So fair comment, but it's a lot easier to see after the fact than it is at the time of giving a series of data points.

And picking the next one is tough, and that's what we're going to do. And in this case, it look like we took a couple of points and thought that they were maybe a little lower than they really were.

I understand. I understand hindsight, 20/20, and I'm certainly not questioning that.

But I guess you should -- we don't really know what the drivers are of -- specifically of severity, but can you tell us kind of where is it -- where the problem states were that you kind of identified last year and earlier this year that the change in whatever problem you're identifying was worse than you expected? Or is it a broader picture across a larger subsegment of the book than just the states that you had kind of focused on earlier in the year?

Yes. Bodily injury, a big part of our business, obviously.

I would say that's a relatively widespread broad response. So while it will vary state to state, I think the message that we're giving, that it is increasing severity and a little faster than maybe we thought in the first quarter, is a fairly general and accurate statement for the vast majority of our states.

And we've talked about PIP previously. And clearly, PIP -- most of issues that we referred to when we were talking about PIP are now Florida specific.

Brian, do you have any things you'd add to that?

No. That's a fair representation.

And relative to severity on the physical damage, which quite frankly, is at the upper end of the kind of numbers that we're talking about, a lot of that is being driven by total loss replacement, the pressure on certain model years as we see the demand for vehicles and parts for years where we simply didn't sell a lot of the vehicles 2, 3, 4 years ago. And used car prices are actually quite high, and part availability reflects that same supply and demand.

In some of the more challenging states, particularly Florida, how many renewal periods do you think it will take to get your profitability back to target?

You want to talk about current results with some specificity?

Yes. I would say Florida, our calendar year results are not to our liking.

But Florida is one of the states where we've had a fair amount of unfavorable reserve development on previous asset years. And that really relates to PIP, and we talked a little bit about that in June.

So actually, on a current rate level and accuracy in the marketplace, we actually feel pretty good about our current rate, and what we're writing today are it adequate rate levels. We actually took a modest rate increase in the past couple of months.

But we actually think, on a going-forward basis, our rate accuracy in Florida is positioned. The -- we still have the remaining exposure on the PIP development, and we'll see how that plays out.

But we actually feel pretty comfortable about our current level in Florida. And the most important thing is how is your current accident period results, and we're feeling okay about that.

And for those of you who are constant students of Florida, that's an important statement. And we feel good about our rate level -- our current rate level, and we are planning and believe we can do whatever is necessary to meet the requirement of a PIP adjustment in October.

Okay. How is the profitability in California?

Is that holding? Or has that been deteriorating slightly recently?

I would say it has deteriorated from last year, but it is really more specific to the agency channel as opposed to the direct channel. We feel very comfortable with our rate level and profitability in the direct channel.

The agency channel, on the other hand, is not meeting profit targets. A couple of years ago, we lowered rates a fair amount.

In hindsight, we may have gone a little bit too far. We've taken a small increase.

We need -- we likely need additional rate, and in the short term, we are doing other things that we can to reduce new business writings. But direct channel, which is the bigger channel for us, actually, in California, we feel very good about, and it's more of an issue on the agency channel.

Okay. And one more follow-up.

As far as investment income, it seems like we're going to in a lower interest rate environment for a while. Is there any thought to changing and going out of that longer on your maturities?

Go ahead, Bill.

This is Bill. Not at the moment, not with the curve as flat as it is and rates where they are.

The story is still the same for us, it's where we can find some value in some non-treasury product in the short to intermediate part of the curve, well, we'll add. We're not trying to hit a yield above and manage total return.

And our view at the moment is that the risk, maybe not near term but eventually, if having a longer duration, with rates as low as they are, this leaves us a bit vulnerable. So that's not a risk we're willing to take at the moment.

And just a slight topic change here. Now that you've have the telematic out a little bit, can you judge if you've actually seen new demand from the telematic itself versus just a retention process?

And I'm just trying to get a sense of sort of the sensitivity of whether or not that's actually driven new customers or if it's -- and if we can tell that yet or if it's really just still very much a retention tool.

Paul, are you referring primarily to Snapshot as for current customers or the test drive version of it that we have recently?

The Snapshot.

Snapshot. I think that John sort of covered that to some degree in the Investor Relations meeting.

And retention certainly is the big issue, and we're happy to give you some sense of exactly what we thought our policy -- not only policy life expectancy but lifetime premium estimates were. And also, that we have, yes, I believe attracted customers.

And the way he was able to do that is to give you sort of the best shot where we have submitted controlled experiment to allow people to sort of take it or not take it. However, once you move beyond the controlled experiment, I really don't have a good basis to be able to tell you that Snapshot itself, net of all other market conditions, is doing x or y.

And I'd love to be able to do that. I just don't have a way to do that.

So the best information is still the information we provided at the Investor Relations meeting on the controlled test.

Excellent. My follow-up is to bang a little bit on the severity.

In the PIP, in Florida in particular, the last time we had PIP problem, it was very much a -- as I understood, a fraud issue, and not just for you but for everybody, particularly in New York but also in Florida. Is that the case again this time?

Because sometimes, it's about policy terms. Sometimes it's about issues that have to do with the way the structure of the benefits are given.

If we've been able to kind of home in on what really caused the PIP issue this time around.

This is a different type of issue. Brian, you're somewhat close to it.

Do you want to sort of give a synopsis of what has happened in Florida subsequent to the rulings?

Yes. I won't go in a lot of detail, because some of it involves litigation, and so I'm not going to comment on all that.

But suffice it to say, the unfavorable development, which is driving sort of poor PIP results and including -- affecting our calendar year results, relates to previously closed claims, claims that we have previously closed, where some court rulings would suggest that we did not pay the right amount. Now we believe we did, and we believe we paid relative to our contract.

And it's an issue in terms of the courts, in terms of actual contract wording, which is more specifically what the issue is about. So as a result of that, we have had to reopen some claims from previous accident years, and that's causing the unfavorable development.

I won't comment a lot more, because it involves some litigation, but that is what it is. It's now -- whether the question is, is there fraud, et cetera in Florida and has that changed over time, I think it goes in waves.

But our specific unfavorable development, I wouldn't pinpoint specifically the fraud. This is more a contractual wording and a litigation issue.

Is that statutory contractual wording? Or is it specific to how you wrote the policies?

It is more court decision and court interpretation of our contract wording, not our contract wording.

So it's an issue for everybody as opposed to Progressive?

We believe it's widespread. It's always hard to know exactly what everybody else has in their contract.

This issue gets a little bit more complex. So you see we're sort of hedging a little, not for lack of wanting to disclose, just because there are issues that it doesn't make sense to talk a great deal about right now.

But the real issue to your question, "Hey, what's the driver of severity?" This one's a pretty weird one and unusual.

It's claims that we believe were closed and settled appropriately are now coming back to us and in some cases, getting additional payment on those claims, driving an increase in severity.

So I'd love to talk a little bit more about the sort of the severity roll-forward and pricing actions. And I -- maybe we'd just sort of start with clarifying something.

I think, Brian, you raised this this morning, I think, and it sort of helps illustrate something that I know investors are sort of confused around, which is with the headline 6% severity number, whether that's a reasonable proxy for projecting future severity or whether we ought to be using the paid severity measures or what the implied severity change was say for accident year 2011 that you guys have effectively disclosed. And in line with that, I think the question then is what does -- if you add sort of a go-forward view of severity that's reasonable and not distorted by reserve changes, what does that -- what, on a sort of pure premium basis, does this look like that you guys actually need in terms of rate?

And is it -- because if you simplistically do this, you'd add up 6% times the larger numbers for positive frequency changes. And you'd say that your guys' rate changes are well below the rate need that actually -- this sort of simplistic loss trend that you'd be seeing.

There's a lot there, Josh. I will try to at least take pieces, and Glenn, feel free -- or Al, feel free to chime in.

The best measure for severity, I believe, over time, is your own assessment of the accident year severity change over change -- year-over-year. Unfortunately, at any given point in time, you don't know what that is, because they develop over time.

So that's why we record both incurred, and I want to let people know that the paid is slightly lower. Now paid can be influenced by a lot of things, your closure rate, how you're closing more simple claims earlier or later versus previous years.

So there are pitfalls in all of those. The other thing, suffice it to say, is that we believe severities have increased versus what we had seen, particularly in the previous couple of years, where our severities for bodily injury coverage and some of the other coverages had been fairly modest low-single digits, low-single digits.

And in fact, last year, we have said 2011, at the end of last year, we thought bodily injuries severities are relatively flat. And we found out and as things have developed that we were a little bit short in that, and that's on the unfavorable reserve development.

So we've increased our severity estimate for 2011 accident year about 2% from where we thought at the end of the year. As we project future severities, that's a select.

If we had a crystal ball as to what future severities are, we -- it would be great. We have increased our select for future severities.

I would not go all the way up to the 6% incurred severity that we are recording right now. The other thing I would just say relative to frequency, for certain, we have started to see an increase in frequency.

In bodily injury, we've seen it for the past couple of quarters. Property damage and PIP is more second quarter related.

Collision was 1% in the second quarter. Frequencies now in 2012 are comparable to what they were in 2010 or 2 years ago.

Last year, we had negative decreases in frequency. So it's not as if frequency is way out of control.

They are now closer to what they were 2 years ago, and some of the things last year that may have affected frequency were gas prices and the like. The reality of it is for us and the industry, in the first quarter, collision frequency was down significantly.

Ours was down like 10% in the first quarter. I think NAI data would say something like 5% for the industry.

And that clearly, I believe, is attributable to very mild weather in Northeast Midwest and sort of a onetime event but nothing that would be expected long term. So on the frequency front, I would say our frequencies are comparable to a couple of years ago, so not overly concerned about that.

One that we will continue to watch more is that BI frequency seems to be going up a little bit faster. And then on the severity front, we've increased future selects.

But to project and say, "Hey, we think the number is 6%," I think that would be on the high end. It'll be less than that.

Okay. That's really helpful, really helpful color.

The other question that I'd love to just clarify is relating to sort of pricing actions. It sounds like your -- if I understood you earlier, you said that what had been sort of 5% by fall sounds like maybe it'll be a couple of points more of rate by year end, make sure I understood that correctly.

And then just generally, I'd love to understand how much -- like your recent rate filings, what rate -- and what average rate increase you've been asking for. And the -- and then if you can expand a little bit on if we sort of model out against future trends, how much revenues and sort of an offset to that should we be expecting from, nationally, your monthly transactors, which I think Glenn was referencing?

In terms of average rate change, think of more in the sort of mid single-digit rate changes. That's not -- they're not all there, but think of them as more in that range.

So as it relates to what percentage of the country have we covered to get to this 2% or 3% that we've taken in the last couple of months, most of the rate changes are most mid-single digits, and that is much more our preferred approach. We don't want to get into positions where we have to take plus 10s, because that will definitely adversely affect retention rates and the like.

So our preferred approach is to take consistent amount of rate changes when we need to. So think of the average rate changes in the mid-single digits for most states where we've taken rate changes.

And in terms of the monthly rate factor, it is not a significant amount of the rate change taken. But it is an increasing portion, because I think some of the experience has proven that that is a good methodology of continued take rate over time.

So we've implemented it in a few more states and then in some states, increased that percentage. But in terms of the overall rate change, I would not say that is a big portion of the 5% rate change through current period or 7% at the end of the year.

Your understanding of the price statement, of how we have taken and how much we're planning to take is correct. Only other point I would suggest to those who are somewhat close to it, most of the changes we're taking now, most are base rate changes.

And the reason I point that out is that sometimes you can take rate changes, and if it's factors and so on and so forth, that may or may not flow through in market based on the mix that you attained. The -- perhaps the more definitive way to get rate is to take it at a base rate level.

We're quite comfortable with the design of our product, though most of these changes are base rate changes, and we would expect them to flow through somewhat close to the amount that we're taking.

Just wanted to shift gears and maybe get a quick tutorial on the gain sharing matrix. And I guess, specifically, how does it react to a combined ratio above the target like you're currently producing?

And then also, are you -- are there any tweaks that need to be made? Are you effectively getting the right behavior but the wrong profitability?

So is it over-rewarding top line at the expense of profitability or anything like that?

Well, the master designer is Brian, so I think you can talk a little bit about the tradeoffs. But -- and I think you'll comment sort of on individual month changes and also the aggregate for the year, which while we're not happy with the last several points of trajectory, still is within target.

Yes. As it relates to that first question about how does it react to combined ratios over target.

And keep in mind, and I've mentioned this before, the gain share, Florida, as you see, is actually in aggregate of 5 component pieces. So there's 2 for direct auto and 1 for each of the agency auto, specialized and commercial auto.

And as they're constructed -- they are constructed relative to growth and profit tradeoffs. And all of the matrices have a pretty big penalty function for being over target, and the penalty function gets greater as you get further and further over target.

So being slightly over target, there is a penalty function that's not as steep. But as you get further and further away from target, the penalty function is pretty severe.

So I think those in particular who have the responsibility of setting prices and ensuring rate accuracy are very cognizant of that. And it is one of the drivers of what you saw in June.

The number had come down fairly significantly from I believe it was 1.42 at the end of May and moved down to 1.29 at the end of June, and that combined ratio of 101 won't -- doesn't sit well in any matrix. So that was the primary driver of it going down.

And yes, there are penalty functions of being over target, and they're pretty severe, particularly as you go further and further over target. We do try to reward and incent growth too, because we want to grow, but we only want to grow at or better than target.

And it doesn't sound like there's any change to the matrix or anything as part of what's happening. It is what it is, and the organization will respond.

Yes. We set these matrices, actually, before -- right at the beginning of the year.

We don't change the matrices during the course of the year. And obviously, over time and over the years, we do tweak them and amend them for the next year.

But the matrices that we set at the beginning of the year, they stay enforced for the full year.

Okay. And then, again, as it relates to this, so if I was sitting in, I don't know, a theoretical state that was earning 100 combined, what would prevent me from just shutting off my advertising spend for the rest of the year and bringing my combined immediately down?

What would prevent you is some degree of management oversight, like anything, but we try to entrust our product managers with a great deal of responsibility. Recognize a large portion of our advertising is not done on a local basis.

Now we -- I would actually expect, in some certain cases, that we actually temper our local advertising. That would be an action I would expect of a product manager if they were running at 100.

And you know the story of how we match advertising to yield, so frankly, I would expect that local level. But again, a big chunk of our advertising is done on a national level.

And while we will likely temper that to some degree, we are not so driven to doing action today that will mortgage the company going forward. We're currently operating, back to your gain share issue, the score of 7% at a 95.9 produces the score that you know about.

It's the directory that's not meeting our expectations, and we'll do everything to adjust that but also try to make sure that we're continuing to get to that rate level in a responsible way so that we can continue our growth and push for a very healthy '13, which we may very well enter with a stronger average written premium largely across the board in both channels and in many states. So the answer to -- the short answer to your the question is as a product manager, you could do that, technically.

There would be management oversight. We're not looking to do those kinds of actions.

That's not what we do, and more importantly, a lot of it is done on a national level.

A couple of ones here quickly for you. I think you just ad -- answered my one with respect to ad spend and whether this is going to -- the rate increases will have any kind of impact on ad spend going forward.

And I guess if yield's going to potentially decline, is that kind of where you're going to manage it and look at it?

Yes, exactly. But don't expect -- if I had to give you a general piece of guidance, and you know we don't sort of do this, I think our ad spend will be roughly the same this year as it was last year.

And under different conditions, it might have been up.

Got you. And then looking at some of the BI severity kind of pushing up here, is there any difference potentially between maybe some of the high risk and then more preferred?

Or is it pretty uniform across the whole book? And then also, any difference between what you're seeing in the direct and agency?

I know agency is where most of the reserve increases are coming from.

Yes. Actually, I was just looking at an instant spreadsheet the other day.

And I would tell you it's fairly uniform. Al, you look at even different data, but essentially the same, the vehicle.

All right. The answer to your first question is really quite uniform across what we would now call sort of our customer tiers.

There's a second piece to it.

Just agency versus direct. And...

We're strengthening rates more in agency than in direct. There is no big, big story there, but it does look -- I mean just 2 books of business running.

We try to treat them with separate indications. We're big enough in both channels and in most states to be able to do that.

And for no particular reason other than just the way the data comes out, we're strengthening rates a little more in agency than direct at this time.

I have one question, which is regarding new applications. And why is the growth faster in agency versus direct?

Yes. Actually, it's been good, but unfortunately, my last statement might be a little bit of a telling statement that we need to strengthen rates a little more in agency than direct.

So you should take from that agents are very responsive to us. They like our product, and they like our rates a great deal.

But we're going to have to take that up, and as we've said on many occasions, agents are facilitated by software that give them very good market comparisons from lots of different companies. So while we've been very attractive, we're going to take our rates up.

It's even hard for me to answer the question. I just have to wait and see what that means.

In an individual situation, every agency has different levels of carriers, different numbers of carriers, different combinations of carriers. I would expect our competitiveness to go down just because that's sort of a pretty clear statement.

You take rates up, your competitive goes down. What I can't really know is how to calibrate that even in a constant situation, if no one takes rate changes and secondarily, if I will just take rate changes.

So really, I'd say just track on the numbers there for the rest of the year. We are very, very happy with our year-to-date with our agents and production, and we hope that certainly stays.

I don't think it'll stay at the current level.

Could you give us some color on the new applications on the direct side?

Brian, you want to take -- you follow that fairly closely.

On the direct side, I would say the application count, certainly through the first 6 months of the year, relatively low growth. It's not certainly as pronounced as in agency, and more of it is related to conversion gains as opposed to prospect gains, and generally, the change in rates will likely have some effect on conversion.

So going forward, we'll see how that plays out. But I would say in direct channel, most of the gains that we have had so far this year are more conversion related as opposed to more generation of prospects.

Just one other thought on the agency piece. And it's still a small piece, so I don't want to overplay this.

But we did tell you at the Investor Relations meeting that we were putting a push on getting more of our message and more of our agents to understand what Snapshot might be able to do for them. That work continues.

There are some favorable results from that. We want them to be even more favorable.

And the reason I bring it up is it's a very interesting way to be able to also accommodate a customer's desire for lower rates. In a rate-rising environment, it may actually give people a little bit more incentive to say, "Yes, maybe I'll try this, because this gives me an opportunity to tell you I'm a good driver and get a lower rate."

So hopefully, Snapshot is a nice time to have -- or it's a nice time to have Snapshot into a rate-rising environment.

I wanted to follow up on Greg's question on additional tools to adjust to the elevated combined ratios. Most of the conference call has been focused on rate as your 1, 2 and 3 strategy to deal with it.

With Greg -- you mentioned to Greg, you suggested advertising might be another tool. I was wondering if slowing down Snapshot because of the elevated first year combined ratios of that business might be another strategy.

Or there would be other things that you would highlight as beyond price?

We've got a couple of them. Brian also mentioned indirectly in California, in agency, which is not the only place we would do this, but bill plans.

So we can certainly control sometimes the flow of new business just by the attractiveness of bill plans. It matters in different classes of business a little more than others.

It's something we will do. And within Snapshot, clearly, we have a discount structure that we're very comfortable with.

But if that proved -- it proved to be a little more aggressive, we have an opportunity to adjust the discount structure there. And we're not just talking 1 point, 0.5 point, not move away, in any way, shape or form, from our excitement around Snapshot.

Snapshot is something that we think actually offers consumers a way to really manage their price, and this is a good time to do that. So no, we're not planning on pulling back on Snapshot at all.

But at $1 billion, which is a number that we put out in the Investor Relations meeting, we expect Snapshot to perform. And when we say perform -- sorry for the long explanation here -- it does mean that we understand new business can still meet our target but numerically be above 96.

And if that were the driver of the only reason we're above 96 today, I'd be telling you more about it. That's not the big issue, but we're not afraid of bringing on new business at greater than 96.

It's the blend that we're looking for, and we see no reason to slow down any attempts to incent new people to come to Progressive.

Okay. Just a -- great answer.

Just a follow-up. You're 100% sure that it's -- the problem center around being run on severity as opposed to adverse selection or any other sort of problem with the book.

Yes, 100% is always a great certainty. But I've seen this a few times in my career, as has Brian, as has John Sauerland and so on and so forth.

We -- I'm not that -- you never want to come across cocky, but I think we know how to handle this one. And while we give answers here that are meant to be full disclosure, the detail that goes into our rate revision is far more intense than we can get into in this type of a setting, whether that be some of our commercial segments or whether it be individual states or product designs.

We do have more aggressive work on pre-binding verification in certain states, where we realize there are some opportunities to be presented with individuals who are not buying the product for the intended purpose, so a lengthy way of saying what you might have said. And we think we're on top of all of those things, but those are not new.

Those are things that we do all the time, and we're continuing to refine them. What we see is the instantaneous sort of action.

I would tell you we are very highly confident that we can fix that with the rate changes that, in many cases, we've already taken.

Glenn, in the past, you've talked about how when there's a sort of turmoil in the insurance marketplace, it ultimately benefits Progressive, because you get your rates to where they need to be faster than the competition. Do have that sort of expectation this time around as well?

I'm always an optimist to some degree. You might not get that from my writings at times.

But here's the one thing I've learned in this business is get your rate right first and then attack the marketplace. And when you don't have your rate correct, you own up to it right away and get it correctly -- get it correct.

So yes, I mean in general, you can take in sort of time series, and you'll see that when the market is a little disrupted and rate increases, our disruption to the marketplace. And go back almost close to 5 years, we really haven't seen a rate-rising environment in auto insurance for a while.

So I don't look at this instantaneously. We have a job to do.

We know how to do it, confident we'll get that done. We'll come out of this with higher average premiums, which is something we haven't had for a good long time.

And I think there'll be some market disruption, and we go into that with more tools than we've ever had before, whether it's Name your Price, whether it's Snapshot. And so we actually feel very good about looking forward to '13 and '14.

Okay. And when we look at sort of the monthly reserve development numbers, I guess there's been, consistently, over the past few months.

Does the schedule of reserves reviews imply that that will continue for a few more months? Or do you think you're done now?

Is the question regarding the actuarial reserve changes or the reserve development?

Or both?

It's about prior year development, not the actuarial adjustment.

Prior yield development. Actually, the last couple of months has actually been pretty modest.

So certainly, in the quarter, we had about $23 million of unfavorable development. Most of that was in April.

May and June have been pretty modest. What to expect for the rest of the year, I can't forecast or predict.

But generally speaking, we have generally -- historically seen more development in the first half of the year versus the second half, which just makes sense, because as you get further and further away from the previous years, more of them have settled, et cetera. So -- but I don't want to suggest that we won't see any more unfavorable development.

There may be. But you see it each month in the last 2 months, it's been much more modest than what we sort of saw the first 4 months.

The actuarial changes, actually, from a reserve actuarial perspective, which we look at by segment by segment, state by state and we have a schedule for that, where almost all the large segments are reviewed at least quarterly, we have at least about -- last month, in June, took a more significant increase than we have recently. And we'll continue to do our actuarial reviews and make some adjustments as we see them or need them.

I just wanted to explore the adverse selection concept again. Is there anything -- and no one really calls it adverse selection.

But just as far as you've been growing in the last year to the types of customers you're growing with, have you found anything to suggest that those cells were a little hotter than maybe you expected or just state a different buying pattern than you expected?

Nothing notable, Ian. New business carries with it an elevated frequency, but that's not new.

That's not unexpected. There really isn't anything.

That's not to say we don't get some adverse selection, and we don't get some customers that perhaps are not looking for exactly what we're offering, but that's not the issue. That's not a big issue.

So I don't feel any less comfortable on that than I have for some period of time.

Okay. And is there -- did anything get online as far as a mix shift that may mean you had feel severity more than others.

And the example I thought of, possibly, as you've moved to more preferred drivers by higher limits, you have maybe a higher limit base than you used to.

I don't think that's the big driver. Brian, do you have any sense of what it might be?

Other than in commercial, we're...

You're right that the more preferred have the higher limits. That's right.

I would say that our limit exposure in aggregate, on average, has changed dramatically. I don't believe that's the driver now.

Okay. Great.

And if I can ask one Snapshot question, what happens -- so just this idea that obviously, you have the trial run, and after that, your discount is locked in. How do you account for things like lifestyle changes, right?

Maybe when someone does the trial, they're taking the train tour, because they're only driving on the weekends. And a year from now, they get a new job where they're driving 45 minutes in rush hour traffic each way every day and their profile has changed, but you don't know it because the device isn't in the car anymore.

Yes. There is no question those types of risks are in the design.

Just period, they are. We also reserve the right and will likely use the right to resample.

So even though Snapshot is a discrete measurement, period, we are actively looking for our programs as to when and how to be alerted to remeasuring and resampling. So it is no question that there's a lot of design positives to being able to do the discrete measurement period, lots of positives there.

There are some negatives, and there's no question that you might capture somebody at a point where their behavior is not particularly representative of their ongoing behavior. We are very cognizant of that.

We don't think that that is a large enough percentage to be overly concerned with, but we will be concerned to the extent of resampling. Now obviously, people bring vehicles onto their policy during a point in time, and that's certainly something we will likely resample for.

A couple of quick questions. How much of the 6% projection for severity would be related to what you're seeing in Florida PIP?

That's mostly the BI. You want to talk about Florida?

The severity that we're referring to was -- on the 6% was more on the BI side. PIP severity has increased as well.

I don't have that here in the -- but I guess the PIP severity also is actually above 6% so far through the first 6 months. But PIP, in particular, you have to look at on a state-by-state basis where Florida is actually a higher than average PIP severity, some of that obviously driven by the unfavorable development.

So PIP severity, on an incurred basis, is higher than 6% so far this year.

Okay. And any other states are causing issues besides Florida?

Relative to PIP, I would say Florida is the main driver. Michigan, which is historically been problematic for us, is getting better.

I wouldn't say we're necessarily all the way there, but we're certainly in a much better position there than we have done in a long time.

Okay. And then I guess just for Glenn, you mentioned you were I guess maybe disappointed at the timing it took to recognize the loss cost.

I guess in hindsight, is there anything you could have done or any other information you had that should have maybe caused you to react more quickly, recognizing this is in hindsight, of course?

Yes. We find hundreds of ways to beat ourselves up on things like this.

But on the severity side, that was pretty tricky. I mean, unfortunately, we have to get the data and see it.

And whether we like it or not, we kind of drive looking out the rear-vision mirror. It has to have happened, and that's -- we just don't have a way to sort of take a great index of health care cost and say this will be directly effective.

The one that I think that maybe, in retrospect, we could have, should have, all the other words that would go along with that is that while frequency had been declining almost monotonically for the entire decade, 11 declining didn't look aberrant. But now it looks somewhat aberrant, and maybe there's a bit more of a frequency plateau around the levels of 2010.

And that one, we'll have to see. I mean, next year, we might be talking about it having gone down again, the safety kind of influences in cars.

You know the sort of the things that are happening to the technology of vehicles. They're all inroads.

They're all driving that frequency. So frequency, as a long-term monotonic reduction, is a fairly good bet, but we probably overreacted or trusted the 11 more than we might have and should have.

So just one quick follow-up on Florida PIP. So if I'm just -- simplistically, you don't think you made a mistake growing there?

Or it was just more bad court decisions? Is that gist I should take away?

Yes. No, Florida is a terrific state for us.

Florida PIP -- and I'd share -- and I'm sure I share this sentiment with many others in the business. PIP is just an awfully difficult coverage there.

Sometimes, you don't know if you're coming or going. We paid claims that we believe were paid.

And when I say we believe we paid, I'm not sort of being an insurance company here. I think a reasonable person would believe they were paid fair and accurately.

And now, we're having to reopen them. That's a very, very unusual circumstance and pay more.

So that doesn't reflect on our feelings about Florida in any way, shape or form. We'll get this one behind us.

PIP reform has taken place in Florida. None of us know exactly what that's going to mean.

But at least, we have a little more clarity going forward, though we definitely have a period of ambiguity that we're working through right now. But no, Florida's a great state for us and for the most part, not a problem.

PIP can be problematic.

Well, Keith kind of got my question a little bit. I'm wondering if there's any commentary you could disaggregate Florida severity from the national numbers and see if there's any relative difference that can help us understand what's going on with the numbers.

Or maybe it's just not possible to do that.

Actually, Al, can you do sort of x Florida BI? I'm not sure it's going to be that much different.

Really, it's not. That change in BI country wide has been pretty consistent going up.

And to say a state would vary between let's say 3 and 5 or 6, that's a pretty tight range, and Florida fits right in that range with what we're seeing in BI.