Welcome to The Progressive Corporation's Third Quarter Investor Event. The company will not make detailed comments related to quarterly results in addition to those provided in its quarterly report on Form 10-Q and the letter to shareholders, which have been posted to the company's website. And we'll use this event to respond to questions after a prepared presentation by the company. This event is available via a moderated conference call line and a live webcast with a brief delay. Webcast participants will be able to view the presentation slides live or download them from the webcast site. Participants online can access the slides from the Event page at investors.progressive.com. In the event, we encounter any technical difficulties with the webcast transmission, webcast participants can connect through the conference call line. The dial-in information and passcode are available on the Events page at investors.progressive.com. Acting as moderator for the event will be Julia Hornack. At this time, I will turn the event over to Ms. Hornack.

Thank you, Chris, and good afternoon. Today, we will begin with a presentation on our horizon three strategy by our Chief Strategy Officer, Andrew Quigg. Our presentation will be followed by Q&A with our CEO, Tricia Griffith; and our CFO, John Sauerland. Also joining us by phone for Q&A, will be our Chief Investment Officer, Bill Cody, and the General Manager of Progressive's Home business, Dave Pratt. This event is scheduled to last 90 minutes. As always, discussions in this event may include forward-looking statements. These statements are based on management's current expectations and are subject to many risks and uncertainties that could cause actual events and results to differ materially from those discussed during the event. Additional information concerning those risks and uncertainties is available on our 2018 annual report on Form 10-K, where you will find discussions of the risk factors affecting our businesses, safe harbor statements related to forward-looking statements and other discussions of the challenges we face. These documents can be found via the Investors page of our website, progressive.com. It is now my pleasure to introduce our CEO, Tricia Griffith.

Good afternoon, and welcome to Progressive's Third Quarter webcast. I feel like a broken record, but we continue to be really thrilled with our results. I started off my letter talking about having 15 consecutive quarters of net premium written over time. And so to me, at double-digits, I should say. And to us, that is really incredible. And I start thinking about not just growth in premium because, of course, that's very important, but it can be influenced by trends or a mix of business. I started thinking about what we also talk about that is growth in units. And so I looked back and realized that, we've had nine consecutive quarters of double-digit growth in auto policies for us. So to me, that's equally impressive. We continue to be bullish on where we've come from and where we're going, and that would be a lot of what we're going to talk about today. Andrew is going to focus on horizon 3. He's our new Chief Strategy Officer, and he'll talk about how we're thinking about the future. And then we'll have ample time for Q&A with John Sauerland and myself. So thank you for being in attendance. You're all familiar with this construct. We call it the 3 Horizons based on McKinsey's 3 Horizons. Today, I'm not going to talk about horizon 3 because Andrew will cover that. He'll also cover a little bit of horizon 2. But as you recall in the last webcast, John Barbagallo and Karen Bailo gave a really deep dive to all the exciting things we're doing in horizon 2, including small business, our BOP coverage, TNC, fleet, Smart Haul, just to name a few. We have a plethora of things that we've invested in a couple of years ago, they're really coming…

Thank you, Tricia. Today, I have the pleasure of sharing some additional details around Progressive's growth strategy. For our agenda, I'll first walk through our four strategy pillars and provide more details on the growth strategy framework, which Tricia discussed earlier. In this area, I'll underscore the mission of the strategy group, which I lead. Second, we'll discuss the important balance we are striking between being aggressive with the opportunities available to the company and our long history of using capital prudently. Finally will be our high level framework for evaluating the Horizon 3 areas of opportunity. So let's get going with Progressive's business strategy. Executives, investors and analysts, all discuss company strategy, but it can take on a variety of meetings and approaches depending on the context. Definitions for the word strategy generally fall in two realms. Broadly, strategy can be any plan. Narrowly, strategy can be known as a plan for military action. But neither of these really correspond to strategy in the context we're discussing today. I feel the topic was well examined in a Harvard business review article from 1996 by Michael Porter, appropriately titled, what is strategy? In this article, Professor Porter asserts that differences in performance between companies is the result of the many, many activities that companies undertake. A company can outperform its rivals if it executes similar activities better. This is operational effectiveness. An example of this is segmentation. Every auto insurer segments their pool of risks. A company can also outperform if they choose different activities than rivals. This is strategic positioning, a historical example for Progressive has been Snapshot. Progressive uniquely invested in Telematics starting in 1998 and continues to differentiate ourselves in this space today. Professor Porter also argues that operational effectiveness is necessary, but not sufficient over the…

[Operator Instructions] Before we get started with questions from the conference call line, Tricia, I have received quite a few questions about the change in dividend policy we announced in December of 2018. So why don't you start with a little bit of that history and the reasons for that change.

Absolutely, so the Board terminated the dividend policy that we had for many years in December of last year. Let me give you a little bit of background. So in 2007, we initiated a very formulaic variable dividend policy. And what we -- the three items we looked at, one was a full year after-tax underwriting profit. We multiplied that by our internal gain share factor. And that is a factor we used -- then we looked in growth and profitability across the board. And it is for our employee bonus that it's on a range from 0 to 2.0. And then Board would decide a percentage on that. So it would be the after-tax underwriting profit times the factor, times at that point 20% divided by the number of shares. And that's how we came up with the very formulaic dividend policy. We made a tweak in 2008 and we said that if comprehensive income was less than the after-tax underwriting profit, there would not be a dividend. And that was, although, just kind of a nuance was very important because in 2009, after the financial crisis, we did not pay a dividend. So it was kind of a prudent change to the dividend policy. 2010, the only change we made was we raised the percentage from 20% to 25%. And then in 2011, we raised it to 33.3% and that stayed the same until the end. We have done a couple of extraordinary dividends. In addition, in 2010, '12 and '13, and we gave $1 per share or an extraordinary dividend in top of the variable dividend. So as we started really growing in earnest around 2016 and going to the debt market occasionally because we needed capital because how fast we were growing. We started thinking about,…

Sure. The trigger around comprehensive income being greater than after-tax underwriting income, certainly, it was instrumental in 2009. In 2018, however, it was potentially an issue for us in that investment income was very low. And you may recall, it was a material sell-off in December, underwriting income was robust. Around a 90 combined ratio, growing 20%, but comprehensive income was very close to becoming less than after-tax underwriting income. And so we are in a place where we were potentially not going to pay any variable dividend or potentially pay what we ultimately paid, which was about $1.5 billion. And we think that sort of binary trigger there isn't optimal, either for the company or management or shareholders. So I think that's another benefit of moving away from the previous approach to the variable dividend. Obviously, the dividend is part of a capital management plan. We've been fairly transparent in our capital management philosophy. Obviously, we have capital requirements from a regulatory perspective, 3:1-ish for auto, about half of that for property. On top of that, we can carry a contingent capital layer, both of those, generally speaking, grow as we grow. So we have additional needs for capital. That said, at this point, we are in a very good position from a capital perspective, and we will continue to follow the path we have, as Tricia noted earlier, around returning under-leveraged capital to shareholders in the form of dividends or deploying them in share buybacks. So I guess, in aggregate, as Andrew noted, I think our financial stewardship has been good over the years. We have, as Tricia showed, achieved fairly impressive returns on our equity and it is certainly our aspiration and intent to continue to do so.

Great. So Chris, can you please introduce the first caller from the conference call line?

The first question comes from the line of Yaron Kinar at Goldman Sachs.

I had a couple of questions. First, we did see a moderation in auto severity this quarter. And at the same time, a little less favorable frequency. Could you maybe talk about both of these shifts and what you see as underlying causes for those?

So are you referring to collision, in particular?

I'm referring to -- I think it was even more broad than that. I think it was overall severity and overall frequency.

Yes, the frequency, overall, was a little bit less, more in the 2% range. And severity is a little bit up. We look at it differently in collision versus PD. So collision was -- the severity was less than in prior quarters. And there could be a lot of reasons, but we believe our mix of business is changing from a preferred perspective. But on the severity front, we know that it is costing more to fix cars and the components of the vehicles continue to be more expensive. And so we relate a lot of severity to that in total losses, but the frequency is still in the range. It's a little bit less -- I mean, the frequency, the frequency is a little bit less. And that, of course, is really hard to predict. That can go up and down, but we don't see anything systemic.

I thought that property damage already actually came in, in the quarter.

The property damage -- may you repeat that?

That property damage severity came in this quarter.

Yes.

Yes. I would sort of characterize it. So a point or two, quarter versus year, we don't normally view that as a material trend. It's tough to project as it is. But if it's 6% severity change for the quarter, 7.4%, year-to-date. We consider that to be a similar severity trend. And from what we can see, our severity trends are fairly consistent with those of our competitors. We obviously don't have full transparency there, but we have pretty good insight, and we think severity is in the 6% to 7-ish percent range right now. Frequency, down 2% to 3%. And that's generally speaking, what we are thinking going forward as well.

Okay. That's helpful. And then my second question is around the agency new application conversion rates this quarter. Seem to be a significant decline there. Could you maybe talk about what drove that?

Well, I think there's -- we've talked a little bit about a soft market, and there's less shopping when people's rates aren't changing. And so if you look over the last couple of quarters, there's been either rate decreases or less increases, so less shopping. And so as we've talked about in the past, we continue to advertise, and that affects both our agency and direct to make sure that we grow and the growth is slowing a little bit. We believe our conversion is still solid based on when people do shop, we have a competitive rate. But yes, we definitely have seen a slowing down because there's either been rate decreases or less increases. And I think we'll have to see if that changes are not based on what we're hearing from the industry, as you just talked about, from severity.

Just to add that... Sorry.

I just would have thought that the lower rates would have driven lower new applications but not necessarily decrease the conversion rate of these new applications.

Well, yes. Well, good version is usually based on -- is the rate. When you do shop, is it the right way for you and some people, if it's not significant, it won't convert.

On a year-over-year basis, actually in both the agency and the direct channel, conversion has been up. And it's been sort of a highlight of our product model introductions. And I would also add that its conversions has been improving in segments that we're targeting. So we're targeting more and more preferred customers, especially in the agency channel. We've seen nice shifts up and preferred characteristics of the business we're writing, and that's predominantly driven by the new product models and the conversion that is driven by those models and that we're getting more competitive by introducing factors for the preferred end of the segment that are more competitive.

Yes, I think the only one we haven't -- where we haven't grown new apps on the agency channel with an inconsistently insured, which we call Sams.

The next question comes from the line of Gary Ransom at Dowling & Partners.

Tricia, you mentioned in your letter on Snapshot, you said that your program encourages customers to improve their driving behavior. That's a little bit different from what I've heard over the last several quarters of measuring for pricing, using Snapshot to measure pricing as opposed to changing behavior. Is there something behind that message that you're doing differently, either in your newer distracted driving measurements or in something else?

No. Not necessarily. I mean, we use that, obviously, as our most predictive variable, but we also do want to encourage people to understand when, from a distracted driving perspective, as an example, when they are using their telephone, hands free, not hands free, et cetera, mostly for the safety of every other drivers. So we've always cared about it, and we've tested different things, whether it's beeps, start giving you the information right away. It's still -- it's mainly used as a variable. But obviously, we want to give the best drivers, the biggest discounts. And so both of those things important, it's understanding our rate to risk as well as of helping people to drive better.

Do you have any intention of changing your -- the Snapshot concept of you just measure for 6 months, and you're done and make it continuous?

You know what, we're actually -- we talk about that a lot, and that might make more sense on the mobile device versus the dongle we have right now. So we're always talking about that. And I'm always talking about different ways with which to use that data, whether it be for claims, et cetera. So we're always kind of in discussion around that.

And the next question comes from the line of Amit Kumar with Buckingham Research.

I wanted to go back to Yaron's question on loss cost trends and maybe build on that. If you look at the frequency and severity data by quarter, it has been stable. But in the queue, you talk about additional pricing actions. So all those being equal wire, why not would that translate into additional margin pressure from here as we head into 2020?

Well, I think that depends on each one of our products. Obviously, I talked about property, and we're seeing some high loss cost on property. So of course, that affects trends. So our focus right now is to get into that profitability and expand our products and our segmentation skills and make some underwriting changes. But I think what we will look at is the continued severity trends, understand those where we feel like we're in a really good position as we haven't, in certain areas, where we've thought that small decreases would increase conversion. We've done that. But we haven't done it to the extent that some of our competitors have. So as things change, we believe we'll have to -- our rates would have to change substantially. And we've always talked about that with our customers because we know our customers want to have stable rates. And so when we think about increasing rates substantially or decreasing rates substantially, for that matter, we try to make them stable. So that it will be less sort of a topsy-turvy. But we'll follow the trends. And we will, as we've said in the past, take small bites of the apple. We've always said if we increase rates, we want to be out ahead of it and understand 1 -- 3:1 is better than 1:3, we used to say, to take small, small bites, so customers stay with us because another big part of, obviously, our growth strategy is retention.

Got it. The only other question I have is on commercial auto. In the queue, you talk about the change in loss cost trend. You talked about the shift in the mix of business towards for higher tracking as one piece. Intuitively, as the book pivots and other pieces versus a monoline book grows, would that translate into a higher initial loss pick as we get into 2020?

I wouldn't take 1 quarter or 1 month of commercial to be -- to something -- we feel like it's not a very alarming trend. In fact, if you look at our year-to-date, we're in the high 80s. So we feel very comfortable with that. And in fact, if you look at accident year, it's even better, about 2 points better. So what we do, again, we surgically look at every BMT, every state, new and renewal, and we watch that closely. We take rate when we need it. As evidenced by -- if you remember, back in 2016, Commercial popped up a bit, and we reacted quickly and got back in the track. So we watch new business. We watch our renewal business. We watch our competition. You saw the delta we had in the combined ratio on Commercial Auto in comparison to the industry. So we are on that very aggressively all the time as we see something change because a lot of this could be affected by macroeconomic trends, we are on that. So if we see loss cost on a certain BMT go up, and we believe that trend is going to continue, we will be on that with rates or changes or underwriting restrictions.

I would just clarify that our loss ratios are always bottom-up. They're not top-down. So you mentioned a loss ratio pick going into the next year. That's not how we reserve. So we actually have average reserves for losses that adjustors set below a threshold. And above that, we use the adjustor's estimate for the loss. We also have inflation factors that apply to other losses in the average layer over time so that if we think generally, severity trend is X, we are baking that in over the course of the year. So we then -- when we actually incur claims and add that to our IBNR, that's how we derive our loss ratio. We are seeing more large losses. And as I'm sure you noted, and Tricia alluded to, we did take a reserve change on commercial lines. In September to reflect those trends. But again, we don't go into the year with a loss pick and then develop that, we take a bottom-up approach.

And the next question comes from the line of Mike Zaremski at Crédit Suisse.

This is actually Charlie on for Mike. First question, one of your competitors recently announced initiatives that cut its expense ratio. Do you feel your expense ratio or loss adjustment expense ratio can fall in the coming years based on any specific initiatives that you are undertaking?

We always look to have competitive prices. When you think about -- what Andrew went over, sort of our cornerstones and our pillars, it's really important for us to have competitive prices. So we constantly look for ways to be efficient. And there's a few things that we've talked about in the past. When you think of LAE, we talk about just gain efficiencies through more estimates that -- we can do through photos, making it efficient and customers are happy with that. That works as long as there's no supplements, et cetera, but we've continued to work our way into having a higher percentage of losses settled in this way. We continue to use our technology to get out in front of that. In terms of expense ratio, we talk about that all the time and my team, we're headed into our objectives for 2020. And we're talking about how we think about both expense ratio and LAE and have a goal to reduce that over the years. I talked about that a few years ago where we wanted to get down to a certain percentage. We continually do that. Of course, you don't want to do it, as an example, with claims, there's that balance of having the right amount of people, the right amount of talent because you could reduce your LAE substantially. But if it increases your indemnity, then it's not worth it. So we look at that all the time. It's really about people. It's about making sure that people have the right technology, and we've obviously used some things on the CRM side with AI to make sure that when our customers call in for really simple needs. We don't necessarily have to have a person on the end of the line. That, of course, makes sense, and it's easier for the customer. And helps with expenses. But yes, that's a big part of it. We know price is big. We know that we have to be competitive, and we will continually try to make sure we do the right thing, but we reduced our expenses across the board. Our non-acquisition expense ratio, I should say, on the other side, because we -- in terms of advertising, we'll spend as much as we think we should to make sure that we have efficient use of our media. And if our level costs are too high, then we won't. But we've been really happy about increasing our media spend. So we look at that a little bit differently. We look at non-acquisition expense ratio.

Got it. And then can you update us on the take-up rate for Snapshot? And what percentage of users stay on the program after the initial period? And whether there's been any changes in that rate recently?

Yes. I think the -- well, the take rate on the direct side is much higher. And probably, I think, around 40%?

40%.

40%, it's single-digits on the agency side, we continue to work with our agency workforce to, kind of, to understand that. We would say, at any given time right now, our policies enforced, someone's been on the Snapshot program at some time, about 15% of our policy enforced would say that. So we're really trying to make sure people understand the discount, understand what we're looking for. A lot of people, it's just -- it's that they don't want to do it because of big brother. I think it's sometimes hard to explain from the agency side, partly because we give a participation discount. But when we have the data, we make sure that you can get either a 20% decrease, which is great, or you could get a 20% surcharge. And so I think some people that try it may know that they're not going to get the discount and that won't work out for them. But we think it's -- we think it's, obviously, a great rating variable. We think it's a great for society because hopefully, you do change your habits. But the take rate has been pretty stable, but much higher on the direct side.

I would take the opportunity to add that In the commercial space, we now have an analogous offering we call Smart Haul, and our commercial business is still largely -- by far and away, largely agency based. And the take rate for the commercial business for Smart Haul has been far more promising than it has in the Snapchat side for the Personal Auto. There we are working with our agents to target the four higher transportation segment. So I think interstate trucking, those premiums are fairly substantial. So if you can offer a customer a material discount, and they're an independent operator many times. It's a material consideration for them because that's a business cost. And we're seeing great -- as in the last quarter, I had John Barbagallo share our commercialized precedent, along with Karen Bailo, shared some of the results from that program, which are very promising. Loss ratio is very predictive from the information that comes from the electronic logging devices. And we're providing, as I mentioned, significant discounts as well as, surcharges. But even -- given that, our loss ratio line, if you will, is still fairly steep. So we think there's tremendous continuing potential with our Smart Haul program in Commercial lines.

And the next question comes from the line of Greg Peters at Raymond James.

In your comments, you mentioned the capital requirements for your property business being about half that for auto. And I'm curious if that changes your expectations about underlying combined ratios or combined ratio targets. And in your letter, you also introduced the ASL agreement. And I was wondering if you could provide color on that.

Yes. No, I think that capital requirements are really from a regulatory perspective to make sure that we have the right amount in case of losses. And then I think John referenced briefly, when we talked about the dividend policy, we also have a self-imposed contingency layer in case things happen, should it be a big catastrophe or whatever. So our capital requirements won't change based on that. Our reinsurance does in terms of us understanding the ultimate loss costs. So we went into the ASL a couple of years ago. And basically, we wanted it help us. It doesn't prevent going over 100 clearly, but we wanted to kind of lessen the downside depending on if we have not just name storms because now that's part of the ASL, but it used to be 63 LAE and non-storm or liability. It's really just another protection factor besides our name storms, our 1 in 100 years name storms. So that's something that we've had for a couple of years now. Just more of a downside protection.

And just to elaborate on the combined ratio part of your question there. We target, generally speaking, return on capital across our product lines in the same neighborhood. So you're correct in that if we have to have more capital around for property, we target combined ratios for property that are lower than auto. You might quickly point out that, that is not right now, our performance, and we're very cognizant of that. As Tricia mentioned in opening comments, taking actions to improve profitability in the property space.

My follow-up question. And I'm using the words from your conference call about an environment where there's less shopping. I'm curious if you can tell us how you might adjust your advertising budget, considering that we might be in this new environment where there is less shopping, as you described it, from your consumers?

Yes. So how we look at our media spend is really, we believe, if -- we have an allowable cost per policy. And if our cost per sale is less than that, we'll continue to advertise if we believe it's efficient. We watch, what we call, NP6, which is new prospects that haven't shopped us in the last 6 months, and then we watch conversion. As long as those numbers, we believe, are within the guidelines that we set, then we'll continue to advertise. We want to make sure that we advertise across the board of all of sort of our customer set. So we offer -- advertise a savings message, protection message for auto and home, and we watch those to make sure how good each commercial is, how good each digital advertising are. We have -- we do such deep analytics from our media spend that we will continue to spend as long as we think it's efficient and if not, we'll pull back.

Yes, we think that some of the lower levels of shopping over the previous periods has been driven by the fact that many competitors have been actually lowering rates. We think that environment has changed, and we're starting to see far less rate decreases and more rate increases. And naturally, as customers receive rate increases, they're far more likely to shop than the rate decreases. So it's a dynamic environment. As Tricia mentioned, we manage spend very dynamically, even daily. And when we're seeing changes in the environment, we can react very quickly.

We don't want to -- you don't want to pull back either. Obviously, we still feel really good about our spend. But you don't want to pull back because as the environment changes, you want to consistently be out there because people will have awareness of you and then ultimately, if they shop you, that would be a consideration. So we even see a couple of the trends or the turning point where some -- there is some increases. So we can't predict what's going to happen. But John and I often talk about and look at data and challenge ourself "is the soft market softening," because we have seen a turning point from that based on what we believe are consistent industry severity trends.

I'm actually going to take a question from the webcast now because it kind of builds on a question that Greg just had. The question is, what specific actions is the company taking to achieve underwriting profitability in the Property segment? Should investor -- why should investors, excuse me, be confident that Progressive will get this line to underwriting profitability, when that has been a challenge during the last several years?

We -- let me start with the end one. If you know the history of Progressive, when we set up to do something we execute it. And so I referred to in 2016, when our trends went up and we were a little bit over 96. We got together. We made changes in our expense ratio, and we got rate, et cetera. And we nailed that in really a short period of time. And Dave Pratt, you're on the phone, you can add in. So what we've looked at over the last couple of years, obviously, if you take out cats, we're profitable. The catastrophes, especially wind and hail have been substantial. And so we are basically -- we're changing some of our underwriting guidelines. We're changing some of our coverages, some of our deductibles in specific states, where we see that hail and wind. We're rolling out our 4.0 product. We have it in, I think, three states now, maybe three more by the end of the year, and then we'll have an aggressive rollout in 2020. If we have to pull back a little bit on growth, we will. We obviously have reinsurance as well. And now I think the beauty of having the ASI Progressive is that we have the auto R&D in pricers, the home R&D in pricers, and we can share that data to really understand loss experience across different consumer segments. So for us, we have all hands on deck. I will leave from here today to go down to St. Pete to Progressive Home. We are encouraged about what we're seeing. It takes a little time because some of these are annual policies. But I feel very confident we will get where we need to go. And as a reminder, when we think about growth and profit, we always want to do both. And this has been an amazing several years where we've been able to grow substantially and make a lot of money. If we have to make a choice and we never want to, profit comes first. So we do what we can to make sure we make our target profit margins.

Tricia, I could add just a little detail on the wind and hail comment, if that's appropriate.

Yes, please.

Because you're right, if you look at our losses, the weather losses in the last few years have been higher than we had priced for. The non-weather claims in water, fire, theft, liability, et cetera, we've been right on our pricing targets. And so we've -- when we price for weather because it's so volatile, we can't really use our own experience from the last few years as a good predictor of next year's losses. So we rely pretty heavily on models to do that. The models for hurricane performed pretty well. But what we've found as we've dug into this problem was that most of the tools we had been using for severe convective storms, so thunderstorms that cause wind and hail. We're not performing well at all. And so we've done a lot of work. We've adopted a new tool from Karen Clark and company that we think, with some proprietary adjustments on our part is a much better predictor of future losses. And we're using that to guide our pricing in states with significant wind and hail exposure. So I feel much more confident going forward that we're going to have -- we'll be priced to the risk accurately. And then we'll manage our risk concentrations in areas that get a lot of hail, just to make sure that we don't have an overconcentration in a place where we don't want to.

Yes. And Dave is down in St. Pete right now, and I'll tell you, he's got an extensive career at Progressive doing many of these things. And if there's anybody that can right this ship, it will be him. And really, we have all hands on deck. So my confidence level is very high.

Great. I actually have a follow-up question, sort of related, but it kind of speaks to the questions and the comments, I guess, that Dave had about weather and weather data. But the question is, what do you think the impact to the company will be in the long-term from the risks of climate change?

I mean, that's really hard to say. And clearly, we will continue use of reinsurance to, again, protect that downside. I think what Dave said is right on target with that question as well in terms of us understanding the models better and understanding how they affect our book of business. And so he said, we have some very specific work we've been doing on severe convexity in the last several months, we feel good about it. We'll continue to tweak those things, continue to -- if we have to pull back, have underwriting guidelines, et cetera. It's really hard to say overall, how it will affect us, but we'll do things to understand it better and protect ourselves better.

And the next question comes from the line of Ryan Tunis with Autonomous Research.

This is Crystal Lu in for Ryan Tunis. For our first question, can you please talk about what is driving the increased severity trends in the parts of your commercial auto book that are seeing higher losses? And why those drivers or characteristics are not impacting the healthy portion of your commercial auto book?

I think the primary driver of a lot of the severity is medical costs, and they are affecting the book overall, whether it's auto or commercial auto, it's just more pronounced in the for-hire trucking.

Okay. And can you also talk about which parts of the commercial auto reserves you reviewed in September and which parts of the book are going to be reviewed in fourth quarter? And maybe whether or not you pulled forward any reserve reviews when you found some trends of increasing severity in certain parts of the book?

Yes. I'm going to let Gary Traicoff, he's in the audience. I'll have him come up because he will know specifically what he's reviewing and, ongoing, what we're going to review. But I would tell you, just as he gets -- comes down here, we review a big portion of the book of our entire organization on a very frequent basis to understand that we're reserved. And when we think there's any trend that we see, we really want to strengthen the reserve. And that's why I wrote my letter this quarter to make sure that we knew we were strengthening because we saw those, and we wanted to get out in front of it. Gary, do you want to add?

Sure. Everybody, I'm Gary Traicoff, Chief Actuary. With respect to commercial auto, we review roughly about 90% on the reserves in commercial every quarter. And bodily injury, for example, where we took the $25 million increase, we look at every quarter. So from the perspective of are we changing the ways in which we're looking at the reviews in terms of timing. No. We cover pretty much everything every quarter. The other 10% would be more of our smaller coverages such as collision, property damage, et cetera, that might be reviewed a couple times a year. In general, just to give you a sense as well, we -- across all of our products, we would look at about 80% to 85% of the reserves every quarter. So we are touching them multiple times through the year. And then most of our injury coverages, we will hit those four times a year.

Thanks, Gary.

I'll just add, while we're on that topic. Again, referring back to our last quarterly webcast, John Barbagallo and/or Karen Bailo. I forget exactly which shared a chart with you all around commercial auto prior year reserve development from, I believe, 2012 through 2018 and showed what Progressive had developed and what the rest of the industry has developed. And I will give kudos to Gary Traicoff and team in saying that the Progressive dots were fairly close to the 0 line. Not always on 0, obviously, we're not going to be perfect all the time. But within 1 to 2 points every single year. This year, we're a little over 2 points year-to-date. And the industry was an average that was far greater, and in some years, was even experiencing 9 points of prior year development. That level of reserve accuracy helps the product managers in commercial auto price accurately. We also reserve down to, I think, a finer level than some of our competitors in that space. So again, that drives more accurate prices that makes us more profitable and more competitive in the marketplace. So we have had some development. It's a little higher than we've seen historically. I think the product managers in commercial lines in the appropriate areas are right on top of taking rate, as Tricia was mentioning, it is their job to do. And I think they've got great data to make sure they're doing it in the right place.

And the next question comes from the line of Jeff Schmitt with William Blair.

Could you provide us with an update on your efforts to move into the small business market? Obviously, a pretty big opportunity. But any details you can give on your growth outlook there?

Yes. We see, obviously, the growth outlook we see as extraordinary, but it's really too early to talk about. So our focus is on the 30 million small businesses with 20 or fewer employees. And we just rolled it out in the agency channel. And so we're going to continue to learn. When we have anything to tell we will, but we believe that really fits in our opportunities and we're excited about it, but nothing really to tell yet.

Got it. Okay. And then just a question on catastrophe losses. They've been left out the last three or four monthly reports. And then looking in the queue, I think they're close to $160 million in the third quarter. It's obviously very useful to look at those sort of underlying loss trends with those backed out. But should we expect those to return to the monthly results? Or are they going to be sort of left out going forward?

Well, we've had a long-standing sort of policy that we will put cat information in the monthly reports if it's 2 points or more on the combined ratio. And so that's been decades old. And so we doubt we'll change that. We feel like we're pretty transparent with the monthly reporting and with, obviously, we share a lot during this webcast. So the likelihood of us changing that, I think is low. We'll always consider it, but that's really kind of the -- where we delineate when we share it on a monthly basis versus a quarterly basis.

And the next question comes from the line of Meyer Shields with KBW.

Well, that really swept my question, but I do want to at least throw in a vote for including the information. Because that way we'll be able to have a better read on what's going on. I understand that you're not expecting to change it, but I did want to urge a reconsideration. I wanted to get your comments on the softening of the softening in personal auto. And to see if you can break down whether -- like how much of to be expected Progressive response is to maybe pull back on the modest rate decreases you've been taking or to anticipate improving top line growth -- or policy organic growth, I should say.

Yes. I mean, we always look to improve on top line growth. We took really modest decreases across the board. And it really was to balance out that growth and profitability in terms of getting higher conversion. Just watching the competition and seeing the whole industry, when you combine everything, things have been pretty stable. And in fact, you have seen a couple of our competitors take a few small increases. And that's what we think about when we think about trends and we're watching severities. Like, okay, if this is turning, how are we positioned? And we believe that we're really well positioned because we didn't say, okay, we have this margin, let's take huge decreases to take on -- to take on growth or maybe not take on growth commensurate with those increases. We're very deliberate in understanding, okay, if we take a little bit, how does that convert to balance that great margin, but also that great unit growth. So we feel like -- and the softening of the softening, that's just John and I talking in the hallway because we're like, what do you think? And we'll look at numbers together. We are just going to watch that surgically. And what I would say is, we're in a really good position because we are -- we feel really good about our rate to not have to take huge increases. If we need to take increases across the board, we always will. But again, if we do, and the trend continues to be able to take smaller increases. And of course, I'll have the caveat we always say, that could be different in terms of states and channels and products, et cetera. Do you want to add anything?

[Indiscernible]

Is that good?

Yes. On my end, it is. I had a quick follow-up question, if I can. I was just curious, when we look at the responsibility relationships between auto data and Homeowner data. I was wondering whether the claims that don't necessarily reflect responsibility, is there any relationship there?

We're just looking into that in terms of sharing data with, specifically, only and only Progressive Home with Progressive Auto. But we do see there's overlap. And so there's some interesting just components that we're looking at to better understand how to segment. So yes, there is a relationship. I don't have a lot to talk about with it publicly now, but yes, we're definitely seeing that sharing that data is gathering insights for both auto and home.

And the next question comes from the line of Adam Klauber at William Blair.

Could you give us an idea how the -- how your digital acquisition is expanding? Just any ballpark over the last three to five years, how much of your auto -- personal auto business is coming through the digital channel, whether it goes to your agents internally to sort direct? What was it like 3 years ago compared to today? And then where are you on the homeowners and bundled as far as your digital rollout?

I would say, the acquisition on the digital side -- I can't tell you exactly percentage, has increased substantially over the years compared to when you look at mass media. We're assuming we actually -- you don't know if somebody sees it if they end up calling an agent. So we believe it likely helps the direct side more often because it may be a different kind of customer who's looking online versus going into your agent and kind of wanting to sit knee to knee for the that consultation. But I would say, it's -- I know that it's increased substantially. I don't know the exact percentage. And then with Home Bundle, we continue to change our message. So forever, it was auto. And our assumption was that when you bundled that it extended auto PLE. What we're finding is from -- when we do the ads where we talk about home actually or auto, that helps to measure that bundle. So I would assume that happens again on the digital side. It's hard to say because you have so many ads coming your way, what actually makes that happen. But we have increased on our spend in the digital portion of our media spend.

To that, I would add -- by far the majority of our auto sales now are digital, and that has continued to climb, as Tricia noted. I think the other important trend there to understand is the majority of those are now mobile. So it's amazing to me, frankly, that so many people are willing to buy their insurance on their mobile device and that trend doesn't seem to be abating either. That trend has yet really to transfer to home as much. We're starting, as Tricia noted, to build the capabilities for people to efficiently quote and buy property insurance online. We're seeing some interest in the mobile space for property. But the property -- I mean, sorry, the auto trends are pretty amazing and very digital and very mobile.

And I was talking about, he was -- John added on the acquisition side. And I think I was interpreting your question more as the advertising, which we, of course, equate to acquisition. But again, we believe that what we're spending with Home will eventually go that way as well. So for years, people didn't buy digitally. On our HQX, our HomeQuote Explorer, we now have Progressive Home, the buy button, in 12 states. We'll have 14, we're expected to have 14 by year-end. So just to continue to make it easier, have less questions, all those things. We're going to continue to invest in that because we believe that will go the way of auto and just trying to make it easy, especially if we have data that is publicly known that we can help to guide the coverage for your home.

[Operator Instructions] The next question comes from the line of David Motemaden with Evercore.

Just wanted to get an update on the process improvements in the 5 states where you were addressing some of the BI severity on the auto side. And sort of how those are progressing? And when we should start to see any improvement there?

Yes. So a couple of things. We've done some deep dives. And there's a couple of things that we've tweaked, nothing substantial. We really see that it is inherent in the industry. And in the spirit of transparency, as we've kind of looked and stepped back and we do a lot of audits across the board in our claims organization. I would say, it's likely now more low double digits. And so we see that even more in the five states that we talked about before. In looking at the industry, we believe that it's really just inherent with increased medical costs. We've also seen an increase really over the years in a attorney rep rate. Not necessarily litigation, but having the attorney there at some point in the claim, sometimes at the actual first notice of loss. When I looked back at data around March 2013, we started to see increase in auto countrywide. And I would say, from 2013 to now, there's been about 6 points of increase in ultimate attorney rep percentage. And so for us, we're seeing that. Obviously, when you have an attorney they need to get their fee, so it's going to be more expensive. And that said, we haven't seen a huge increase in general damages, more on the medical side. So we feel like it's more usage. We've seen more injectables, et cetera. So what we're really trying to dig into is, is this a trend? Is there something we can do about it? Because if someone's treating and they're with an attorney, we can't necessarily influence that. We always try to get out in front of the customer to make sure we take care of them and walk them through the process because it is very emotional when you're injured. But some people get attorneys right away, which does make the ultimate cost more expensive. Again, not necessarily litigation, but attorney repped. So we're working on those things all the time. But really, as we have watched the results come in for this quarter from the industry, it seems very consistent with what everybody is seeing in terms of higher medical costs.

Got it. Great. And just switching gears. I noticed in the queue, there is just some discussion of a higher portion of bundled auto policies that have 12-month policy terms. Just wondering how you think about this going forward as you shift your mix more towards the preferred segment? Are we going to see this continue -- the 12-month policies continue to increase? Any sort of difference in profitability in those products versus others? And I guess, just how you view -- it obviously can enhance retention if your mix moves in that direction. But then it also kind of prevents or limits your ability to reprice as quickly based on experience.

Yes. That's true if you were following us back in 1999 or 2000, but we are very cognizant of being able to be nimble when you have 12-month policies. One of the things when we went with Progressive Home in the agency channel is we wanted to make sure, for competitive reasons, too, that we could have yearly policies on the home for the agents, which was consistent with the other preferred companies that they were purchasing. We watch those very closely and make sure that -- especially because sometimes they'll get multi-policy discounts. We watch this closely to make sure that we maintain profitability and flexibility. We do have a higher percentage of those, which we think it is good when you're priced right. But again, we watch it closely, and we'll continue to likely talk more about that.

One fact that I'd point out on that question. As Tricia noted, we had some bad experience back in '99, 2000, where we overweighted our annuals. And we learned from that. And today, we don't price our annual policies, generally speaking, at twice the 6-month. We actually price them higher than that to offset or mitigate the trend risk that you mentioned. So we are, yes, growing. It's still not a large percentage of our policies, and it's just in the agency channel right now where we have annual auto policies, and we think we're managing it pretty well.

And I think, too, I just -- I've been -- I've had a couple sessions with some big agents. And we have been really for years, wanting to make sure that we're seen as the preferred company or 1 of their preferred companies in agency. And with that, for them to get bonuses, their loss ratio has to be within a certain percentage as well. So it's -- it behooves them to put on really good business.

I'm going to finish up with maybe another one or two questions from the webcast. The first one is can you quantify the improvements you've seen so far with Model 8.6 in the five states where it's been rolled out?

Early read is that it's doing very well. Again, that's early, but we're happy with the results.

Okay. And then the last question I have in here is what is the growth prospect for commission income at Progressive? So this is the service revenue line in our income statement. This investor makes the assumption that it's typically a high profit margin and is given a high multiple by investors. So what will Progressive do, if anything, to highlight this expanding revenue stream of the company?

Well, the commissioning that we get from a lot of non-affiliate carriers that we work with -- and I think Andrew talked about the life insurance company we work with, which is great. But we also offset that to make sure that we fund what we call the Progressive Advantage agency. So we use that sort of that contract spends to fund that. So we continue to grow that because we continue to grow. I don't -- we don't have actually any goal with that. It really is about making sure that we have the right amount of carriers for the customer base that comes into our door. And obviously, we want to be competitive with commissions based on the companies that our unaffiliated partners work with.

Yes. So the commission revenue is great. You're right. Service revenues are becoming material. But I hasten to remind that the primary intent there is to keep those auto customers longer and to keep the most profitable auto customers longer. So the commission is great. We are also offering Progressive Home in that same space, but the primary intent is to, as we say, bundle and extend. Meaning, get more of the household into the Progressive family and continue to ensure we meet all their insurance needs for decades to come.

Great. That actually is our last question for today. So I'm going to give it back to Chris for the closing scripts.

That concludes the Progressive Corporation's third quarter investor event. Information about a replay of the event will be available on the Investor Relations section of Progressive's website for the next year. You may now disconnect.