The Progressive Corporation (PGR) Q2 2020 Earnings Report, Transcript and Summary

Welcome to The Progressive Corporation second 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 will use this event to respond to questions. Acting as moderator for the event will be Progressive's Director of Investor Relations, Doug Constantine. At this time, I will turn the event over to Mr. Constantine.

Thank you, Jason, and good morning. Although our quarterly investor relations events typically includes a presentation on a specific portion of our business, we will instead use all of the 60 minutes scheduled for today's event for a question-and-answer session with members of leadership team. Questions can only be asked by telephone dial-in participants. The dial-in instructions may be found at investors.progressive.com/events. 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 today's event. Additional information concerning those risks and uncertainties is available in our 2019 annual report on Form 10-K and our first and second-quarter's quarterly report on Form 10-Q, where you will find discussions of the risk factors affecting our business, safe harbor statements related to forward-looking statements and other discussions of challenges we face. In particular, note that our quarterly report on Form 10-Q for the first-quarter includes discussions of the risks and uncertainties that we face, including specific risk factors arising directly and indirectly from the COVID-19 pandemic and these risks are further referenced in our second-quarter 10-Q. Before going to our first question from the conference call line, our CEO, Tricia Griffith, will make some introductory comments. Tricia?

Thanks, Doug. I really wanted to actually introduce you to everyone. I know this is your first IR call and it's in a weird circumstance due to COVID. But I thought -- I know you've talked to a couple of people over the last month or so, and I've really enjoyed working with you. A little history of this role being really additive for the person in it, as well as Progressive, as you all know, Julia Hornack took her talents to St. Pete for Progressive Home. Patrick Brennan was a prior director of investor relations, who's now our treasurer and Matt Downing is our HR controller. So we pride ourselves on movement around the company. We think this is such a good role for Doug at this time in his career, it really helps both analysts, investors, Progressive and the individual. So Doug, could you tell us a little bit about your career?

Sure. Well, I've had the opportunity to work throughout the organization in my eight years at Progressive. Spent five years in Commercial Lines, where I managed five different states and included [indiscernible] team for two of our trucking BMTs. Also had an opportunity to work in national accounts, where I worked with some of our largest agency relationships. And most recently, I've been a Progressive Personal Lines PM, where I managed two different states. Really looking forward to working with many of you on this call, and I'll do my best to feel the huge shoes that Julia has left behind.

Great. Thanks, Doug. We are thrilled to be working with you. And of course, we do miss Julia. But we're glad to have you here. Jason, we're ready to take the first question.

[Operator instructions] Your first question comes from the line of Elyse Greenspan from Wells Fargo. Your line is open.

Thank you, good morning. My first question, I was hoping to get a little bit of color on how you're seeing frequency and also severity trending in July. Obviously, the trends in the second quarter were pretty favorable. And just as individual start going back to work and perhaps we see a rise in miles driven just due to folks driving for summer vacations, etc. Just wondering if you can give us a sense of how the trends are, and if they started to stabilize in a direction toward pre-COVID level?

Thanks, Elyse. I won't talk about July, but I will talk about when we think about frequency, I'll state that June was lower than the full quarter at about 24%. So we are seeing them stabilize. We still -- it's very different depending on the state as well. So you can see vehicle miles travel go up. And then immediately, if things close down, they go back and forth. So we formed sort of a macroeconomic dashboard. We have a lot of data to look at that, and we react to that. And of course, I'll go into severity because the calendar period severity is very distorted by the drop in new features. So the mix difference compared to last year is very distorted as you see, unlikely with a lot of our competitors. They talked about this last quarter with PD. And remember, we report incurred versus paid. And some of our competitors report paid, as well as PCI. Same story on PD incurred, it's -- the supplements drove PD up about nine points. So when you look at those payments from prior quarters that were we're paying supplements on that frequency, the supplements in the numerator. And so those dollars are increased, where the incurred accounts are the denominator, and so that's where you'll see the difference. Same thing with BI, it's age inflate -- incurred severity about nine points this quarter. Average age of BIs are up about 7%, a little bit higher. So we look at the accident year trends being about 6 to 7%. PIP incurred is really reopens on supplements from PIP. And It accounts for about 25 points of frequency from -- again, from prior periods when frequency was normal. Now the oddest one in this quarter is really our collision incurred [indiscernible], and that is a -- the severity is negative on that. Mainly because when you look at Quarter 2 '20 -- or Quarter 2 '19, the frequency is down about 36%. And then as we thought about our plan because we had excess capacity in claims, we redeployed many different claims individuals. I think I talked last quarter about giving Ohio 100 members of our claims organization to adjudicate unemployment claims. We've also deployed claims people in our CRM organization as people are trying to kind of get their arms around their bills. We also redeployed about 100 people to our subrogation unit, which means that money is coming in. We have kind of full-court press on collecting money when we have liability disputes with our competition. And so that money has come in this quarter because we have 100 extra people actually doing that. And so that's really the -- how the severity trend has gone negative this quarter. So frequency, in June, we saw it abate a little bit to 24%. We'll watch it closely as things go, and we'll just react accordingly.

That's helpful. And then my second question, it seems like players in the space have kind of responded differently in terms of rate changes, right? I mean we've seen rates slowing, but obviously, there's a degree of magnitude between the different players. Are you seeing the difference in rate taking and obviously, part of that -- part of that is impacted, right, by the rebates, which all depends on the different company. But are you seeing the rating environment starting to have an impact on your new business trends? It seems like new business trends started to rise in the second quarter kind of from some of the COVID lows, kind of that we saw at the start of the quarter. So I'm just wondering if the rating environment was having an impact or it is more just folks there just see more shopping going on independent of that?

Yes. I mean we're seeing more shopping, I think, specifically in the direct part of our business. And how we look at rates, when we took the -- when we had the credit, the $1 billion credit in April and May, that was sort of a, hey, we're -- this thing happened very quickly. COVID happened very quickly. We wanted to react. And now we're in the mode, really, we're back to grow as fast as we can at our target profit margins. But we're really trying to leverage, we believe we have, and that's industry-leading segmentation. So we are looking across the country, each product manager, state by state, channel by channel, product-by-product and being much more surgical when we're thinking about our rates. So as an example, in quarter two, we lowered rates and states that made up about half of our auto premium. And if you want to go from the beginning of COVID, take April through August rates that are in play, we'll -- we'll have reduced rates in 35 states that make up more than three quarters of our auto premium. That 35 states, that doesn't mean it's 35 revisions. It could be a few more because maybe we take smaller bites of the apple. I always think about -- think about segmentation and especially increasing rates, unless it's something where we need to react very quickly. I think about my predecessor, Glenn Renwick, who always said, three ones is better than one three. And so we are very, very surgically looking at each state and trying to determine the best rate to continue our growth and make sure that we also have our target profit margin. Does that help, Elyse?

Yes, that is helpful. Thank you for the color.

Your next question comes from the line of Mike Zaremski from Credit Suisse. Your line is open.

Hey, good morning, thanks. First maybe, Tricia, you did mention Michigan auto reform in your letter. Maybe you can kind of give us an early preview of kind of how you see it playing out? Like directionally, do you expect pricing to fall a lot? Are direct writers like Progressive maybe in a better position versus the agency writers as you guys can kind of blanket targeted customers certain discounts, just curious how we should think about that [Inaudible] does move the needle, and it's very profitable for Progressive.

Yes, Thanks, Mike. What I can say is we know it was -- it's a herculean effort to get ready for that event, and we were ready. I will tell you -- and this is, of course -- this is really early because it went into effect July 2nd. I can tell you that we are very pleased with results in terms of growth. Again, we're going to have to watch this closely because individuals are going to choose to have PIP as it was before and not. So it's a little bit early to tell anything. We'll have a lot more data next quarter. What I can say from just looking at the early results, from a growth perspective, we are pleased.

OK. So I'll put that maybe next quarter. Moving maybe to Telematics. We're hearing industry some participants talk about an increase of a take-up rate. And those industry participants are typically on the agency side. Have you seen any increase in uptake in either direct or agent for your Telematics program?

Yes. Mike, we have. We've seen it actually more on the agency side. So our direct side went up a couple of points, but leveled out. And of course, we already have a high percentage of take rate on our direct side -- and direct side. And our agency side, yes, it did tick up. And I think that makes a lot of sense because this is a great time to understand your rating equals one, and it's up about 12% on the agency side. So overall, about 40% on the direct side, about 12% on the agency side and so a little bit more than 20% overall.

And Mike, I would add that we have UBI in our commercial business as well. And take rate there is very strong. And we're seeing great trends around our for-hire segments in terms of their take rate and their own renewal rate as well. So we are broadening our UBI deployment and really pleased with what we're seeing in commercial lines.

And then, could you give us the kind of the recent take-up rate like Tricia gave us on personal or commercial?

We have not provided our take rate on Smart Haul and our commercial business. But I would just tell you, it's exceeded our expectations out of the gate. And the other thing I would mention is that these truckers are required to have the recording devices in their vehicles by federal law. So we don't have the barrier of getting a device into the car or getting an app downloaded. We get the information directly from third-party providers of those devices in the trucks. So it is a pretty good proposition to a trucker who can get a significant discount on a premium, which is a pretty significant premium. So again, I would point to those items without pointing to a number and say, we're pretty pleased with [indiscernible].

And I would say, I think, John, we a few quarters ago talked about -- or maybe as Karen talked about insurance being like the top -- one of the top three costs for those truckers. I think that that is really important. And we've seen trucking increasing based on COVID. I will also add, John talked about Smart Haul. We also have Snapshot Preview, which is available for our business auto and contractors. And although we don't have the data, it gives same upfront savings and fleet management. But we're encouraging -- we're encouraged, I should say, with the demand, and we think that will only increase. And we have that in about 40 states in our agency channel so far.

Your next question comes from the line of David Motemaden from Evercore ISI. Your line is open.

Hi. Good morning. Just a question on just the bundled business, the Robinsons. In the 10-Q, it sounds like you had some pretty good growth in the Robinsons, particularly in the direct channel, but it also looks like you had some good growth in PIP in the agency channel. Just wondering where we're at in terms of the mix, what percentage of your book is now bundled policies? And where do you expect that to get to?

Yes. So we're happy with our Robinsons growth. The agency channel is up a little bit over 18%, and the direct channel is up about 46%. I think our mix there is right around 10%, maybe a little bit lower. Our goal is to continue to have more bundled customers and so the goal is to have as many as we can. We want every auto customer in our book to be able to bundle when they have either a rental or a home, should that be other demographic. So we continue to push hard. You can obviously see it in our advertisement that we talk about protection in home. So our goal is continue to increase our numbers of Robinsons because we know they're stickier and give them reasons to stay based on the products and sources that we're able to provide.

And a little bit on that percentage. So a little higher in direct. We've been at it a little longer in the direct channel, but we've been really pleased with the Robinsons growth we've been able to achieve with Progressive Home, but that's behind where the direct channel is today. But if you blend them together, you come to a number, a little below 10%.

OK. Great. And then just my other question is just a higher-level question on -- just in terms of how you guys are thinking about miles driven and accident frequency. I know there is a lot of uncertainty, but just wondering what you guys are thinking about, will -- do you think that will ever get back to a level of miles driven and accident frequency that we were at pre-COVID and I guess, just how are you thinking you'll adjust given whatever environment you think we'll be in?

Yes, I'll start, David, by the caveat being, we really don't know because of all the different things that go into it. So schools reopening, cases going up, cases going down, vaccines, unemployment, work from home, all those things. So I'll start with that caveat. I will say with our Smart Haul data, we did see signs of congestion kind of flattening more recently. And we're starting to sort of look at data with that. And so we have some initial trends. And again, we're going to have to fine-tune these. So these are -- aren't perfect, but we're continuing to understand the measures of congestion, both on our Telematics on the commercial side, as well as our UBI on the auto side. So I'll give you a quick example. And again, these are going to change, but this is kind of how we look at it. So we've got -- vehicle miles traveled are up, not as up as much as they were this time last year. And obviously, losses haven't followed in particular. So we're trying to figure out the delta between those two. And what we have seen from a congestion perspective is that that does tell some of the delta. So on the UBI side, on the auto side, we believe that that congestion, the gap is about one point during morning rush hour and about nearly two and a half points on afternoon rush hour, which would take into account some of the differences in actually miles driven and accidents because less accidents happen when there's less congestion. Again, we're just digging into this as we have more and more. And the data changes and is very influenced about what states do depending on their rise or fall in cases.

Your next question comes from the line of Meyer Shields from KBW. Your line is open.

Great, thanks, good morning. Tricia, you talked in the past about how Progressive responds really quickly to frequency. And I was wondering whether there are any constraints. I mean, obviously, we're seeing frequency swing around a lot more than ever before. And from an internal perspective, are there stability considerations that would minimize the amount of rate change that you would pursue based on frequency?

Yes. It was a little bit hard to hear, but I think what you were asking was how we would react to frequency from a rate change perspective. I think we take all the data into account. Frequency right now is just so hard -- it's always hard to predict, but it's even more difficult now with all the different inputs with COVID. Again, I'll go back to our stated goal, and we're going to try to grow as fast as we can. We'll look at all the trends and try to understand surgically, by channel, by product, how to continue to put our pedal on that growth mode while making sure we have our profit target margins. We never want to grow and just -- and not also have that profit come with it. So that's an important part. We want to make sure that we have competitive prices and growth. So the predicting frequency is going to be a real challenge for us in the near term just because of all the different inputs that are in a constant state of flux.

Yes, that's helpful. I was just wondering whether there is a limit to how much embedded frequency change you'll include in a rate filing just because of how rapidly it could fluctuate?

I can elaborate a little bit there. So a rate filing, depending upon the size of your state is going to look at data, sometimes going back a year, sometimes going back three years. But you're trying to project the trend going forward because you're trying to price to a point in the future. So you can look backwards, which we do. But ultimately, what we're trying to do is price forward. So it's a little tricky at this juncture given the anomalies we're seeing due to COVID. So it's a great question as to how much of that we actually include on a going-forward assumption basis versus exclude because we think it's a one-off that will not be in plate down the road. So it really depends upon the robustness of the program we're pricing to and I look forward as to what we think we're going to be experiencing when those premiums are in effect, which can be for the next year to two.

Follow-up question on the homeowner side. I know in the past you talked about correcting maybe the pricing for non-cat weather. And obviously, weather has been pretty bad this year, so if you could tell us internally how that progress has -- or how that's been progressing?

Yes. I mean you saw the results. So we're -- we need to make sure that profit is a big part of what we're thinking about in terms of property. And so as you know, we have our new product 4.0 model that we're rolling out. The hail and wind that's happened from a cat perspective has been pretty -- very high. What we're concentrating on right now is to continue to have segmentation. So we continue to roll our 4.0 product out. In addition, we're making changes to the products. And that includes minimum deductibles and ACVs on roofs because that's what we see, especially in hail-ridden states. Underlying ex-CAT, we feel better about our movement in property. We feel pretty good about it, as well as the growth. But again, we want to make sure that we continue to roll out our segmentation. We want to continue to roll out minimum deductibles. And more importantly, make it easy for people to shop. So we have property in 45 states. Now in 17 of those states, you can buy through a tablet or a phone. So we're pretty excited about our investments in technology around that. John, do you want to add anything?

Sure. Yes. Unlike the vehicle programs, especially personal auto, where we've been getting rates more competitive. We're taking rates up in home. So the segmentation is really important. And I think we've got about 18 states out with our next-generation 4.0 product, which is fantastic. We're also increasing rates. We're taking them up almost 6% year-to-date. And we're obviously not hitting our profitability targets. But you also recognize that if you're looking at cat losses year-to-date, we changed from the aggregate stop-loss agreement that we employed last year, which was essentially a cat loss plus LAE ratio to a retention of 375 million of hail losses, effectively hail, wind, a nonnamed storm, and we're yet to hit that retention level. So if you adjust that, and if you assume that we are under the same agreement we had last year to a combined ratio that would be below 100, but still well above our target margins there, so we are taking actions, as Tricia noted, and we're also changing rates.

Great, thank you so much.

Yes. In those 18 states on 4.0, I think 16 of them have the mandates that I referred to in terms of min deductibles in ACV. So we feel good about where we're going. We need to continue to concentrate on that.

Your next question comes from the line of Greg Peters from Raymond James. Your line is open.

Good morning. My first question is around technology. I know you guys have been innovative with technology, but there's also been some new platforms that have hit the market like Root with usage-based insurance and lemonade and renters. Can you talk about your competitive position relative to some of these start-up companies that are gaining a lot of attention in the marketplace?

Yes. I feel like we're in a really nice competitive position for a couple of reasons. And we've had usage-based insurance for literally decades. And they've come in different forms, and we've continued to evolve as technology has evolved. So we have a lot of data to really understand that variable. And we continue to evolve. That's the best part. And it's not something that we sit still on. So we continue to evolve. I will say that start-ups are good competition because they're doing things that the consumers want. And we've always believed in the ability to rate a driver on their individual driving behavior. And so I think competition in that aspect is good. I think also where we have the benefit is in acquisition costs. If you are start-up, those acquisition costs are very expensive. We have a good base of auto customers, for renters in home, etc., and we're a known brand. So I feel like we're in a great position and we continue to add technology advances in our UBI, and I don't think you'll see that abated.

I'd just add a little bit to that. The usage-based model today that we predominantly employ applies discounts in the personal space basically at the first renewal point. We gave a participation discount upfront. Increasingly, we are getting that data upfront. And in Commercial Lines, we're always getting that data upfront and that's really the Root model. You might recall that, I don't know, five years ago or so, we rolled out what we called Snapshot Test Drive, which was a model where you get the driving behavior before you price the policy. We are doing that now with third parties that are -- such as OEMs, such as other app providers. And we are now deploying what we call Snapshot Road Test, which is a model where you download the app, we see your driving behavior, and we employ those discounts at new business, similar to our Commercial Lines model. So I think you'll gradually see migration to employing the driving data where available upfront and we think we're very well positioned for that transition. On the property side, I will point out that we have for a number of years now been the No.1 provider of homeowners quotes and the No.1 provider of renters quotes online. And as Tricia mentioned, brand is a big driver in all of the insurance space, and we think it is a key lever to have acquisition costs in a place that are feasible to make money over the long term.

And my follow-up question around your comments on acquisition costs. I realize you have the Snapshot product, the dongle, but then there's also a lower cost alternative of having the app on the phone. Can you talk to us about how the mix has shifted within your auto business from heavily Snapshot to more blended, where we are in that cycle and where we're going to get to?

Yes. I think as we introduce the mobile device, that has continued to increase. I think it's easy. I think people get it. They can see their information more readily. We still have a fair amount that we're on the Snapshot dongle. I think as we continue to look at the model of senior information upfront, and we're very early on in that, and we haven't really even advertised that, we've seen a really high take rates. I think people -- I think a couple of things. One, we're trusted brand that understands UBI 2, the situation with COVID. I think people are going to -- want to give their data more and more, especially if they're not driving, and they're working from home more often.

Your next question comes from the line of Yaron Kinar from Goldman Sachs. Your line is open.

Thank you, good morning. Maybe a couple more questions on Telematics. First, when you look at the dongle versus the mobile app, do you find that they are equally precise and accurate?

I'll take the last part.

I think you're asking about the accuracy of the data from a dongle versus the mobile app, you're breaking up a little bit. There are differences, for sure. Technology, as you can imagine, continues to evolve in the mobile space and pretty quickly. So there are definitely differences, but we are very confident that we can adequately price folks based on the mobile data. And the data that is coming directly from vehicles is growing as well. And that's obviously as robust as you can get. So we're very comfortable between mobile and dongle in terms of ability to price accurately. And like it was just discussed, it is a lower cost option, and can afford also continuous monitoring, if we so chose to do that as well. So a little different, but not materially. So it's way I would characterize it.

Yes. I would agree with John. I would say when we look at the data we have for dongle also versus the mobile device is a little bit different as well because we're able to understand handheld versus hand free and not just -- in addition to the time of day, miles traveled and hard breaking. So we're also getting a little bit more data, but there's a little bit of a difference, but nothing that we're concerned about.

Got it. And then for the customers who do use UBI to [indiscernible].

We're having a lot of trouble hearing you or understanding. Can you say that one again?

I'll try. I'm asking for the customers who sign up for UBI, what weighting do you assign the UBI data in the grand scheme of pricing those customers?

Sure. Sure. Some commentary there. It's very powerful rating variable. However, today, because not everyone takes the Snapshot option, we solve that last in terms of our algorithms. So we solve for all the other rating variables because all the customers will be rated on those rating variables and then we solve secondarily for Snapshot. So even though we would see it being our most powerful rating variable, one could surmise that if we solve for it first, it would be an even more powerful rating variable. But again, because we don't require everybody to take that option, we solve it last. Does that answer your question?

It does. But can you also maybe offer like -- does it account for like 5%, 50%? How important is it?

That will vary at the customer level. So based on where you live and the other demographics of the household, as you can imagine, for more preferred households, it's going to be different than for more nonstandard households, young, old, etc., and urban and rural as well. So we don't provide an absolute percentage for you.

Got it. Thank you.

[Operator instructions] Your next question comes from the line of Brian Meredith from UBS. Your line is open.

Yes, thanks. I was hoping, could you talk a little bit about the competitive landscape right now in personal auto? And particularly, as I look at your largest competitor on the direct side, they're offering discounts to kind of new and renewal customers instead of these credits. Do you anticipate that having any impact on your ability to kind of grow new business in the near term?

I mean I think everybody took a decision when we went through COVID on what to do and no one company had it perfect because the data was ever changing, it was very new. So I feel like GEICO is a great competitor. They took a different stance than we did in terms of taking the two credits in April and May. I feel very comfortable in our ability to continue to grow. And especially as the end came to the second quarter. So we've increased our advertising in auto about 12% this quarter. So that kind of tells you that we're very much in play for new business. We believe times like this, where there's disruption is really when we win in marketplace. And so we're -- although we hate the fact that this is happening to our country, we believe a lot more people will shop. And our goal is to have very competitive rates and great service once you're with us.

Great. And then my second question is with respect to the Commercial Lines business, you continue to have some adverse reserve development in that area. Is that related to some of the shared economy business? And kind of what is your experience on that business? Is there any changes going on?

Well, the severity and the development is really -- it has been based on similar topics that we talked about before in terms of increased medical cost in the marketplace, higher attorney rep and newer features and then our mix of business going to for-hire trucking, which is a higher severity than business auto and contractors. And when you look at the quarter two reserve development in Commercial Lines, about 44 million really came from about four states, those states being big states. And so when you look at our overall -- year-to-date development overall, you can see Commercial Lines is the biggest part of that 116 million at 98 million. So we continue to watch that from the Commercial Lines perspective. Frequency is down as well. But we -- it's very much like I've talked about in prior quarters with specific states and those variables in terms of medical costs, attorney rep and our mix of business changing.

So it's not specifically related -- it's not necessarily the rideshare-type businesses, it's just generally?

Yes. I mean, T&C, we talk about our T&C, and we took our premium down by 29 million this month. So we're definitely seeing less driving, but in terms of the development piece, that really is about higher BI limits and our adjusters' reserve going up as well. It's a thing we talked about before.

Your next question comes from the line of Ryan Tunis from Autonomous Research. Your line is open.

Hey, thanks. Tricia, I was hoping you could share with us an actual average rate number for the quarter and also through the year so far?

I don't have it with me, but it's been -- it's pretty flat. I think our average written premium. The rate, you mean or rate of increase?

Average premium.

Between 1 and 2%.

Minus 1%. A little over minus 1% [indiscernible] for personal auto.

Yes, personal auto.

We've taken commercial up a little over 2% and property I mentioned was almost 6%.

Yes. Sorry.

Perfect. And then my follow-up is, I guess, looking through the 10-Q, in direct, you've got quotes up, but conversion rates down. I'm curious how you guys are interpreting because it sounds like you guys are being proactive in terms of making your rates more competitive. How are you interpreting your lower conversion rates?

Yes. I mean I think that when we look at -- we also want to add in new apps there. So our new apps in direct are also in conversion, we had some increased ad spend as well. I think conversion is down about 2% with quotes up 6%. It's -- the data is ever-changing because we do believe more people are shopping. But like I said, we spent more in the direct auto. And we feel comfortable with our new apps being up at 4% on the direct side and especially increasing toward the latter part of the quarter.

As Tricia noted, we are taking targeted rate decreases. So we have product managers generally who are managing a state or two, and they are very focused on where they sit competitively or watching conversion in their markets, and they're taking targeted cuts largely at this juncture where we think we should be more competitive and have room to take the rate again on a longer-term basis.

And again, I'll reiterate. For the majority of the states, the 35 states where we're taking targeted rate decreases, they're small bite to the apple to kind of see what happened. Some states needed deeper decreases, but we're really looking at, like John said, the average is between 1.5 to 2% decreases. We're looking at all the data that could be changing. And what we know from the past and what we've always talked about with rate changes overall is our customers want stable rates. And obviously, that can always happen when you need to get rate like we do in the homeowners product, but that's really our goal here to understand the data as it changes and just be lockstep with what's needed to be competitive and grow.

Thanks for the answers.

[Operator instructions] Your next question comes from the line of Michael Phillips from Morgan Stanley. Your line is open.

Thanks, hey Tricia, I'm just curious on your thoughts on given the -- maybe the longer-term implications of the pandemic and stay-at-home and online shopping on how that might change the mix longer term of agency versus direct? And maybe, I guess, what used to be agency customers maybe more willing to become a direct customer?

Yes. It's hard to say, Mike. I think that we're obviously seeing more now and the recovery is faster in the direct business on both Personal Auto and Commercial Auto. I think about it almost like I would have never bought groceries online before because I go to my little suburban town grocery store where I see my friends and it's comfortable. Well, now I'm doing that. I see where people, one, because it is -- a lot of these customers, a lot of these agents, I should say, are small businesses. And so they're reacting to get themselves set up and socially distant, and some people might not be as comfortable coming in. So I do think it's definitely changing now. I think it depends on how long this goes and how comfortably -- how comfortable people are. But it does show, I think, especially on the commercial side, that people are getting -- and then the majority of our business, by the way, in commercial has been through the agency channel. People are more comfortable buying small business insurance, etc., on the direct side. So we have seen it change. I can't commit that it will be a long-term change, but I do think it could be one factor in having people be much more comfortable buying insurance across the board in direct channel.

OK. And then separately on the bundled product topic that's talked about a lot by many companies. Can you say, if you've been approached more by other -- other insurance companies, monoline companies, have you been approached more today than in the past to partner with them to offer home or to offer auto when they don't have auto?

Yes. I think we're approached a lot in terms of that. And in our HomeQuote Explorer and our BusinessQuote Explorer, we work with a lot of different companies, unaffiliated partnerships to sell our home to our auto customers through, not just Progressive Home, but many different carriers and the same thing on the commercial side. So yes, we obviously want -- we want to care about the values of those companies, the brand of those companies to make sure when we partner that we feel good about it. And we always want to expand those, if we think it will be better for our customers, to be able to bundle their auto, even if -- auto and home, even if it's not with us. So yes, I think -- we're approached very regularly.

[Operator instructions] Your next question comes from the line of Yaron Kinar from Goldman Sachs. Your line is open.

Thank you. Just one follow-up. I hope you can hear me better. I think in the June results, you said you were positively surprised by the renewal auto applications and lower policy cancellations. Can you maybe -- now that you have a month of hindsight, can you maybe talk about what happened there? What's led to the surprise?

Well, we're happy with retention. I think, obviously, during the COVID period through May 15th, for the most part, there were moratoriums on any cancellations and we had leniency going into play. And so we knew at some point, our customers on both the commercial side and the auto side would either have a big bill coming due or need our help to really kind of get through what could be a big hump and kind of get on course for their future payments for auto. So we had a process a really very detailed process in both our CRM organization on the commercial side and the auto side to personalize those things what our customers called, knowing that they would have what we call kind of big bills coming up and how we could help them to get through that payment plans, forgiveness, etc. And we were really happy to say that through that plan that we started in May 15th and has wrapped up more recently, on the private passenger auto side, we were able to salvage over 50% of those people that might have canceled. And maybe they would have canceled because they were shopping anyway, maybe they would have canceled because of finances related to COVID, but we were just happy that we were able to personalize that process and have those individuals maintain their coverage with Progressive.

Thank you.

Your next question comes from the line of Mike Zaremski from Crédit Suisse. Your line is open.

Hey, thanks. Can you hear me?

Yes.

Okay great. Just a follow-up question, kind of along the lines of thinking -- starting to think about whether there's any kind of secular trends post COVID. Anything you guys are seeing on the efficiency side, whether it's LAE or expense ratio directly that you think might be done differently, maybe consumer preference, too, that could lead to some kind of benefits for Progressive or just broadly the industry going forward?

Well, I think the efficiency automatically go to claims, which is the big organization, that's really our product once it happens. And it can be -- you got to balance it with accuracy. So we've been testing photo estimates and video estimates from our customers for quite a long time. And this really -- this fast forwarded it because we weren't going to body shops. And so we have a much larger amount, much larger percentage of our auto claims going through, what I would call photo method of inspection. Some of it is from our customers, giving it themselves and they may or may not get it repaired. I would say, that's about 25%. And after COVID, post COVID, it's been about 55% coming from our network body shops. And so of course, you want to balance that efficiency of reps not having the windshield timer going out with accuracy. And so as we're starting to kind of come through the first wave of this, we are seeing some accuracy trends that we want to be able to have our people eventually site of car. So although it's not as efficient because they're going out, we think it's more accurate. So as an example, recently, we -- a lot of our managed repair reps that got into nonnetwork body shops to do the estimates, really want to be able to get out there. When you have a really hard hit incoming -- I was a claims adjuster, so you have a really hard hit, doing it from a photo or video are really tough because underneath that sheet metal, there could be a lot of damage. And you can see a little bit, but you can't really get there. So that causes more supplements, which, of course, is inefficient. So we've just recently talked to our managed repair reps because they've been asking, can we go back out? We have given them all the materials they need to be safe, whether it's mask or gloves and both. And it's completely voluntary. So if you do not feel like you want to go out, that is no problem. We -- our first protection is our employees, but they're going out to some of the nonnetwork shops to do the estimate by the car, and we think that will be a good balance of the efficiency with the accuracy. And of course, we have it set up where the car is outside, no one's with them, etc. So we're really protected. And if they go to a body shop and they see the people aren't wearing mask. We ask them to turn around. So I think we'll learn more about the efficiency of how many estimates post COVID can we do with technology? And we're always testing technology and artificial intelligence to understand that we have so many years of millions and millions of photos. Could we ultimately have very simple estimates actually almost ride themselves and so we've been testing that for a while. Again, I hope I'm answering the question, but it really is a balance of accuracy and efficiency. We're having a lot of learnings from COVID, which is you always want to take advantage of something that's not good to say, what did we learn from that? And how will we come out better and more efficient? And overall, we have goals for LAE and NAER. During this time, there's so much noise in the data because of excess capacity at this juncture and etc. But we care deeply about efficiency and care deeply about our cost structure because we know that in order to have competitive costs, we have to be very competitive on the expense side.

That appears to have been our final question. So that concludes our event. Jason, I'll hand the call back over to you for the closing scripts.

That concludes Progressive Corporation's second-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.