Ladies and gentlemen, thank you for standing by, and welcome to the Allstate Second Quarter 2020 Earnings Conference Call. At this time, all participants are in a listen-only mode. After the speakers’ presentation, there will be a question-and-answer session. [Operator Instructions] As a reminder, today's program is being recorded. I would now like to introduce your host for today's program, Mark Nogal, Director of Investor Relations. Please go ahead, sir.

Thank you, Jonathan. Good morning. Welcome, everyone, to Allstate's Second Quarter 2020 Earnings Conference Call. After prepared remarks, we'll have a question-and-answer session. Yesterday, following the close of the market, we issued our news release and investor supplement, filed our 10-Q and posted today's presentation along with our reinsurance update on our website at allstateinvestors.com. Our management team is here to provide perspective on these results and further context on our strategy to grow personal property-liability market share. As noted on the first slide of the presentation, our discussion will contain non-GAAP measures for which there are reconciliations in the news release and investor supplement and forward-looking statements about Allstate's operations. Allstate's results may differ materially from these statements, so please refer to our 10-K for 2019 and other public documents for information on potential risks. And now I'll turn it over to Tom.

Good morning. Thank you for joining us this day on Allstate. So let's start on slide 2 of its Allstate's strategy and our second quarter highlights. So our purpose is to protect people from life's uncertainties and to be a positive personal good. And as you know, our strategy has two components: increase personal proper liability and market share and expand into other protection businesses, which are shown in those two ovals. Richest Allstate brands, customers and capabilities to both of those components. This strategy is an adaption the coronavirus pandemic led to excellent operating in the second quarter, as shown on the right hand. The enterprise customer experience score increased as employees and agencies did an excellent job of working remote, and we benefited from leading on the shelter-in-place payback and helping customers in other ways. Profitability was good with adjusted net income of $2.46 per share. We also made progress at our multiyear transformative growth plan by leveraging the direct sales capabilities of Esurance and lowering expenses. Allstate Protection Plans, which, as you know, we acquired 3.5 years ago for $1.4 billion and high-growth and high returns, section plans [Technical Difficulty] at the end of 2017 and estimated adjusted net income of $35 million in the quarter and $69 million for the first six months of 2020. The performance-based core losses in the quarter reduced reported net investment income despite solid total loan growth. Our profitable growth it depends overall [Technical Difficulty]. We move to slide 3. I'll say quickly excellent operating results, total revenues of $11.2 billion increased 0.5% of the prior year for gain and growth in the property-liability premium earned were -- that more than offset the decline in investment income for performance-based losses. Net income of $1.2 billion increased 49% to the prior year quarter, as you can see from the middle of the table market value in just $380 million 2019, $2.46 per share, 12.8% of other than repurchase program. With adjusted net income mature line at 17.8%. [Technical Difficulty]

Thanks, Tom. Let's go to slide 4, where we'll discuss the strong performance of our Property-Liability segment. Premium and policy growth continued with excellent recorded and underlying profitability. Policies in force were $33.8 million at the end of the quarter, and Allstate brand policies reaching an all-time high for auto at $20.5 million. Underwriting income of $904 million increased by $537 million compared to the prior year quarter. The chart on the lower left shows the second quarter combined ratio of 89.8% and the impacts driving the 6-point improvement over the prior year quarter. Starting on the left the underlying loss ratio improved by 15.9 points on lower auto insurance losses from fewer accidents, due to significant reduction in miles driven. The underlying loss ratio on homeowners insurance also improved due to increased premiums and lower non-catastrophe losses. The underlying loss ratio improvements were partially offset by Allstate's efforts to help customers during the pandemic. This included the 15% shelter in place payback on auto, which totaled 8.3% of premiums across all lines of on business, plus a 0.5% of premiums from increased bad debt expense, due to billing flexibility related to the Allstate special payment plan. The chart on the right breaks down the expense ratio components. The expense ratio was 23.0 and improved 0.5 point compared to the prior year quarter, excluding the $738 million shelter-in-place payback and bad debt expenses in the quarter. As you know, one portion of our transformative growth plan is to deliver better value to customers, in part by reducing operating expenses. While expenses vary by quarter, as you can see from this quarter, the long-term trend has been down as you look at the adjusted numbers, which are 2.1 points below 2018 expense ration. So let's now move to slide 5 and…

Thanks, Glenn. Let's go to slide 6, which highlights investment performance for the second quarter. We take a proactive and holistic approach to managing the investment portfolio. After reducing public equity in the first quarter, we increased our allocation to investment-grade corporate bonds this quarter. The chart at the left shows net investment income totaled $409 million in the quarter, which was $533 million below prior year, due to a decline in market-based income and losses in the performance-based portfolio. Market-based income, shown in blue, was below the prior year quarter by $77 million. As interest rates have declined, reinvestment rates are below the average interest-bearing portfolio yield, reducing portfolio income. We recorded $211 million loss on our performance-based investments in the second quarter, as shown in gray. As you know, we proactively adjusted valuations in the first quarter in response to the significant decline in equity markets. In the second quarter, we followed our standard process for recording performance-based results, which generally recognizes valuations on a one-quarter lag. Given this lag in income recognition, the second quarter improvement in public equity markets did not have a positive impact on this portfolio in the quarter. GAAP total returns are shown in the table on the right. The second quarter return of 5% primarily reflects tighter credit spreads and the impact of higher equity valuations on the $2.8 billion public equity portfolio. Year-to-date returns were 2.7% and the latest 12 months was 5.9%. Performance-based investments had a 2.4% and 4.5% loss for the quarter and first half of 2020, respectively. These investments are expected to generate higher returns than the market-based portfolio, and consequently, typically have higher volatility. This portfolio has generated an annualized rate of return of 7.4% over the past five years, as shown in the bottom right of…

[Operator Instructions] Our first question comes from the line of Yaron Kinar from Goldman Sachs. Your question please.

Thank you very much. Good morning, everybody. So my first question is just looking at the expense ratio, adjusted for bad debt and the payback improving by 50 basis points, maybe a little more with – when you adjust for the ad spend. How much of that is from the transformative growth program? And how much is just simply COVID-driven?

Yaron?

Yes. Can you hear me?

Yes.

So Tom, I think we're having difficulties hearing you. I was able to hear Mario and the others on the call, but I've not been able to really hear you. I'm not sure if there's somebody on the line – if my line is open. I can't hear.

Your line is open. His line is definitely – the phone is definitely not coming through clearly.

Mario, do you want to jump in on that one?

Why don't we have this? And then let's – Tom, maybe if you could dial back in. Go ahead, Mario.

Oh, you want me – okay. Thanks, Yaron. So, I guess, what I'd say is, as we've talked about transformative growth, I'll just remind you, there's three core elements, as Glenn talked about. There's expanding customer access, enhancing the customer value proposition and continuing to invest in technology and marketing to support the broader transformative growth program. And improving our cost structure is a key part of how we're going to achieve that second objective of enhancing customer value. We continue to make really good progress on this front. Our underwriting expense ratio, as you indicated, is down 0.5 point once you adjust out the coronavirus impacts to 23 in the quarter, and it's down eight-tenths of a point compared to last year. And when you look at where the improvement came from, it's in acquisition-related expenses and operating costs. And that's an area that, as we've talked about, those are two of the areas that we're obviously focused on taking cost out. So we've been at this now for over a year. And I'd say our cost reductions, again, are a core part of transformative growth. So I'd say that's really what we're focused on doing as part of the plan. And we're – we would expect to continue to take costs out over time. And we'll continue to look for ways to improve processes and enhance efficiency and through a variety of means that support the transformative growth program.

I guess, what I was trying to get at is, how much of the improvement this quarter is sticky? And how this quarter is just reduction in T&E and in office costs, given that we were in a shelter-in-place environment?

Yeah. And, I guess, what I'd say, Yaron, is we're going to look to continue to drive cost out of the system. So I don't think – the expense ratio may bounce around a little bit going forward, as we invest in technology. You saw us – we invested more in marketing in the quarter. But I guess the way we're thinking about it is, our intent is to drive the expense ratio down over time, and that's what we're going to do.

Okay. My second question, I think in the past, you've said that you're not looking to cut rates, but we are seeing some of your competitors starting to cut more meaningfully here. Favorable frequency experience probably also helps. Are you holding to your decision?

Yes. So I'll jump in and take that one. This is Glenn, but thanks for the question, Yaron. I think, first of all, we're seeing – there's one competitor that took some meaningful action more broadly. I don't think we're seeing a huge rush in direction, for one. Our competitors have been aggressive. And it's a competitive environment, but not irrational. And the second point I'd make is we manage this on a local level. And I know I'm a broken record, I say this every quarter when we talk about how we manage the – both the margins and our pricing. Over the course of the last 60 days, this could look like, with our flat rate environment, that our product people have just sort of been hanging around and waiting for something. We've actually made 180 filings over the last over the last two months. And what we do is we're constantly, in each state and each market, looking at how we pull and push levers to be competitive in the environment. It's why we saw sequential growth. And we've continued to grow in spite of a tough environment. And those filings were things like, making new business more competitive, making our telematics offering more competitive with some incentives in that area and across all 50 states, pulling different levers like that. So I would say that, while we don't favor, as you said, sort of a broad base like we're just going to cut X amount everywhere. We do favor being very surgical, very detailed and very thoughtful about how we manage our competitive position and our margin in each market. And I think you know from over time that we react quickly, whether it was the frequency spike back in 2015, or it's the frequency decline now of 2020 that we move really quickly on these things, and we have a very responsive system. And we react market-by-market and make sure that we return a very good return to our shareholders. But at the same time, we stay competitive and we're able to grow profitably.

Thank you and best of luck.

Thank you. Our next question comes from the line of Elyse Greenspan from Wells Fargo. Your question please.

Hi. Thank you. Given the strong results that you've been seeing, obviously, pretty favorable frequency results in the quarter, are there any plans for further rebates within your auto insurance book?

So, this is Glenn. Thanks, Elyse. I'll take that as well. I think what we're looking at now, and as I mentioned in the last question to Yaron, we're looking at sustainable and more sophisticated instruments. We're looking at how we get competitive with those as opposed to what I think was the right decision and a good way to go about it with the shelter-in-place payback at the moment. It needed a broad and blunt instrument because of the big move in frequency at that time. But when you wrap-in frequency and then the mix shift that moves severity and the expansion of coverage, the increase of bad debt because we get more flexible for customers in need, put all of those things together, I think the go-forward approach is going to be more in the lines of allowing our losses to flow into our rating as we always do, and to be more precise on a state-by-state and market-by-market basis versus a broad shelter-in-place payback. Things could change because depending on how the frequency and how the virus moves and all of that, but our thinking right now is to be more -- look at more sustainable tools than a shelter-in-place payback going forward.

Okay. That's helpful. And then my second question, could you give us a sense of how frequency and severity trends have trended in July versus what you saw in the second quarter as we kind of -- as folks that started to go back to work and, et cetera, we've seen a change in loss trends?

This is Mario. Glenn, can talk about the second quarter, but we're not going to talk about July trends at this point.

Yeah. So I'll talk about the second quarter a little bit and to say, first of all, thanks for mentioning frequency and severity because they do tend to move in opposite directions with one another. And one thing, I'll just proactively address is that I know some of our competitors report a recorded number on severity. We report a paid number and paid calendar quarter is, I think, a more transparent and better metric to use in normal times, but we're in anything but normal times. So it's much, much more volatile and subject to mix shifts. So you'll see the number move more just because of the mix of what's being paid inside the quarter where you have fewer short-tail, small claims being paid in the period and everything, so that you'll see the two things move a little bit opposite one another. But what we saw throughout the quarter, and we did disclose that June was less of a difference in gross severity -- frequency, excuse me, than the other months in the quarter -- is that it has come down. It's starting to normalize in some places. There's a lot more variation by state early on like what we would see in April and May is that while there was some variation, pretty much all the states were down within a relatively close tolerance to one another. That has started to diverge more with a really rural state like Montana, for example, was actually up slightly in the year-over-year. And then you've got places that are more significantly down, particularly states that are more heavily concentrated with metro areas and that have had more spikes of the virus. So we're watching things on a very local level into how they're moving.

Okay. Thank you. I appreciate the color.

Thank you. Our next question comes from the line of Greg Peters from Raymond James.

Good morning. I'm going to pivot and I'd like to get an update on your homeowners business. I was looking at some of the statistics that you provide, like on page 17 of your supplement. And I was struck by the improvements in frequency, both gross claim and paid claim frequency, not just in the first and second quarters of this year, but it seems to be a longer term trend. And I would be curious about your impressions about what's driving that and how we should think about those trends as we look beyond just this year and think about next year.

Greg, it's Don Civgin. I'll ask Glenn to answer it in more detail. And we've obviously worked hard on our homeowners business over the last number of years, and so we're pleased with the returns we're getting from that business. But Glenn, do you want to get into more specifics?

Yes. Yes, absolutely. Thanks, Don. So yes, as Don said, I'm always -- I always preach that -- look at the long term and look at the reported combined on home. We've got year-to-date and 88.6% combined ratio recorded. Our 12 months is 83%, and our 5-year is 87%. So we've sustained that. I think, Greg, your point on the frequency is a tough one because it changes almost every quarter because the weather drives so much of it. So you'll have more caught up in catastrophes, and so it's sort of a lower underlying frequency, and then you'll have maybe a good weather quarter and you have few weather-driven and fewer cats. They move around a lot. What I would say recently, and this is probably an underreported part of the COVID impact, is that we've seen a shift in frequency severity on the home side because -- and it's pretty logical is we're seeing fewer like death claims and break-ins because everybody's home. And those tend to be lower severity claims, and yet we're seeing about the same number of fire points, and those tend to be higher severity claims. So we saw our frequency go down in the quarter. Our severity average go up because the ones that went away were smaller claims. And so we get into the details, and we look at all of those components, but I wouldn't necessarily draw a conclusion that there's something in home driving long-term frequency down that will be sustainable. I think it's been different factors each quarter.

Got it. Thanks for that answer. And then I guess I'll circle back your transformative growth plan and the integrated services platform. And as you guys know, I always like to track your agency data that you provide on page 9 of your supplement. And so I was wondering if you could just, from a big picture perspective, we're watching some directional changes in some of the numbers, whether it's a slight decrease in total Allstate agencies and LSPs, but we're also seeing an increase in the independent agencies and Encompass independent agencies, can you walk us through what your view is on how that's going to change further as we continue to roll out these plans that you've mentioned previously?

Yes. Glenn, why don't you go ahead?

Okay, great. So you're always into the detail, Greg. So you definitely picked up on the right trends there. We're focused on growing with agencies looking to invest, and this ties back to something Mario talked about the expenses. It's not that our acquisition costs have gone down because we don't want to grow and we don't want to invest in agents looking to grow. They've gone down because we've moved the money really in a way that incentivizes growth and it's more efficient to do it that way. So coming into this year, we moved some money that was incentivizing retention, and we moved it to incentivize new business. And -- or I should say renewal versus new business as opposed to retention because how it gets paid. That's one change. And that changed some agents, who wanted to invest in that way, and some that didn't and perhaps voted themselves out on that one. We also increased our baseline performance requirements for, I'll call it, our lowest producing agency and required that folks sort of be more open for business than they have been. And so that has driven the number down a little bit. And the other thing is integrated service, which you mentioned in the question. That number will start to look strange over time. We have about 500 agents in there right now. So those agents won't show staff members that are being augmented or provided by Allstate. And so it kind of moves the number to a different place because we don't report it in as part of the agency staff as licensed sales professionals because we're doing the work on the back end for them in those centers. So there's a number of pieces moving it. But I think the headline to take away is that we've grown in the agency plan. I mentioned in the prepared remarks that we've all-time high in Allstate brand auto policies in force. So the agency force has grown, and we want to grow in all of our lines. We want to grow all state agents. We want to grow our independent agent channel, which is the pending acquisition that we've got with National general, and we want to grow in direct, which is the combination of Esurance and Allstate Direct.

Hey, Greg, this is Tom Wilson. I'm – hopefully, I've reconnected.

We can hear you.

All right. Let me highlight what Glenn just said. So, transformative growth, so we're going to sell as much as we can through everybody as we can. So we want to have more Allstate agents. We want to have independent agents. We want to sell more. You will see, over time, some increases decreases and how much goes through each of those channels as we transition to or a profit model. So – and Glenn explained what we're doing on new business versus a retention. But we're really working on different ways we can get to customers. We may not need as much real estate in the future as we have today. That means local offices are different. So you'll see changes, and you're seeing some of that change now because we're not new Allstate agents at this point as we build out these new models. So it's really about very distribution force.

Got it. Thank you for the answers.

Thank you. Our next question comes from the line of Mike Zaremski from Credit Suisse. Your question please.

Hey, good morning. Thanks. Could you guys – and maybe I missed it, did you guys touch on auto severity? I think, we – investors have expected it to increase during COVID. Maybe you can discuss whether we should expect the severity or you could talk about it, so we can maybe better understand whether it should continue? And I'm curious whether, like, the underlying severity trend is still kind of higher for longer due to issues we talked about pre-COVID? Or is there a change to that potentially post-COVID?

Mike, let me have Glenn answer that, but give you an overview first. In a time like this, where the numbers are changing rapidly in terms of the number of claims, they – the statistics in the method you use for any individual component of loss costs bounce around a lot. So it's – the best thing to do is to look at overall loss costs and say, are your loss costs still the right percentage of your premiums? We obviously have done attribution on severity, and Glenn can talk about that.

Yeah. So it's a great question because, as Tom said, I mean, they do bounce around a lot. Overall, we feel very comfortable and confident with where we are on severity. PD severity, paid severity was reported at 20% up, which number jumps off the page. Let me give you just an example of why that is on a paid severity number like that. So paid severity is – as you'd expect, it's a number of claims you settled and closed during that calendar period divided by the total number of claims, and you get your average. Well, during the second quarter, we settled just as many six-month-old claims as we would in any normal quarter, because six months earlier, there was no COVID and there was no drop in frequency. We settled just as many five-month-old claims and four-month old-claims and three-month-old claims, but we settled a lot fewer three-day-old claims. And one-week-old claims, because they just weren't there. They weren't happening with the same frequency and as you'd expect, the claims you settle in three, four and five days are lower severity on average. They're quick. They're easy, and they're low cost. The ones that are six months are typically ones that have like they've gone through their own carrier. The segregation comes over from that carrier, they average a much higher severity type of claim, or they've been a total loss, and so they take longer to process and so on. So it's just a pure mix issue that drives that type of number. When we get down underneath that, and we're able to look at the things that come in, in the quarter and how they're resolving and are estimating practices and what the true severity looks like, it feels like normal inflation that we've seen in the single digits that we're very comfortable with. And the claims team has done and the finance team a terrific job of getting underneath the data so that we understand where and if we have any pressures, but we've seen it run very predictably.

Okay. That's very helpful. And lastly, my follow-up. Services segment, it feels like, driven by Allstate Protection continues to do, I think, better than expected, and it's actually kind of moving the needle a little bit these days. You talked about diversifying into – outside of the U.S. I mean, just since it's growing so fast, should we expect the rate of growth to temper? Any additional color there would be around what's going on would be helpful.

Let me provide just overview like why we talked about it, and then Don can talk about its future. So, when we bought it, one of the things we said was, we'll come back and talk to you about how we're doing every so often. And that's what we wanted to do here. And it's meeting and exceeding our expectations. I think there are a number of people. We're trying to understand why we were into that space, and you can see it's a huge growth. And we've had huge growth both domestically, and we're starting to get some good growth internationally. So, it was really about coming back to you and saying, we told you we'd come back, and here we are. And it happens to be a good news story, which Don can talk about.

Yes. Thank you. So, first, thanks for noticing. The service businesses, I think, in total, have made a lot of progress and has improved the values that they provide to customers. And so I think across the Board, they're doing really good work, more specifically around the Allstate Protection Plans. I mean, Mario laid out the three goals that we set and how we're doing against them. But I'd remind you, we set those goals the first quarter after we did the acquisition, and we communicated those, and we've been consistent about them. And we have had a really strong trajectory since we acquired SquareTrade, both on the topline and improvements in profitability all along the way. This particular quarter was good, but I'd remind you we expected to have a good 2020 to begin with. And that's because we were continuing to pick up new customers. We were doing a really good job working with our partners to make the product more available and accessible to their customers. And so we expect the growth not only for new business-to-business customers, but also more availability in existing customers and our international business continues to grow substantially. So, we expected 2020 to be good. We have benefited from the virus in some ways, and topline is one of them. When you look at the partners that we do our largest business with, they tend to be the larger one-stop shops. Which, as you would know, have been in favor in the retail world -- retail is doing mix these days. But if you're a large one-stop shop, we've done quite well. In the categories that we are particularly in consumer electronics and so forth and home office, have also done very well. So, we benefited because of the virus with the topline. And then, obviously, the topline has a benefit to the bottom-line with respect to scale of the expenses and so forth. So, we're very happy with the results. The results would have been strong even without the virus. They have clearly been helped to buy the virus, both on the topline and the bottom-line. But we're still very happy with the underlying performance as well. The only thing I'd add is, it's hard to tell what the impact of the virus is going to be going forward. And that's just going to depend on consumer behavior, where they buy, what they buy. So, while we benefited from impacts from the virus at this point, it's hard to tell what's going to happen in the future. But all that said, the underlying business has been and continues to be quite strong and we're quite bullish on it.

Thank you.

Thank you. Our next question comes from the line of Paul Newsome from Piper Sandler.

Good morning. I wanted to ask about bad debt trends within the quarter itself. And I guess I'm wondering whether or not we could see more when the Federal subsidies for unemployment go away or if maybe the bad debt is more tied to some of the state level moratoriums on -- that were in place. Just your thoughts on that would be great.

Mario, do you want to take that?

Sure. Paul, this is Mario. So, just to give you a little context of the driver of bad debt, as we mentioned earlier, one of the ways that we provided additional benefits to our customers during the pandemic was to offer them the opportunity to opt into a special payment plan, which essentially gives them 60 days of coverage without having to make a payment on the policy. That started in March. And we've – as we've worked our way through the quarter, we've gotten more and more experience relative to how that subset of customers has performed and then the bad debt activity within that customer segment, because as much as we had historical context around offering similar-type programs during catastrophes, this was obviously much more widespread and different. And you see the impact in the quarter where bad debt was $44 million. We'll continue to get more and more experience on that customer set going forward. We'll continue to update our analysis, and then we'll update the numbers in the third quarter, if we need to. But based on the experience we've seen so far through the end of Q2, that's what we recorded in the P&L.

And Paul, I would wrap that into just the overall pandemic impact, right? So we've obviously had lower frequency, which we've talked a lot about. This is another cost associated with the pandemic. And when we're thinking about what we're going to do in the future, we factor all this stuff in.

Makes sense. Second question, I wanted to ask about whether or not you think the regulators will take into account sort of the, hopefully, one-time nature of the pandemic? Or if we put through – when you put through rate filings in the future, will you kind of have this middle of the Python kind of situation where the unusual low frequency is embedded in what you can file for rates prospectively until that kind of rolls off through the system.

Well, it's always hard for us, of course, to speak for somebody else as to what they will do. And of course, it varies by state, which oftentimes relates to the political environment in that state. What I can say is, we were out early, no regulator forces to do shelter-in-place payback or expand coverage and all the other things we did. And so we got out ahead of it. I think we got positive feedback from people that we were doing what was not required, like no one had to come in and say, you must do this. And so I feel like we're on a good basis to work through like what's the right price. And if frequency were to stay at these abnormal lows, obviously, regulators, customers in the competitive market wouldn't have you charge the same amount. The good news is our loss cost would be a lot better. So our focus, really, Paul, is on maintaining our overall competitive position and our margins. And we've been able to do that with regulators. I think there'll be a whole bunch of stories. But the good news is we have good math, an operating model that adjusts locally in a way in which we can continue then grow profitably. Glenn, anything you would add to that?

No. I think that covered it well.

Thanks for your help. Everyone, stay safe, guys.

Our next question comes from the line of David Motemaden from Evercore. Your question please.

Hi, good morning. Just a question for Tom and Glenn. Just thinking bigger picture, I'm just wondering how you guys are thinking about miles driven and accident frequency over the longer term. And I know there's a lot of uncertainty, but are you guys thinking -- how are you guys thinking about miles driven and accident frequency? Do you think that we'll ever get back to levels that we were at pre-COVID, especially given what seems like increasingly prevalent adoption of remote working arrangements? And relatedly, how are you guys going to adjust to potential lower frequency levels on a more sustained basis?

Let me start, Glenn and then jump in. I'm going to start, David, the most macro which is of course, our strategy includes a couple of different components, the top over Property-Liability and then the bottom of other stuff. In the Property-Liability stuff, there's auto, big driver of that. And as -- and that's a key part of our both revenue and profitability. But the other thing is, we have a really profitable homeowners business, much more profitable than other people. We're good at it. We're precise at it. We sell a bunch of other stuff. And then the circle protection selling things like the service plans and everything is about making sure the company has multiple things we can sell to customers to protect them from whatever goes wrong in their life and leverage those customer bases. So we're dealing with -- because you had the same question that was about the same question about the future of auto was really around autonomous vehicles 3 or 4 years ago. And so we've been on that path from a strategic standpoint as to how to deal with that. And, of course, what's happened is, actually, premiums have gone up, not down in the last 4 or 5 years, which people were afraid nobody was going to drive. So it's really hard to predict. What you can do is have multiple options and take advantage of those options and leverage your skills and capabilities. As it relates to the auto insurance business, you want to have a broad-based approach. And that's what transformative growth really is. It's improve the customer value proposes, give people lots of access, connect to more, make your products more simple. And so we still have a lot of the share of that market we can pick up. And so even if you were to say fewer miles driven, lower average premium, we still think we can grow that business. Glenn, you might want to have some -- that was obvious sort of macro and longer term. How would you react to the question on more really as it relates to personal Profit-Liability in the shorter term?

Yes. And it's a great question because, obviously, we think a lot about this. Let me take you into how the team works on this. And I'll go -- like Tom did, I'll go before even the virus. We had built out a model and assumptions for the change in frequency driven purely by the change in the car fleet out in the market. What does each component of ADAS, safety equipment on cars, what does each component do to change each type of loss, like how many fewer sideswipes or rear ends or intersection actions are we going to have based on the blind spot warning or based on this autonomous feature or so on? And you add those up and you apply them to your fleet and how the fleet is changing over time, and we actually -- we've built into the model, our assumptions were what changes. But at the same time, we built the other side of it, which is the severity. Because all those cars getting into fewer accidents are a heck of a lot more expensive to fix when they have all that equipment on it. And so it goes to what Tom said about, really, it's hard to predict, but we go after it and we look at with as much specificity as we can, the component parts. So now take that into the current scenario and the question you're asking. And we're looking at, so what percentage of the market works in jobs that don't have to commute? If you work in a restaurant or your dental hygienist or like -- but you're going to have to commute because those things cannot be done remotely. But there are a lot of jobs, most of us have jobs that can be done more remotely than they have been historically. So you get that percentage. And then you look at, well, what percentage of people will do it, and then what percentage of your accidents occur in like high drive time, and you start making your assumptions and picks over what's going to change. And similarly, that then changes severity because one of the things we've learned is that the losses that go away quickest are the bumper-to-bumper accidents in high traffic, and those tend to be low severity. So while a 10% decrease in frequency is great, it's not necessarily a 10% decrease in loss costs. And the reason I'm getting into or micro, since Tom took the macro on this and laid it out, is hopefully just to give you confidence that we look at this in a very detailed way and a thoughtful way, and our teams work through what the expectations are, and then we're able to move quickly on the fly when we're either above or below it with what our prediction is.

Great. Thanks for that comprehensive answer. I appreciate that. And if I could just sneak in just one more on just a quick numbers question. If I think about that 50 basis point improvement in the expense ratio, if I take out the one-time-ish items in 2Q, and then it was roughly 80 basis points in the first half versus the first half last year, is that -- and I know there are a lot of different moving pieces there, but is that the sort of level of year-over-year improvement that we should expect to see going forward?

Yes. I would say transformative growth is about reducing expenses. In part, it's just one component, but it's about reducing expenses, so we can improve our customer value proposition with better prices, without impacting our margins. But it's going to bounce around a lot by quarter. You got advertising. We're going to have – we have new advertising. We're going to launch – you have some technology spend. But you should expect the overall trend going down. I don't think we could predict how much it will change by every quarter because – or set a target that way, because then it leaves unnatural consequences. So – but you should expect it to keep going down, yes.

No. Thanks fair. Thank you.

Thank you. Our next question comes from the line of Suneet Kamath from Citi. Your question, please.

Thanks. Good morning. I wanted to go back to the frequency benefit issue, if I could. I get that you want to be more surgical in terms of how you're approaching this, as opposed to more broad-based, which makes sense, but, I guess, the question is, do you – are you comfortable or confident that you won't see an impact on policy growth to the extent that some of your competitors actually stick with much more of a broad-based approach?

Well, it's always hard to tell what customers will do. We've always found that precision works for us on the long term, and that those people who rush to grow and not have precision to it, end up having to fix it later. So if you look at homeowners, people trying to grow in homeowners are taking huge losses. Like, we just don't think that's the right way to build the business, because you get the customer for a price that's not appropriate, and then you either have to lose them or manage them to a much higher price. So, all I can say is, what we'll do. But it's a reasonably – it's a highly competitive market, but it's a thoughtful market, right? Like people are not crazy in this market and trying to use bad economics. So all of our major competitors understand their business well and why they might make different bets at different times. I feel like, the basis of competition is still going to stay the same. Whoever is the smartest, win.

Got it. Okay. And then I just wanted to ask one on the National General deal, if I could. I see that you're still guiding to this, I guess, high single-digit accretion, I think, was the original guidance. But when we were running the math, just based on where expectations were for Allstate and Nat General based on consensus, we were getting something close to high single digit or in a high single digits EPS accretion, without factoring in any cost synergies. So I'm just trying to understand, is there any help that you can give us in terms of what you're building in with respect to cost synergies? I don't think you talked about on the last call, but any help there would be appreciated.

What, I would say is, we don't own it yet. So our projections haven't really changed. We do have cost saves in the middle of it that we figured out in terms of what we factored in when we bought – when we agreed to pay the price we did. And it did lead to accretion, as you point out. But right now, we're working on -- Glenn and Barry are working hard on what is the transition program. So as we get more specifics on things that we can lay out for you that are, here's our measurable goals, we'll do that. But, right now, we're just like a month into it.

All right. Thanks, Tom.

Thank you.

Jonathan, I think we have time for one more question.

Certainly. Then our final question for today comes from the line Ryan Tunis from Autonomous Research. Your question, please.

Good morning. This might be a question for Glenn, but it seems like there's some decoupling between trends in miles driven and trends in frequency with, I guess, the latter declining quite a bit more. I guess my question is, what indicators that you're seeing in your book do you think are actually the best predictive indicators of what frequency is doing in this current environment?

That's a tough question. I will let Glenn do that one.

Thanks. So it's a good question. And you're right about the decoupling, so I'll give you a little background on that. Not all miles are created equal. If you're driving in a high-traffic environment, you are much more likely to get an accident because it's less forgiving. You take your eyes off the road. You're distracted driving, which, by the way, I don't recommend at any time. But if you're doing it in loose traffic and you make a mistake, you're less likely to bump into somebody than when you're doing it in high traffic. So what we're seeing is with commuting miles being way down, that has a little bit of an exponential effect on frequency. So if miles came back to a new normal for non-commuting, but were lower only on commuting, it would have a disproportionate effect on the frequency. So hopefully, that answers your question. It's – we're looking at the data in that way. And our – with the partnership with Arity and the partnerships they have with a lot of the non-insurance companies out there, and they get a lot of miles-driven data on those, and that what they do for us with our own data, is extremely helpful in that.

And Ryan – and Glenn's example also is goes for rural and urban too, right? So think of the same thing, not just time and day also. So Montana, the frequency is not down as much as it is in urban areas because there's fewer people. There's always less congestion. So it is a difficult question. But the good news is we're on top of it. I know you all have another call. Thank you for participating and listening to our story. We'll talk to you next quarter.

Thank you, ladies and gentlemen, for your participation in today's conference. This does conclude the program. You may now disconnect. Good day.