Allstate Corp (The) (ALL) — Q3 2019 Earnings Call Transcript

Well, thank you, Jonathan. Good morning and welcome everyone to Allstate's third quarter 2019 earnings conference call. After prepared remarks, we will 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 on our website at allstateinvestors.com. Our management team is here to provide perspective on these results and discuss the strengths and leading competitive position of Allstate's homeowners insurance business. 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 2018 and other public documents for information on potential risks. Beginning in the fourth quarter of 2019, Allstate plans to announce catastrophe losses every month, removing the current $150 million reporting threshold. The enhancements to our catastrophe announcement process increase transparency for analysts and shareholders. As many of you know, this will be my final earnings call as the leader of our Investor Relations team. As I've transitioned to a new role in our P&C finance area, I’m leaving Investor Relations in the capable hands of Mark Nogal, who will be a great partner for all of you going forward. I'll now turn it over to Tom.

Well, good morning. Thank you for joining us today to stay current on Allstate. Let's begin on Slide 2 with Allstate's strategy. Our strategy has two components: increase personal property liability market share and expand into other protection businesses. Starting with the upper oval, the personal property liability market provides consumers protection. We insure their autos, homes, motorcycles, boats, and personal liability. We use differentiated products, sophisticated pricing, claims expertise, and a building in integrated digital enterprise to lower costs, which I'm sure will come up later. We're also diversifying our businesses by expanding our protection offerings, and that's highlighted in the bottom oval. Allstate offers customers a circle of protection. It's a wide range of products from Allstate life, workplace benefits, commercial insurance, roadside services, car warranties, protection plans, and identity protection. These growth platforms have extremely broad distribution, including major retailers, insurance brokers, work sites, auto dealers and manufacturers, telcos, and directly to consumers. They now comprise about 75% of our policies in force, although a much smaller percentage of our overall premiums. We leverage the Allstate brand, customer base, and capabilities to drive growth in these businesses. Some of these businesses also support the property liability businesses. This strategy creates shareholder value through customer satisfaction, unit growth, and attractive returns on capital. It also ensures that we have both sustainable profitability and a diversified business platform. We move to Slide 3. Allstate’s strategy is delivering growth and attractive returns. Revenues exceeded $11 billion in the third quarter of 2019; Property Liability earned premium, which grew 5.6%. Strong operating capabilities enabled us to generate net income of $889 million in the quarter. Adjusted net income was $946 million or $2.84 per diluted share, as you can see in the bottom of the slide. Returns were also attractive with an adjusted net income return on equity of 14.2%. We turn to Slide 4. We also did really well on all of our 2019 operating priorities. And we have five of those, as you know they focus on both near-term performance and long-term value creation. The first three priorities, better serve customers, grow the customer base, and achieve target returns on capital, are all intertwined to ensure profitable long-term growth. Customers were better served, as our Enterprise Net Promoter Score improved. Property liability policies increased by 664,000 from the prior year quarter to 33.6 million, as the Allstate and Esurance brands grew 1.9% and 5.9% respectively. Allstate protection plans, which of course was formerly SquareTrade, were at 89.8 million. Total policies in force now exceed 136 million, an increase of 40.7% compared to the prior year quarter. Returns remained excellent with most individual businesses performing well. The $89 billion investment portfolio had excellent total returns and generated $880 million net investment income in the quarter. Shareholder value is also being created by building long-term growth platforms. We increased telematics usage in the property liability businesses, which is supported by having industry-leading insurance solutions from Arity. Allstate protection plans are achieving the acquisition goals we established two years ago and sales in Europe are growing. Allstate identity protection, which we acquired about a year ago, is integrating its products into our customer value propositions. Mario will now discuss our results by segment in more detail.

Thanks, Tom. Moving to Slide 5, you can see that property liability results continue to reflect strong operating capabilities. Net written premium increased 5.8% in the third quarter, or $1.5 billion for the first nine months. This reflects policy growth in the Allstate and Esurance brands and higher average premium for auto and homeowners insurance across all three underwritten brands. As you can see in the middle of the left table, total policies in force increased 2% to 33.6 million. Underwriting income of $737 million was substantially better than the prior year quarter, due to lower catastrophe losses. Moving to the bottom of the table, the property liability recorded combined ratio of 91.6, which was 2.3 points better than the prior year quarter, reflecting a planned improvement in the expense ratio, offsetting an increase in the non-catastrophe loss ratio. The underlying combined ratio, which excludes catastrophes and prior year reserve re-estimates, was 85.0 through the first nine months of 2019. Moving to the right-hand table, Allstate brand auto and homeowners insurance net written premium increased 4.5% and 6.7% respectively, compared to the prior year quarter, due to increased policies in force and a higher average premium. Esurance auto insurance policy growth was 5.5%, which combined with average premium increases resulted in total net written premium growth of 8.3%. Encompass written premium increased 2.6% as higher average premium more than offset a small decline in policies in force. On the bottom of the table, you can see that underlying combined ratios remain strong across our brands. Esurance reflects primarily auto insurance, which has a higher combined ratio than homeowners when catastrophes are excluded, while Encompass, on the other hand, reflects a 60:40 mix of auto and homeowners insurance premiums. Let's go to Slide 6, which highlights investment performance that benefited from overall market returns and proactive risk and return management. The portfolio generated a strong 7.8% return over the last 12 months, of which 1.9% was in the third quarter. Approximately half of this total return came from interest income on the fixed income investment portfolio and returns on the performance-based portfolio. The remainder was due to portfolio appreciation, reflecting lower market yields and higher equity values. The chart at the bottom shows net investment income for the third quarter of $880 million, $36 million higher than the third quarter of 2018. Market-based investment income shown in blue, increased to $727 million from $683 million a year ago, reflecting investment at market yields above the portfolio yield. Performance-based income, shown in grey, was $202 million in the third quarter, $12 million lower than the prior year quarter. Slide 7 highlights results for Allstate life, benefits, and annuities. Allstate life, shown on the left, generated adjusted net income of $44 million in the third quarter, $31 million lower than the prior year quarter. This is largely due to the write-down of deferred acquisition costs, driven by lower interest rates and model refinements in connection with the annual actuarial assumption review. Excluding the impact of the non-cash unlock charge in both periods, adjusted net income was $86 million in the third quarter, an increase of $6 million, or 7.5%, compared to the prior year quarter. Allstate benefits adjusted net income was slightly lower than the prior year quarter, as higher premiums were more than offset by increased DAC amortization, driven by lower projected investment returns related to our annual actuarial review of assumptions. Excluding the impact of the noncash unlock charge in both periods, adjusted net income was $32 million in the third quarter, an increase of $1 million or 3.2% compared to the prior year quarter. Allstate annuities on the right generated adjusted net income of $16 million in the quarter, which was $4 million lower than the third quarter of 2018, due to higher contract benefits and reduced investment income. Adjusted net income of $43 million for the first nine months was substantially below the prior year, reflecting lower performance-based investment income in the first quarter of this year. Let's turn to Slide 8. Service businesses continue to grow with the number of consumers protected with policies in force increasing 67.7% to $95.9 million. This is largely due to Allstate protection plans. Revenues increased 27.1% to $418 million, as you can see from the lower left table, due to growth in Allstate protection plans and Allstate dealer services, as well as the acquisition of Allstate identity protection last year. Revenues for the first nine months now exceed $1.2 billion. Adjusted net income was $8 million in the quarter shown in the lower right, a $7 million improvement over the prior year quarter, largely due to improved loss experience in Allstate protection plans and Allstate dealer services. Slide 9 highlights the continued strength of our capital position and financial flexibility. In the third quarter, we issued $1.15 billion of 5.1% fixed rate non-cumulative perpetual preferred stock. The proceeds from this issuance were used to redeem $1.13 billion of fixed rate perpetual preferred stock with an average dividend yield of 6.54%. These actions will lower annual dividend costs by about $16 million. We continue to deliver excellent returns to shareholders. In the third quarter of 2019, we returned $775 million to common shareholders through a combination of $166 million in common stock dividends and $609 million of share repurchases. We have repurchased 6.7% of common shares outstanding over the last 12 months. Book value per share is up over $9 over the last 12 months. Now, I'll turn it over to Glenn, who will discuss our special topic of Allstate brand homeowners insurance and how we are positioned to generate industry-leading returns while growing market share.

Thanks, Mario. Homeowners insurance is a great business for Allstate, as you can see on Slide 10. Allstate is the second largest homeowners insurance provider in the United States with 6.6 million policies in force. We have written premiums of $7.6 billion in the Allstate brand and over $400 million in Encompass. Esurance is also expanding into homeowners insurance, using Allstate's capabilities, and now has over 100,000 policies in force. A significant portion of our customers bundle home and auto, which improves retention and overall economics of both product lines. The key message on homeowners insurance is that it generates substantial underwriting income and attractive returns on capital. To achieve these results, we target an underlying combined ratio in the low 60s to handle volatility that comes with catastrophe losses. Since 2012, we've generated over $1 billion of underwriting income on average annually, including catastrophe losses. As a result, returns on economic capital are in the mid to high teens. This profitability also provides diversification to auto insurance profitability. The graph at the bottom of the page shows homeowners insurance combined ratios for Allstate and the industry since 2012. As you can see, Allstate has consistently outperformed the industry. The result is that we've earned over half of the industry-wide underwriting income in that period. Turning to Slide 11, Allstate optimizes returns through sophisticated portfolio management. We've improved returns and decreased homeowners insurance volatility through advanced catastrophe modeling, geographic diversification of business, and strategic use of reinsurance. Our spread of the business across the country works to our advantage by providing a significant diversification benefit, as timing, type, and magnitude of weather events differ based on geography. As you can see on the U.S. map, we have a top three market share in 20 states which are shown in green, but much lower shares in states like California and Florida, prone to catastrophes like wildfires, earthquakes, and hurricanes. We take a proactive approach to managing our exposure to different types of risks. We substantially reduced our exposure in California to earthquakes by helping establish the California Earthquake Authority in 1996. We've decreased our underwritten policies in force there in the last decade. In Florida, we reduced our market share from about 10% in 2003 to less than 2% today. We also helped shape the Florida hurricane catastrophe fund, which provides reinsurance. We use a separately capitalized company there, Castle Key, and process external reinsurance. The overall objective is to meet customer protection needs while optimizing shareholder risk and return. We underwrite risk directly where we can achieve target returns. We also broker nearly $1.4 billion of other insurance property policies. This allows us to meet our customer protection needs, leverage our distribution strength with more customers, bundle additional Allstate products, but not directly underwrite risks outside of our underwriting appetite. In total, we manage our portfolio of states to target a combined ratio that generates attractive returns. For new competitors in homeowners insurance, the state level profit dynamic makes it difficult for them to achieve the same level of overall profitability or have the resources to expand. We shift a substantial amount of catastrophe risk to reinsurance markets which reduces our capital requirements and protects annual returns. The reinsurance program covers individual large events, utilizing traditional reinsurance and alternative capital. The current nationwide reinsurance program provides over $4.3 billion of limits above a $500 million retention for any single event. We also use an aggregate cover in case there are multiple events below $500 million. This provides additional protection in case the accumulation of those events throughout the year exceeds $3.5 billion. Moving to Slide 12. We're not standing still and we're constantly innovating in this space. We're focused on customer value and ease of doing business to accelerate growth. We streamline the homeowner quoting process by using both proprietary and third-party data sources to increase efficiency and accuracy. Using this information, we've reduced the number of questions asked in the quoting process from over 40 to just three when bundling homeowners and auto insurance together. In many cases, after our quote is complete, there is an inspection of the home. Technology, such as aerial imagery and predictive modeling, has enhanced the speed and efficiency of those inspections and lowered our expenses. We continue to enhance the design of our homeowner’s product while increasing our pricing sophistication. Our homeowner’s product, House and Home, is better able to address severe weather risks and unique customer needs. For example, the product includes a graduated roof coverage schedule, though it still provides the ability for customers to purchase full replacement if they choose. House and Home now represents 90% of our new business and about 45% of our total policies in force. The pricing of House and Home is more sophisticated than traditional homeowner’s insurance products with more occupant and resident characteristics. We've also improved the efficiency and effectiveness of our claims handling through technology and innovation. We leverage our scale, data, and analytics to rapidly deploy more than 700 full-time catastrophe resources to quickly help customers when they need it most, while mitigating damage and managing costs. We use aerial imagery to improve our efficiency and customer experience. We've expanded virtual claim handling capabilities, including the use of drones, airplanes, and satellites, so that now nearly 70% of all wind and hail claims have some aspect of the claim handled virtually compared to less than 10% in 2017. The bottom line is Allstate has a significant competitive advantage in homeowners insurance. We'll continue to leverage our scale, pricing sophistication risk management, distribution system, and claims capabilities to deliver industry-leading returns and market share gains. We'll now open the line for your questions.

My first question, I wanted to spend some time on the expense ratio and property liability. It's been low now for a couple of quarters and I know last quarter, you guys had pointed to a combination of improvement in processes, automation, as well as incentive comp driving down the second quarter expense ratio. Could you just give us a sense if it's kind of the similar components that drove down the ratio in the third quarter, and just how we should think about modeling through that expense level going forward?

Yes, thanks Elyse. It's Glenn. We're definitely focused on improving expenses. We've been fast at this and going after it hard. I mentioned a few of the examples last quarter where we've been able to reduce customer inquiry calls, which is a win-win, because it's a better customer experience, because we've eliminated the need for those calls on the front end, but it also is more efficient on the back end. I mentioned in the special topic there some of the aerial imagery and data analytics we're using in both claims and to reduce inspection. So we've gone after some real tangible ways we can manage expenses. As with any quarter there is a mix of things in there. So similar to last quarter, there are some components to compensation, some components of marketing, and some components of sustainable improvements in the baseline of that. But we've made some real tangible improvements that we will sustain and we're going after expenses in a real way because we think it's a path towards growth where we can maintain margins.

Can you break down and give a little bit of color on what might be in a sustainable bucket versus maybe what was kind of one-time in nature in this quarter and the expense level?

I don't know that there was a lot of one-time in this. So we've got where it is sustainable. It says operational improvements, which is a meaningful chunk of the change that we've got. When you look at some of the compensation components, we've managed our employee compensation and agency compensation over time and you reset every year with a new program. So there's a little bit of benefit in there from that, that we were short to our growth aspirations in the year, but otherwise we've got sustainable expense prudence in here.

Elyse one place this time, one place we would like this quarter was marketing, which we will dial up as we talked about last quarter, we're dialing out there, there are some select markets where we know that we can generate economic growth. That said, that won't change the overall trend line of expenses should be coming down over time. But every quarter is a new quarter.

And then my second question on, we've seen some of your peers that have seen higher bodily injury severity trends within their auto book of business, or just could you just give us an update on what you're seeing on the BI severity side and just if you've seen any pockets where trends have picked up in the U.S.?

Let me give you, first we can't comment on everybody else's numbers because bodily injury; those are long-dated claims, it takes three, four years to really get them paid out depending. The little ones get closed early, the big ones get closed late. So you always have some bounce of mix in the paid BI that you have to get underneath. That said, we feel good about where overall bodily injury reserves are set and our trends Glenn can talk about the pay trend.

Yes, so we feel good about where we are in bodily injuries. We put in the Q, you know, we’re running around medical inflation. As Tom mentioned on the reserving, it's all in the numbers I think would be a headline there because in our reserving actuaries, they are talented folks, they work very closely with our claims team, our underwriters, our product organization; they all work together to ensure that we have the right reserves on the book. If you look at the trends over time, we've had a lot more favorable development than unfavorable development throughout the years, which is a good bellwether for you to look at in terms of our overall trends.

My first question, I'd like to revisit how your guidance is going to look for 2020. John I know you mentioned in your prepared remarks that you're going to start disclosing monthly cat loss numbers. I also remember from a previous presentation that you said you were going to shift from an underlying combined ratio target to a target return sort of guidance. I'm just wondering if this is a trailing ROE target, do you plan to adjust it for changes in interest rates, etc? Just looking for some color there?

Well Greg, thanks for the question. We love the fact that you are paying attention to what we say; it makes real good sense. The monthly cat numbers address what people don't like, is it above 150, what if it's 149? We felt like it was confusing. So we're just putting out monthly; you can do what you will with it; everyone has different things they use it for. So just to make your lives easier and that’s more transparent. As it relates to the return on equity as we're replacing the annual underwriting - underlying combined ratio guidance with longer-term ROE goals. ROE is a better measure of the overall business results because it includes our businesses and includes investment incomes and ties directly to reported results, which includes catastrophes. The underlying combined ratio of course only reflects the property liability businesses and it excludes catastrophes which bounce around a lot from quarter to quarter and year to year. But on a long-term basis, as a management team we're accountable for making sure that we get a good return on homeowners with catastrophes included because that's the risk our shareholders take. When we report full-year 2019 results, we will give you what we think a long-term ranges on return on equity and debt management.

What do you define as long range and is this adjusted earnings that it's on?

Yes, it will be on adjusted earnings just because the accounting - as we move to fair value accounting the book value and reported income bounce around a lot with equity investments. So we feel like adjusted net income return on equity is a better measure of what we want to do. Long-term would be sort of what you expect to do over two to three years, but the goal really is to get shareholders focused on what's the overall return we're generating and the capital you have, and focus on the overall side of the business, as opposed to just one component. While auto insurance and homeowners insurance are extremely important to our business, they're not the only things we've got going, and that shouldn’t be the only thing we should be held accountable to. So ROE is just a much better measure under which you can judge how well you think we're doing as a team.

Great, thank you for that answer. My second question is, I just a follow-up on the expense. I noted from Page 9 of your supplement that your agency count was up year-over-year and sequentially, and your licensed sales professional account was up year-over-year and sequentially. I'm trying to reconcile this with the fact that you're not reporting any restructuring charges, which is unusual, considering the improvement you've realized. And then secondly, on a previous call, Glenn has mentioned something along the lines of an integrated services platform, so the numbers are kind of moving contrary to what I think that would be. Maybe you can provide some additional color there?

Let me deal with some components; Glenn can talk about what we're doing with our agency platform to make it more effective and efficient, including things like integrated service and what we're doing in compensation to drive that growth. As it relates to restructuring, we do end up with some minor restructuring charges, but they're not big numbers. As we move forward through integrated digital enterprise, we record what we think we have to record under the rules. If headcount goes down because we're using integrated digital enterprise, we have to record a charge. But we don't feel like there is one big bucket that's needed at this point, that we then carve out and don't come back and hold ourselves accountable for because it's some other charge in place.

So yes, thanks for the question Greg. As Tom said, you’re definitely paying attention to the details. If you look at the agent count, that reflects the investment folks are willing to make in the business based on the opportunity, and there's a good opportunity with Allstate. We are growing items and are profitable and have been successful. Agents are putting their money down; these are small business owners who are opening up a shop and going out and selling products and serving consumer needs. In terms of licensed sales professionals, that's reflective of their hiring, mostly on the sales side. That said, I’ll mention as Tom said, a compensation component leading into next year as well as the integrated service you brought up. From a compensation standpoint, we're leaning more toward new business production. We're interested in growing, so as we increase marketing, we've improved our expense and we’ll be more effective and cost-effective for customers. We lean more into new business production in terms of the compensation as we shift in that direction. We think that's a good system-wide approach to drive growth. In terms of integrated service, we have talked about it before; it's in the early stages. It's a big system. We have 40,000 people in that agency system, when you take the agents and all of their employees, licensed and unlicensed, sales and service. So today we do that service in a decentralized way. A lot of the service is done in individual offices, which are not scalable, which is not ultimately going to be as cost-effective as you can do it in a more scaled way. So we built an integrated service model. We're doing that with - we're getting into hundreds of agents, not thousands at this point. So from a scale perspective, you wouldn't see it showing up in the numbers that you looked at there at this point.

And Greg, just the change in agency compensation; we don't expect to raise overall compensation as a percentage of premiums.

Correct. Yes, it was a shift.

Can you talk about the increase in property damage frequency and Allstate brand auto this quarter? Are you guys sort of just a little surprised to see that given the more recent trends?

So first of all, I always like to go back to the overall profitability of auto, and we're doing very well. The combined ratio is at 92, some of that is the expenses, and I'll come back to sort of some of the intentionality there. Frequency, first of all, it's hard to predict. You can't predict what is going to happen in the next year, next quarter, but you can understand the trends and where things are moving. So miles driven are up. We saw some change in what had been a declining trend to frequency. That said, it was a little bit mixed. So I'll point you to a few different numbers. Gross frequency was up two points, as you pointed out. But paid frequency, which on a short-term tail line tends to be pretty accurate, was flat. BI frequency was slightly down. So I look at the overall frequency picture and say: It's kind of a mixed story in here and something that we're keeping a close eye on. Now you go a level below that, and at a local level, every state manager is out there looking at their competitiveness, their trends, their frequency, their severity in every state. And we're managing that profitability at that local level. So that's what rolls up into a really favorable return, as we've delivered. Lastly, on expenses, I want to keep going back to that because part of the reason or the main reason you go after expenses the way we are is actually to allow you to let the loss ratio float up and deliver the same return. And that's what it looks like in our ratios when you're giving back more value to customers.

And then the second question going to homeowners, also saw an increase in severity there. I guess just given the amount of pricing that the companies have pushed through over the last year, I was still surprised to see the deterioration, actually your loss ratio from a weak prior year quarter to begin with, even when that increase in severity. So can you maybe talk about that dynamic of where the severity has come from and why the price increases we've seen to date have not been sufficient to offset it?

Homeowners is one of those businesses where it's really difficult to look at it by quarter regarding severity. Because it's obviously impacted by weather; sometimes it's catastrophes, sometimes it's not a catastrophe. So you have to really look at it on kind of a rolling 12-month basis. When you look at our business, it's got great returns; we feel really good about it. If you look at the average premium, it's way up. And when you go underneath that, as Glenn said, we'd like to segment this down, then we segment it by state, we segment it by coverage. In homeowners, depending on what kind of loss you have, it changes your paid severity a lot. And again, those tend to be relatively short-tail lines. But so a fire loss has a much different severity if the house burns, than obviously if somebody runs into the garage door, or hail damage can tend to be much more expensive because it could take out a whole roof than some leaky pipes. It really depends on the mix; theft is not as big as water damage. So, what I would say is Glenn's comments on this special topic, we feel really good about homeowners. It's a really good business; making really good money, much better. I want to underline that with a 9% share of the market, we've captured half of the overall profits generated in homeowners in the industry. We feel like that's indicative of good operating expertise and capabilities.

My first question was on commercial auto and the ride-sharing agreement with the major providers there. One other competitor in that market has had some issues negatively. So just wanted to understand how Allstate is going to position that business?

So let me start off and say that we call it a shared economy business, so it's not just ride-sharing and car-sharing. We look more broadly. Obviously, you've focused on the largest customer we have, but we have a handful of others that we're building an entire team around. We are well aware of other competitors in the industry. The particular competitor you talked about strengthened their reserves in 2016 and 2017 when the industry was kind of early in development. We started last year, and as I think in prior calls I've talked about, we triangulated our loss reserves on the basis of the prior history provided by the transition network company, that we used our own internal personal lines and commercial auto experience. We've also looked at industry experience and our telematics type of information that we get – we're just at the state level. We’re comfortable with where we're sitting. We're still recording essentially at the priced amount for the reserves, primarily because the vast majority of the coverage is for long-tail coverages. It's still early, 19 months since that first month. I think long-tail coverages historically take longer than expected. We look at this every quarter. As you may remember, independently, our reserves are reviewed and set by a team that works in claim reserve; looking for Mario and not for me. So we have different eyes on it from our actual department, our financial department, and obviously in the business too.

It does, it does. And then on a separate topic, thinking about the underlying loss ratio in the protection business, having that show an increase year-over-year. My sense is some investors are a little concerned about what's going on in terms of price competition in auto and then also the potential for claims inflation to increase. So can you just remind us how Allstate is managing that issue to restrain underlying loss ratio deterioration?

Yes, Jay, let me provide some overall context and Glenn can jump in on the relationship between expenses and loss ratio. Our auto insurance has the largest product line. It comprises about 65% of our total premiums and it's obviously an important lead line as well because it helps us expand into the homeowners insurance business, which is also highly profitable. So it's a really important question: Are you going to maintain your profitability on auto insurance so you can keep growing it? Before we go into the specifics on expenses and loss ratios, let me just talk about principles for a minute and a little bit of our history. We have some guiding principles on our growth. First, we only want to grow when we get a good return for our shareholders. It seems simple, but not everybody follows that path. As you know, auto business generates really attractive returns, it's in the low 90s when you add that and homeowners' insurance to the equation; that's a big driver of our return on equity, which was 14.2% for the last 12 months. When you adjust out annuities, it's over 17. So we're getting a really good return on that business and what we've put in place has been working. Secondly, we segment the business by growth and profitability by business, geography, customer group, and risk to get attractive margins in each segment. This gives us a solid base from which we can be sustainable in terms of growth. For example, if auto margins drop in one state or one type of coverage, we have the benefit of good margins than auto insurance in another state or on another coverage or in the homeowners business. Our third principle is we create shareholder value by focusing on profit and growth simultaneously, but we default to profit if necessary. We employ this philosophy across everything we do. For example, when we first bought insurance, it was running a combined ratio well over 100. We invested heavily in that. We believe and know now that that growth created shareholder value because the return on the business we were writing was higher than our cost of capital. The accounting, however, required us to write up all the advertising expenses upfront, which resulted in an underwriting loss. After the first year, of course, that advertising expense goes away, you capture all the underwriting profit, and it's very profitable. It took a while before that profitable business could offset the negative of first-year business, especially when you spend $200 million a year on advertising. As a result, that growth reduced underwriting even though it’s positive for shareholder value. We try to think about whether we are creating shareholder value. Our first principle is we create shareholder value when we're doing this. We can do it, and we chose to do it on insurance because we had strong profitability from the Allstate brand, auto, and homeowners business. The concept of portfolio growth, Glenn talked about how it applies at the auto insurance level; it applies for homeowners. It also applies across the whole company. Now, insurance is two and a half times the size when we bought it. A similar situation exists for Allstate protection plans. We acquired Square Trade, which has been renamed Allstate protection plans a couple of years ago for $1.4 billion, equivalent to about $4 a share of the Allstate share. Now their business is over twice its size, we’ve got 89 million policies in force, and it has great distribution; it's growing the European cell phone business, and now it has reported profits albeit small. Valuing that business on an earnings basis obviously understates shareholder value. Our challenge has been to ensure we give you the information to see value creation but not get everybody so focused on the 65% of the business that that's the only thing that matters in terms of creating shareholder value. It’s an important question; how are you growing that two-thirds of the business to ensure you make money. I also wanted to take this opportunity to say don’t forget about everything else because there's a third of the rest of the business that drives shareholder value. When we focus solely on auto insurance margins, it takes people's focus off other vital aspects. This is why we moved to ROE. We are very focused on auto insurance profit. We feel good about it. We need to ensure we keep doing that. Glenn will talk about what Mario mentioned: our intentional strategy to grow the business and provide a good value to our customers by reducing expenses. This enables us to allow the loss ratio to drift up, which allows us to maintain margins and continue to clock that high teens return on equity from that business.

Yes, so thanks, Tom; as you weigh out the principles there, clearly we're in an environment in auto that meets those growth principles, so we want to go after that. That said, it's an environment we've chosen to grow in, but it's a competitive environment; we’ve got significant increases in advertising out there. We have an extremely low CPI. We acknowledge all that in terms of the competition, and we're doing things smart as to how we grow as opposed to just chasing it. I think you could call our combined ratio now a little bit of a restructured combined ratio because, as you point out Jay, the loss ratio was up; the expense ratio was down. I would argue that it's a lot better than the other way around; that if we were here having this conversation and we were like a point or two up on expenses but we had a huge tailwind because frequency dropped through the floor in the quarter, I think we’d be having a different conversation. So I look at this as a positive, sustainable way for us to go after growth and show that we can do so in a way that continues to provide that mid to high teen return.

I guess I'm staying on the expense ratio improvement which is enabling you to grow more. There was a media report about shifting certain customer service responsibilities out of the agencies and into call centers which can serve customers, potentially more cost-effectively. I'm curious if that's part of the reason agent count is falling and whether this could be a permanent decline or just correct me if I'm off on that.

First, I may go up for a minute, Mike, and say we want to use technology and our people to do a great job for our customers. With what you can do in technology today, you can make your people a lot more productive, provide great service, and spend less money. We've talked multiple times, and the productivity of a claims adjuster is now many times greater than it used to be because they're not driving around from body shop to body shop. They're sitting in front of a computer, looking at pictures, deciding what should be done with their customers. This innovation is running through our business. We have lots of ways we’re working on doing that, and that gives us the confidence we can do a better job for our customers with fewer expenses. Glenn talked about integrated service, which is what you're referring to. You don’t see it in the numbers. It would be what I would say today. It's not that we started with 50 agents this year; we're getting the processes down. We're not going to turn this loose on 22 million customers until we ensure it works really well. Because the system we have works well. Our net promoter score was up again this quarter. We like what our agents are doing, that said, we think we can do better by leaning into innovation rather than just going full throttle. So you're not seeing anything on integrated service that’s reported in the media. The reality is we're doing everything possible to provide great service to our customers while supporting our agencies but doing it in a more cost-efficient way. We're all in on that.

Okay, great, that's helpful. And lastly, Glenn, in your prepared remarks, you talked about the state-level profit dynamics making it difficult for competitors to achieve Allstate's homeowners profitability. Maybe you can elaborate on that. I'm curious if there are parallels to auto insurance as well. Thanks.

Yes, thanks for the question on that. I think the basic premise is, you know, we have a lot of scale, and we've got breadth across states in our blend and our mix across the state not only reflects a 50-state view. We’ve also been thoughtful and made choices about where we're larger and smaller, given the types of risks we face. This is really challenging for the smaller or newer carrier going into it. It’s getting that type of breadth across a system where you can offset your highs and lows. It's similar to what you do with an investment portfolio. You're mixing your investment portfolio in different ways so that when one thing is up, another is down, and you're getting an overall good return. That's really challenging if you're starting out and you're only allowed to buy four stocks.

Let me also add to that. Let's just compare homeowners to auto insurance. Homeowners is more volatile on an individual location basis than auto insurance. So you should put up more capital for homeowners insurance than you should for auto insurance. In addition, with homeowners insurance, you get very little investment income because it's relatively short-tail, whereas auto insurance gives you a decent amount of investment income off of it. As a result, you have to run a combined ratio in homeowners insurance that's below what you'd run in auto insurance. So your combined ratio in homeowners should always be below your combined ratio in auto. That's not true for everybody, but in total, you've got to get there. The problem is it’s like if there is hail storm in Dallas one year and not for two more years, there is hail in Oklahoma, because it got dumped in Oklahoma before it got to Dallas the next year. If you're only in Dallas and you’ve got to earn that low combined ratio, and the hail happens to get you in that year, it's pretty hard to have the money to expand into Oklahoma. That's what Glenn is talking about. The businesses, when you look at the homeowners business, as more people get into it, I think it's worthwhile focusing on what is their actual return on capital in that business, as opposed to just growth.

This is Crystal Lu and for Ryan Tunis. Our first question again on the expense ratio. It seems like the expense ratio has improved a lot, but it doesn't seem to be translating into auto policy growth acceleration yet. It seems like some of that expense improvement came from advertising, which tends to drive growth. So I'm wondering what actions are being taken right now, where you're investing in growth and going to see that policy growth acceleration in the future?

Well, first, if you look at our auto insurance growth, it's about two points, which we think is more than the number of cars and miles that have grown in the United States. So we think we're capturing some market share, albeit a small amount of market share, and we would like to have more. We are investing more in marketing, but that doesn't change the overall trend that Glenn talked about. I mean, our overall trend is to reduce expenses, be competitive at price, and maintain margins at levels that are attractive. Is it going to be the same level each quarter? It bounces around a lot depending on the frequency. However, we like the returns. We're good at getting returns in auto insurance, and we've proven that we can grow it, too. We're going to continue to work on producing expenses so we can continue to be competitive.

On the auto rate increases that you guys have been getting this year, it doesn't seem to have slowed very much in the first nine months, despite efforts to pass along more savings to customers and grow the business. So can you kind of describe the auto pricing environment right now and how you're thinking about auto pricing in terms of growing?

Yes, thank you, Crystal. First of all, we manage all of our pricing on a state-by-state basis. In terms of slowing, I would look at it relative to the loss cost trends. It's slowed. Over the trailing 12 months, there's been a 2.2% auto rate increase, which translates into an average gross premium of 3%. However, if you look at the severity trends running, it’s about five this quarter; higher than that, we've not had a rate that has kept up with that. As we previously discussed, we've offset it with lower expenses, and that's where you provide more value to the customer—by not taking a rate that has to keep up with those types of inflationary factors. But as Tom has said several times, we're managing this for profitable growth, and we're committed to a strong return. When we look state by state at that, we scrutinize our competitiveness, what our price looks like relative to our competitors across different customer cohorts in each market, and also, what our return on all of the cohorts are within each state. Our pricing has lagged behind the loss trend; this is part of what we're offsetting with expenses.

It's hard, Crystal, and I'm not trying to be evasive, the level of sophistication in pricing today in auto insurance is so high that while the overall numbers are important to maintain, looking at the trend in margins and definitely watching those, the growth largely depends on the sales you're growing in and what your new business discount is. There are a lot of things we and our competitors do to capture business, which generates long-term economic growth. It's a pretty uneven playing field; we're not noticing a dramatic increase in total pricing increases right now, you know State Farm has taken some decreases but given where they are in overall price, it’s likely they need to take some price reductions because they are pretty high price. We feel competitive regarding pricing right now.

That makes sense. And then one more quick one on the auto bodily injury reserve releases this quarter. Could you just give a little bit more detail on what accident years those are coming from and whether that reflects the level of severity being in line with medical inflation, is that running at a lower level than you were expecting in those years? Thanks.

So Crystal, this is Mario. First, I'd just reiterate what Glenn said regarding our processes; we establish reserves and look at reserves every quarter. We have some pretty robust processes that we follow. We take all relevant facts and circumstances into account, both internal and external trends. We also consider any changes in claim handling practices. We have a thorough reserve-setting process and we tend to be conservative in how we set those reserves. What you saw this quarter in terms of the releases, they are predominantly in the Allstate brand, auto injury coverages, and we continue to feel good, not only about the severity trends that we're seeing in the current year but also seeing favorable development in the prior years that is better than we expected when we established the reserves. It's coming across a number of accident years, but we continue to feel really good about our reserve position and our injury severity trends.

The amount from each prior year wouldn’t really help you. It’s what we picked as opposed to the trends—the absolute trends are in the triangles and so forth.

Yes, thank you. I was looking at the policy count numbers in auto, and it seems like there's a slowdown, but I'm wondering if we can sort of break it down to gross policy ads versus gross policy declines. Are Allstate customers sticking with you, and have the new customers slowed? How should we parse that out?

That's a pretty detailed level of question, Josh. So first, overall retention was flat this quarter, basically for statistical reasons; it’s down slightly but it’s basically flat. This means we're keeping in total as many or as a percentage of our customers as you correctly assumed that new business doesn’t have as high of a retention rate as people who've been with you for 10 years. By the time somebody has been with you, customers generally don’t wait two, three, or four years. They tend to stay with you at really high retention levels. Growth will drive your retention down despite the fact that we were growing over the last couple of years. Our retention has gone up because we're doing a better job on customer service with our net promoter score up. So I don't - I think in terms of our overall growth, you should expect to see more of it come from new business in the future than further increases in retention, if that's helpful.

And when I look at the decline in ad spend, I guess you can correct me if I'm wrong, or is that directly tieable to the amount of new business coming in?

You sound like you're looking at a specific number. What we can do is Glenn can answer for you what the new compensation looks like. We said we could spend more money in advertising. We're working on ramping that up and we expect to continue bringing our overall expenses down as we do that over time. That may not be every quarter, but we're headed down in that direction. Glenn can talk about what we're doing on compensation.

A quick add I'll make to the marketing because this has come up a couple of times. It's during a year-over-year basis, the marketing was lower. On a sequential quarter basis, it was up. We're comparing to a quarter last year where there was a new brand launch, and you know, so we're ramping up and we'll continue to do that with marketing. From a compensation standpoint, as we talked about earlier, it really is about shifting towards new business production. Ultimately, you compensate agents for the service they provide to customers and for going out and getting new business. Within the confines of that total budget, we're shifting money toward new business production to incentivize that more in the compensation plan going into next year.

What we're trying to do is go on a great customer value proposition. We want a good price. We want good service. We're lowering expenses you should expect to see us continue to lower expenses as we move forward from here. As loss ratios go up, that just means you're offering greater value; they’re paying less for expenses and more for fixing issues.

Okay. First, thank you for being on the call. Let me thank John for being both a transparent and direct source to you for a little over three years. He's done a fabulous job. We are excited to see him move on in his career and I know Mark will do a great job. He's worked with John, so it'll be seamless for you and us. As it relates to Allstate, yes, we made really good progress this year on our strategy; our operating priorities remain focused on profitable growth. We'll talk to you next quarter. Thank you.