Good day, ladies and gentlemen, and welcome to the Allstate third quarter 2015 earnings conference call. At this time, all participants are in a listen-only mode. Later, we will conduct a question-and-answer session, and instructions will be given at that time. If anyone should require assistance during the program, please press star then zero on your touchtone telephone. As a reminder, today's program is being recorded. I would now like to introduce your host for today's program, Pat Macellaro, Vice President of Investor Relations. Please go ahead.
Thank you, Jonathan. Morning. Welcome, everyone, to Allstate's third quarter 2015 earnings conference call. After prepared remarks by Thomas Wilson, Steven Shebik, and myself, we'll have a question-and-answer session. Yesterday afternoon, we issued our news release, filed our 10-Q for the third quarter, and posted the results presentation we will use this morning, along with our third quarter 2015 investor supplement. All of these documents are available on our website at allstateinvestors.com. Our discussion today will contain forward-looking statements regarding Allstate's operations. Allstate's results may differ materially from these statements. Please refer to our 10-K for 2014, the slides, and our most recent news release for information on potential risks. This discussion will contain some non-GAAP measures for which there are reconciliations in our news release and our investor supplement. We're recording this call. A replay will be available following its conclusion.
I'll be available to answer any follow-up questions you may have after the call. Now I'll turn it over to Tom.
Well, good morning. Thank you for investing time to keep up on our progress at Allstate. I'll cover an overview of results. Then Pat and Steve will go through the details. Our operating team is also here to provide additional perspective when we get to the dialogue section. Let's begin on slide two. We had good overall earnings this quarter, primarily reflecting the continued strength of our homeowners insurance business. As you know, we repositioned this business so the underlying combined ratio would support good annual returns even with high catastrophe losses. This quarter, we also benefited from lower catastrophe losses than this quarter last year. This continued strength in homeowners has also led us to reduce the capital requirements for this line, which has now become a competitive advantage. We also made progress on improving auto insurance returns.
As you know, in the fourth quarter of last year, the combined ratio on this business began to rise, which reflected an increased frequency of losses and higher severity per claim. As a result, we instituted a comprehensive program based on the business model and practices that have been highly successful really the last 14 years. That, of course, includes raising auto insurance rates. For the first nine months of 2015, the approved rate increases, this is in absolute dollars, was about twice the amounts achieved in the average of the two prior years. That increase in approved rates has begun to be realized in premiums earned, but of course, the impact will continue to increase over the next year as policies renew. We also made underwriting standards more restrictive, which has the effect of reducing the higher loss ratio of new business.
As a result, Allstate brand auto policy growth declined to 3.1%, with larger reductions at Esurance and Encompass. Given the cost trends we and others are experiencing in auto repairs, there is also a heightened focus on both effectively and efficiently managing claim loss costs. We also reduced advertising expenses and took some other actions so that the underwriting expense ratio in the quarter declined by 1.4 points to 24.9. All of these actions are well thought out. We balance short-term profitability and long-term economic value creation for shareholders. In addition, they are tightly integrated and implemented in a highly targeted and local manner. The recorded combined ratio for the quarter was 93.6, which generated $491 million of underwriting income. The underlying combined ratio for the first nine months of 2015 was 89.1. That is slightly above the outlook we established at the beginning of the year of 87 to 89.
We now expect the full-year underlying combined ratio to be no higher than 89.5. Common shareholders have received $2.6 billion in cash so far this year, due in part to an 8% reduction in the number of shares outstanding. If you move to the box on the bottom of the slide, operating income was $610 million, or $1.52 per share, which is 9.4% higher than the prior year quarter, which you can see in that little red box. Premiums earned were up 4.7%, reflecting an increase in average premiums and a 2.3% increase in the number of items in force. The overall return on equity was 12% on both an operating and net income basis. Let us move on to slide three. We will go through the results for the individual segments. Our five operating priorities for the year are shown at the top, and those have not changed.
The Allstate brand in the lower left is our largest segment, of course, and it comprises about 90% of premiums written and serves customers who prefer a branded product and value local advice and assistance. This business continued its moderate growth across all the product lines, as you can see from the first line in that box. Total policies in force are 2.5% higher than last year's third quarter. Auto insurance growth did decline slightly in the quarter. Homeowners policy growth was up slightly, and other personal lines policies grew at 3%. The underlying combined ratio for this segment was 88.3, as you can see in the red box at the bottom. Esurance, on the lower right, serves customers that prefer a branded product but are comfortable handling their own insurance needs. Growth continued to slow in this segment and was 3.7% versus the prior year.
The reduction in auto insurance growth as a percentage absolutely was not offset by the high percentage growth in other product lines, such as homeowners insurance. The underlying loss ratio did improve and was 105.3 for the third quarter. Encompass in the upper left is the smallest segment that we underwrite for and competes for customers that want local advice but are less concerned about the choice of insurance company. Encompass primarily sells a packaged auto and homeowners insurance policy through independent agencies. This business has gotten smaller, and policies in force are down 5.7% from a year ago due to lower retention and less new business. We took aggressive action to improve profitability in both auto and homeowners insurance. The underlying combined ratio is 90.9 for the quarter.
Answer Financial in the upper right serves brand-neutral self-serve customers and competes in a relatively small segment of the market with aggregators such as Google Compare. Total non-proprietary written premiums of $443 million for the first nine months of 2015 are 11.2% above the same period in the prior year. Overall operating results were in line with our expectations, we remain committed to increasing shareholder value by generating good returns, growing profits, and providing cash to shareholders. Pat will now go through the property liability results in more detail.
Thanks, Tom. Let's start by taking a look at the property liability P&L on slide four. Starting with the chart on the top of this slide, property liability net written premium of $8.1 billion in the third quarter of 2015 was 4.2% higher than the third quarter of 2014, reflecting the combination of policy and average premium increases predominantly from Allstate brand auto and homeowners. For the first nine months of 2015, net written premium grew by 4.5%, while policies in force grew by 2.3%. Property liability policies in force were 34.7 million at the end of September. Note this excludes 5.6 million Allstate Financial policies and 2.1 million Good Hands Rescue relationships. Catastrophe losses of $270 million in the third quarter of 2015 were $247 million lower than the prior year quarter.
Recorded combined ratio for the third quarter of 2015 was 93.6, which is one tenth of a point worse than the prior year quarter. For the first nine months of 2015, the recorded combined ratio was 95.8, which was six tenths of a point worse than the first nine months of 2014. The underlying combined ratio for the third quarter of 2015 was 89.3, the underlying combined ratio for the first nine months of 2015 was 89.1, both elevated over the results we experienced in 2014. Property liability operating income of $550 million in the third quarter of 2015 was a half a percent below the prior year result, while the $1.3 billion of operating income through the first nine months of 2015 was 5.9% below the first nine months of 2014.
Bottom of the slide contains growth trend information as well as a view of the property liability recorded and underlying combined ratio trends. Premium and policy growth trends are in the chart on the bottom left. I mentioned the drivers of premium trends shown in the blue line earlier. The red line represents policy in force growth and shows the slowing policy growth trend that's being driven by auto profit improvement actions in all three brands. Policy in force grew by 772,000 or 2.3% from the third quarter of 2014. Exhibit on the bottom right displays the property liability recorded and underlying combined ratios for the third quarter of 2015. Both the recorded and underlying combined ratios were impacted by higher auto losses in the third quarter of 2015 versus the prior year quarter. Slide five highlights the margin trends for Allstate brand auto and Allstate brand homeowners.
The chart on the top left of the slide provides a view of quarterly underlying margin performance for Allstate brand auto. Third quarter results were impacted by higher frequency and severity than the prior year quarter, resulting in a 5.2 point deterioration in the underlying combined ratio from what we experienced in the third quarter of 2014. On a sequential basis, the combined ratio was three tenths of a point higher in the third quarter of 2015 versus the second quarter of 2015. Chart on the right highlights the trends driving the change in the underlying combined ratio. Annualized average earned premium per policy, shown in the blue line, is beginning to pick up momentum given rate increases implemented throughout 2015. Average underlying losses and expenses per policy increased compared with the third quarter of 2014, influenced by higher frequency and severity, but lower expenses per policy.
Similar information is shown for Allstate brand homeowners on the bottom of this slide. On the bottom left, you can see the impact of low catastrophes in the homeowners combined ratio, which was a 72.5 in the third quarter of 2015. The underlying combined ratio of 60.9 in the third quarter of 2015 was nine tenths of a point higher than the prior year quarter and about equal to this year's second quarter. On a trailing four-quarter average basis, the underlying homeowners combined ratio is a favorable 61.8. The components of the third quarter homeowners' underlying combined ratio are in the chart on the bottom right. Average earned premium per policy increased to $1,079, or 1.5% over the prior year quarter.
Underlying losses and expenses per policy increased 3% in the quarter compared to the third quarter of 2014, but the difference between premium and underlying losses and expenses per policy is essentially the same in the two periods. Slide six provides an update on our multifaceted auto profit improvement plan. Last quarter, we discussed the four components of our plan to lower auto margins, three of them to adjust to higher accident frequency and one of them to adjust to higher claim severity. First, we sought approval for higher auto rates across the country in response to higher loss trends. Second, we implemented underwriting changes wherever we identified specific underperforming segments of business, including ongoing correct classification programs to accurately price policies. Third, we focused on claims operational excellence and precision given cost trends. Fourth, we reduced expenses across the organization.
While these actions in total will improve auto margins, they will also slow growth. Detailed results for these initiatives are shown in the bottom two tables. Allstate brand approved auto rate increases in the third quarter of 2015 were 1.6% of prior year-end total Allstate brand auto net written premium or $277 million. This brings the total for the year to 3.4% or $600 million in net written premium. An outcome of these profit actions that auto growth in the Allstate brand is beginning to slow. Property liability expense ratio decreased by 1.4 points in the third quarter of 2015 compared to the prior year quarter, reflecting expense actions taken across the company. You can see the impacts by underwriting brand in the chart in the lower left.
Bottom right-hand chart shows the net written premium amounts generated by the auto rates we've received approval for in the first three quarters for the past three years for the Allstate brand. You can see the total we've implemented through the first nine months of 2015 is significantly higher than both 2013 and 2014. We know that not every customer will renew their policy, and that some customers will decide to change the level of their coverage, which will result in lower levels of premium in aggregate than what's shown on this chart. This is common customer behavior and why we believe premium change is a key process that's enhanced by an Allstate trusted advisor. Cumulatively, the rates which we received approval for in 2014 and through the third quarter of 2015 will be worth $816 million in earned premium through the third quarter of 2016.
This analysis only includes rates approved through September 30th. We continue to evaluate and run our business on a local market by market basis. We continue to aggressively pursue rate increases and adjust our actions going forward to ensure appropriate auto returns. Slide seven highlights combined ratio and top-line trends for both Esurance and Encompass. Chart on the top of this page includes combined ratio results for both companies. Esurance's recorded combined ratio of 106.5 in the third quarter of 2015 was 10.1 points lower than the same period a year ago, given decreased investment in marketing, along with a 3.7 point improvement in the loss ratio. Esurance's underlying combined ratio of 105.3 for the third quarter improved by seven points.
Encompass' recorded combined ratio of 101.3 in the third quarter of 2015 was 8.4 points lower than the prior year quarter, and benefited from an 11.1 point decline in catastrophe losses. Encompass' underlying combined ratio was 90.9 in the third quarter and 92.7 through the first nine months of 2015. As you can see in the two charts on the bottom of this page, growth is being impacted by profit improvement actions. In Esurance, policies in force and net written premium both grew by 3.7% in the third quarter of 2015 compared to the prior year quarter. These growth rates are lower than recent quarters, given impact from rate increases, underwriting guideline adjustments, and decreased marketing in select geography to manage risks. Both new business and retention in Esurance have been impacted by these actions.
In Encompass, net written premium declined by 3.5% in the third quarter of 2015 compared to the third quarter of 2014, as a 5.7% decline in policies in force was more than offset by higher average premiums from increased rates and underwriting actions. As with the Allstate brand, we continue to evaluate our results and adjust actions to ensure we're generating appropriate returns in both of these businesses going forward. Now I'll turn it over to Steve, who will cover Allstate Financial investments in capital management.
Thanks, Pat. Slide eight provides an overview of Allstate Financial's results for the third quarter, as highlighted on the top of the slide. Premiums and contract charges increased 5.1% when compared to the third quarter 2014, driven by growth in Allstate Benefits accident and health insurance business, as well as increased traditional life insurance renewal premiums. Operating income for the third quarter was $138 million, 10.4% higher than third quarter of 2014. The increase in operating income compared to the prior quarter was driven primarily by higher returns on performance-based investments, which were partially offset by higher mortality and a lower return from the fixed income portfolio. We reduced the maturity profile of Allstate Financial's investment portfolio in the third quarter by selling longer term fixed income securities that back long-dated immediate annuities.
The proceeds from these sales will be invested over time in higher returning performance-based equity investments to improve the long-term economic results from this block of business. This will include private equity, real estate, infrastructure, timber, and agriculture related investments. Although these investments are expected to deliver higher returns, the timing of those returns is difficult to predict and will result in a greater degree of variability in our earnings. Additionally, we will have to increase the capital allocated to this business. While these sales generated net realized capital gains, investment and operating income will be reduced prospectively by lower yields on the reinvested proceeds. Moving to investments results on slide nine. The portfolio total return shown in the chart on the top left was flat for the quarter.
The consistent earnings from the interest-bearing portfolio were offset by lower valuations driven primarily by wider credit spreads, disproportionately on high-yield bonds and a global equity market sell-off, as you can see reflected in the chart in the upper right. The charts at the bottom provide investment income and portfolio yields for property liability and Allstate Financial portfolios. The property liability yield reflects prior duration shortening and ongoing investment in a low interest rate environment. The Allstate Financial yield is higher and is more stable due to its longer duration and use of its cash flows, primarily to fund annuity reductions. Slide 10 illustrates the strength of our capital position and highlights the excellent cash returns common shareholders received in the quarter and for the first nine months of 2015. We are executing our customer-focused strategy from a position of financial strength and strategic flexibility.
Our deployable holding company assets total $3.1 billion at September 30th, 2015. Book value per common share for the third quarter of 2015 of $47.54 was down slightly from the same quarter a year ago, reflecting primarily reduced net unrealized net capital gains. During the quarter, we returned $920 million in cash to common shareholders through the combination of dividends and common share repurchases. We repurchased 12.9 million common shares for $798 million in the third quarter, and have repurchased 8% of our beginning of year common shares outstanding for $2.2 billion in the first nine months of 2015. Since the beginning of 2011, we have repurchased 32% of year-end 2010 outstanding common shares, representing an $8.2 billion return of capital to common shareholders. As of September 30th, we had $1.1 billion remaining on our current repurchase authorization, which is expected to be completed by July 2016.
Let's open up the call for your questions.
Certainly. Ladies and gentlemen, to give a question at this time, please press star then one on your touch tone telephone. If your question has been answered and you'd like to remove yourself from the queue, please press the pound key. Our first question comes from the line of Ryan Tunis from Credit Suisse. Your question please.
Hey, thanks. Good morning. I guess my first question is just on the expense ratio here in Allstate brand. It was down over a point year-over-year. I think the previous guide was you're trying to get about four tenths of a point from your expense save initiatives. Does that imply that the expense ratio improvement should moderate from the point plus we saw this quarter over the next few? I guess along those lines, if trend continues to be adverse, how much room is there to further cut expenses beyond the four tenths of a point that you've called out last quarter? Thanks.
The number should decline. It won't stay up at 1.4 decline versus prior. There were a couple of things that happened in there. First, we reduced the advertising expenses in the quarter because we just don't want to grow as much, there's no sense advertising if we don't want to grow. Secondly, there were some incentive compensation adjustments, which were adjusted the full year, which has really nine months worth of change in it versus the three months. We can talk about the things we're doing. Matt can mention what we're doing to keep overall expenses in line. It's obviously always a component of focus for us.
Yeah. Hi, it's Matt, Ryan. Thanks for the question. As Tom said, there were some one-time items that we saw in the quarter. There are also some more systemic, longer-term work that we've been doing. In addition to the advertising cuts Tom referred to, we've had a multi-year effort on continuous improvement and process efficiency, which is really starting to take hold now, that's allowed us to get some increased efficiency out of the system. It allowed us to absorb some of the growth without increasing personnel costs, we should expect that to continue long term. Some of the cost reductions had to do with some technology costs that we slowed down some very long-term initiatives that we had, while we retained the focus on the shorter-term ones that were essential to continuing to operate the business.
As Tom said, we won't be able to continue at the current level, we do believe that long term, there are some additional cost reductions and efficiencies that we can continue to drive in the business.
Okay, that's helpful. Then just shifting gears, I guess, on frequency, I figured I'd try. I guess thinking back to last year, I think October was when we first started seeing the elevated frequency trend. I was just wondering if maybe you could comment on how this October sort of compares to when we first started seeing it a year ago.
We couldn't comment on that, Ryan.
Okay. Thanks so much, guys.
Thank you. Our next question comes from the line of Joshua Shanker from Deutsche Bank. Your question please.
Yes, thank you very much. Looking at the monthly numbers, Tom, do you see improvement in terms of the rates that you're getting in? How should we think about 4Q a little bit in terms of trend and balance that with seasonality of 4Q? I know that you're not going to give us guidance for 2016 just yet. I thought, three months ago, you were confident you could probably maintain the 89% ceiling, and I guess you're going to go a little bit above that. How should investors think about positioning themselves going forward?
Okay. You're correct in saying we'll wait until the full year is printed before we do an outlook for next year. As you point out, we had about a year ago, actually, the
Frequency and severity started to take off. As always, we're highly focused on maintaining returns on capital. That said, we always try to do that with looking at the long-term economics, and of course, that's driven by retaining customers, which leads to higher profitability and higher growth. When you have a rapid spike in costs like that, we need to obviously recover those costs quickly. I feel very good about the pace at which Matt and his team are executing, and Don and his business are going at it. I would say this is the first time it's really happened across the country all at once in a long time. That said, we see this movie every year, multiple times in individual states. It's not like we haven't been through this movie multiple times before. We know how to manage it.
It does have an impact on growth, which you can see more dramatically in the Encompass and the Esurance brands than you see in the Allstate brand this quarter. You should expect to continue to see some impact on growth. That said, we're not doing it so aggressively that we're giving up long-term returns. As it relates to the combined ratio and the guidance, obviously after the second quarter, we were at the upper end of the range. We were at 89.1 versus a range of 87-89. We had a range of forecasts at that time, and we said we thought we could be at the high end of the range. We thought we still had a chance to be in the range, and we continue to work aggressively to try to achieve that.
That said, we did indicate we didn't think the low end of the range was achievable. Since then, we've now printed another three months, and we're now still at 89.1. We also have a range of forecasts. When we look at that, we said there are some outcomes where the fourth quarter could be higher so that we could end up the full year at 89 and a half, which is half a point higher than the upper end of the range before. As most of you know, the fourth quarter is historically quite volatile, particularly as it relates to winter weather. We feel comfortable that all of our forecasts are going to leave us below 89 and a half. I feel good about where we're at. I think we're aggressively going after short-term actions. I don't think we're throwing out long-term value creation.
That said, we're getting after it because we know that the biggest driver to shareholder returns is return on equity.
That's great, Tom. Just about the winter weather in the fourth quarter, I've thought about this a few times. December may have some inclement driving conditions, but it seems to me that October and November should be pretty mild driving months. How do you account for the difference maybe between fourth quarter and first quarter seasonality in the loss ratio? I would think that first quarter winter weather would be worse than fourth quarter, or maybe I'm not thinking about this right.
I would say there's a whole bunch of things that relate to that. Obviously, you've zoomed in on weather. In addition to weather, there's the darkness. It gets darker earlier, so more of the traffic is driven in the dark in the fall, and obviously in the first quarter. You have weather, you have precipitation, you have near ice conditions versus ice conditions in the first quarter. There's a whole bunch of moving factors there. I tend to look at both of them and just say they're both highly volatile. Those are busy times of year when people are driving a lot, and you do get more action, so there's more range. I don't think there's any El Niño effect or anything like that that we're factoring in. It's going to be what it's going to be.
Well, thank you for the answers.
Thank you. Our next question comes from the line of Dan Farrell from Piper Jaffray.
Hi, and good morning. Just wanted to talk a little bit more about the shift-
Hey, Dan, we lost you.
Hey, Jonathan, let's go to the next question.
Certainly. Our next question comes from the line of Sarah DeWitt from J.P. Morgan.
Hi, good morning.
Morning, Sarah.
On the frequency trends, has the absolute level of auto loss frequency stabilized versus the second quarter? To what extent are you concerned that it could increase further, and you'll need to take rate increases to another level?
I'll let Matt answer that question.
Yeah. Thanks for the question. Components of it have stabilized, components of it appear to still be volatile. At this point, we are assuming the trend line continues to go up at its current levels. We'd like to see multiple quarters at the same level before we call it as stabilized and take our foot off the rate lever. We continue to operate as if the frequency will continue. Our plans are to continue to take rate wherever justified and indicated and wherever we're able to in those states after we go through our examination. I would say that we haven't leaned back and relaxed yet and said it's all over. It appears to be operating in a narrower range than it was, I'd say, this time last year when we saw such an out of proportion spike in one quarter.
I would add on to that. What you see are our numbers across the country. Matt and I were talking last night about how certain parts of the country have not had as big an increase as other parts. It's possible that while some portions of the country may level out, other portions may increase to catch up to where the rest of the country is. We're, as Matt points out, our team's being very cautious about calling a victory.
Okay, great. Thanks. On the last call, I believe you said you thought you could get back to target auto margins of a 94%-96% combined ratio in mid-to-late 2016. Now that you have one more quarter of frequency data, do you still feel confident in that?
Sure. I remember doing, yes, we can get to the 94-96 because we ran there for 14 years. I don't remember making a call on the quarter. I think the call on the quarter will be dependent on a couple things. The trend you just pointed out, which is we'll have to see where that goes. Secondly, as we increase rates, we're having great success in doing that today. Our competitors are doing it. There's not a lot of noise in the marketplace yet. If frequency continues to go up for two or three years, obviously, there'll be increased scrutiny. We have complete confidence that we can run this business with the margins that it's been run at historically. We don't have a good, clean call as to when we will be at that level.
Great. Thank you.
Thank you. Our next question comes from the line of Alison Williams from Bank of America Merrill Lynch.
Hey, it's Jay Cohen, actually. Just to follow up on the last question, the frequency you saw in the third quarter, is it fair to say it was generally within your expectations? You weren't surprised by what you saw?
Jay, it's Matt. That's a tough question. As Tom just said, on a countrywide basis for the entire system, I would say that's accurate. I don't think there was anything that caught us off guard. There are state-by-state geographic fluctuations in there that do continue to intrigue us, I won't say surprise us. We're trying to figure out why some of the fluctuations exist in certain geographies. Overall, as a system-wide basis, I think you're right. My opinion, third quarter was very similar to what we had seen in second quarter from a frequency standpoint. I had actually expected or hoped that we'd get a little more stabilization in some of the components. It certainly didn't shock us or look to be a spike that was totally unexpected.
It also varies, obviously, by coverage. We were talking about BI, body injury, looks a lot like the level that it started to achieve in the fourth quarter last year. Physical damage is a little bit higher. As Matt points out, the good news is we have great visibility and transparency, and we act to it. The bad news is, like all things in frequency, you can't predict it.
Got it. The second question is, what are you seeing your competitors doing now? Obviously, some have talked about higher claims trends as well, others have not. Are you seeing a general broad competitor response at this point?
What we do is we look at our numbers. We make the changes based on what's happening to our book of business so that we can earn the appropriate returns. Matt, you might want to make a comment about what you're seeing in the marketplace.
As Tom said, we make the call based upon what we are seeing in that market and our rate need and indication. However, in every single case, we also do run recent rate filings from our competitors to get a sense for whether or not we're out of sync or whether or not we should be looking closer at the numbers. What I've been seeing as I look at those is that we have a broad range of our competitors taking very similar rates to us. It obviously varies depending upon their starting point, but we've seen several competitors taking much more in certain geographies and some less in others.
On a countrywide basis, I would say that we do not appear to be out of sync with our competitors, which is why, in addition to the fact that our agents are able to help our customers manage through these premium increases, I think one of the reasons that retention is held and that our quote and close ratios have held is because we're not alone in this rate taking.
Thanks for those answers, guys.
Thank you. We do have Dan Farrell from Piper Jaffray back in the queue.
Thank you very much, guys, and apologize for the issue with the line. My question was with regard to Allstate Financial and some of the changes you're making with the shift in fixed income to risk assets. I was just wondering, what's the lag going to be in getting some of the income from those assets? I realize there'll be more volatility. Secondly, you mentioned some additional capital being allocated. How do you view the ROE impact for that segment with the additional capital? Thank you.
I'll provide some oversight as to the strategy, and Judy can talk about the timing of income. Dan, we have about a $5 billion block of really long-dated liability structured settlements where people are severely injured, buyout annuities immediate annuities that have maturities of 30, 40 years. If you look at the right way to invest behind those liabilities, it would be to, A, make sure you have enough cash that in the next five to seven years, you always got enough money to pay those people. After that, then you want to invest for long-term return. When we look at that second portion of the investment pool, we said, given today's interest rates where they are, we think buying 30-year bonds, to satisfy those liabilities is not the right use of our capital, that we should invest in higher return assets.
They have more volatility on an annual basis. When you look at the volatility on a 10 or 15-year basis, it's actually lower than the volatility in investing in theory, lower risk assets like fixed income. In doing that, though, there is a higher capital charge from the regulators. We've chosen to do the right thing long term to generate capital for our shareholders and economics for our shareholders, even though it requires putting up more capital and having a negative impact on return on equity now because we believe that's the right thing to do. Judy can talk about the shift. What we do is we sold a bunch of long-dated bonds and had some capital gains, and now Judy is going to get that capital reinvested into other investments.
You might want to talk about the reinvestment plan and then how long it will take to get the money back.
Sure. As Tom said, we did sell some longer duration bonds out of the immediate annuity block. Initially, what we're doing is we're reinvesting into primarily public equities and shorter duration fixed income. The long-term goal, as Tom said, is to get it into performance-based assets, which we like because it's a little bit more idiosyncratic, brings some idiosyncratic risk into the portfolio. Plus, we underwrite each one of those investments and bring them into the portfolio, as Tom said, for the long term. The kicker is that it takes a while to get those investments, especially the ones that we really like in this environment. What we're doing, though, as I said, was initially putting it into this other mix, public equities and shorter duration fixed income. We're ready when we do find those higher performing, more idiosyncratic investments.
In terms of timing, we have had some success over the past couple of years finding those investments. What we're finding in this environment is that we're getting money back almost as quickly as we can put it out because the environment has been so favorable for realization. We're keeping our head down and trying to find them, and we'll continue to do that going forward.
Dan, we've made a choice that choose the best long-term economics. If that has a short-term negative impact on operating income, which this will, or return on equity, that's the best choice for our shareholders because we believe in long term generate cash. That's what drives shareholder value. We decide not to make economic decisions based on keeping operating earnings per share up on a quarterly basis. As long as we're transparent, we found our shareholders to be quite receptive to that approach.
Great. Thank you very much for the details.
Thank you. Our next question comes from the line of Bob Glasspiegel from Janney. Your question, please.
Good morning, Allstate. I'm going to push Jay's question, which Matt answered a little bit harder on your upping the range. When you were at Barclays in September, Tom, you said if you saw anything, you would have made an adjustment by now, which I thought was sort of a pretty optimistic comment, given that you had seen two months of the third quarter. Was there something in September or October that caused you to now do it? Or is there something that isn't taking hold as quickly as you'd like? I think all your actions are right, and I have 100% confidence you're going to get the underwriting back. Just want to make sure there wasn't anything in the short-term data that caused a change from the Barclays conference commentary.
When we had the Barclays conference, we have a range of projections, and those projections were largely based off what we had seen through the second quarter. Even though we're a couple of months into the quarter, you don't have things like severity and reserve changes and as clean a shot as you'd like on what the full year will look like. Obviously, you get three quarters of the way through the year, you ought to have a pretty good sense of where the year is coming out. The range has come in some, Bob, is what I would say. It's narrowed down. The frequency continued to be up in the month of September. That also impacted our view. I don't read it as a huge change from where we were before.
It was very clear when we were at 89.1 for six months, we weren't going to get to 87 because we didn't think we were going to run at 85 for the next two quarters after we had run at 89. I was very clear, we're not going to be at the bottom end of the range. We still think we got a shot at the upper end of the range. If I had a clear indication we were going to be above 89 at that point, I would have said so. We didn't have a clear indication. Now we look at it and say we have a pretty clean indication we're going to be below 89.5 Who knows what October, November, and December will bring to us. We're constantly updating severities.
There are some things that are, in terms of severities, that whether it's on the physical damage side, where given the dramatic spike in frequency, it had both a stress on the number of people we have in the system and just the system itself. Body shops are busier, auto parts. We had to kind of sort through what that was doing to cost. Matt and his team are working hard on that. On Bodily Injury, as you know, those are long-dated claims, so you want to make sure you get the reserves right. Given the bump, the volatility in both frequency and severity in Bodily Injury, we've had to look hard at those. We feel good about the forecast. I don't think that you should read it as a huge change.
We provide the outlook to give you a sense for where the business is. As you know extremely well, it's not the only impact to operating EPS. We try to just do it to give you a sense for the business.
Well, I had you at 893, it certainly isn't a big change from my perspective. I just wanted to make sure there wasn't something startling in September that got you off your, "If we had changed it, we would've done it by now" comment. If I could just pivot to life with a quick follow-up. What's the size of the redeployment, and what's the yield loss?
It was roughly $2 billion of long duration fixed income. If you look at the portfolio yield on the portfolio, Bob, it's about 560. These were longer duration assets, you probably should think of them as being a little bit higher yielding than the 560 overall yield. The redeployments into three-year corporates and equities. You're probably looking at a proxy of around two and a quarter until we get the performance-based assets into the ground.
Those are pre-tax numbers, right?
Yes.
Great. Thank you.
Thank you. Our next question comes from the line of Cliff Gallant from Nomura.
Good morning. I had a question about, I think one of your competitors in some states are allowed something called premium trend pricing. I was wondering what your comment is on that as an underwriting tool, particularly in a time like this.
As an underwriting tool, Cliff, or as a pricing tool?
As a pricing tool.
Okay. Sorry.
It's Matt, Cliff. I can say that we've evaluated the use of monthly rating factors in the past. We continue to look at it. In some situations, when you have a slow and steady increase in the cost of insurance, it works to your advantage. In situations like we're in now, where you have spikes in either frequency or severity, it really doesn't do anything you can't do by just taking rate as indicated in those states. We continue to look at it. We also know that not every state would permit it. We believe that most of the states that permit it are file and use states anyway.
If we're on top of our indication, if we're monitoring it as we do on a monthly basis, we should be able to do anything that technique allows you to do and do it more targeted, more segmented, in a more responsive fashion. That kind of puts everything in auto-play mode. Ours is more of a hands-on-the-wheel mode, and that's what we believe is most appropriate in times of volatility and significant changes that we have as we do today.
Cliff, I think, often it's interesting as you watch people analyze our results and other people's results. Everybody's looking for an answer to why.
Yeah.
Like, why yours? Why theirs? I would say this is like one of these complicated thousand-piece puzzles. There's a variety of things that we all do differently, and the thing to look at really is the way in which Matt's talked about the way we run this business in a highly targeted, highly local, highly analytical fashion. We use as many tools as possible to compete effectively and get our returns back. I don't think there is any one silver answer as to, is it geographic spread? Is it monthly rating thing? Is it telematics? There's a whole bunch of things that go into this complicated business. I would just say we're highly focused on making sure we make money.
Okay. If I could be a lot of follow-up, I was curious about the Accident Forgiveness program. Are you contemplating any changes to that? Do you think that that might be contributing to the frequency numbers?
No. I can give that as a short answer.
Yes.
It's just not used to such an extent. It's priced appropriately. We monitor it carefully. That's not a causal factor.
Okay. I'm sorry, I'll ask one more on Esurance. We've seen such a slowdown in the growth rate, and I'm just curious from a how is that slowdown affecting the organization? I know sometimes when you have such a go strong growth company and it slows, it can have an impact on how things work day to day.
Yeah. It's Don. Let me try to address that. I think it was an interesting twist to the question on how it's impacting the company. When we put the two businesses together, obviously the strategy was to use Esurance to address the self-serve market that was still brand sensitive. The business is roughly twice the size it was when we bought it, and we've been running it with an eye towards investing for that growth. We've said consistently, we're running it based on kind of lifetime economic value. That said, the business is about twice the size now. As a larger part of the Allstate portfolio, we do want it to get to the point where that size begins to generate appropriate profitability as well. I feel really good about where they are as a business. They had a great quarter. They're on the right path.
They've done a terrific job of getting, in spite of the drop in marketing, getting the company still to grow. If you look at their combined ratio and loss ratios, they had a really strong trend over the last few quarters. Loss ratio is down pretty dramatically again this quarter. Expenses are down largely due to marketing. I think they've done a good job. Now, having said that, they're still growing. They're still expanding states. They're still expanding products. In the 10-Q, we disclosed we've still got almost 2.5 points of investment in expansion of their business. I don't view this as a business that should grow 3% for the rest of eternity.
I think this was an important inflection point to consolidate the gains we've had from a volume point of view and get the business to the point where operationally and from a loss ratio perspective, it was profitable. We have not stopped investing in that business for growth in any way. If I circle back to then, how does the team feel? I think they feel fabulous. They're proud of the combined ratios they're running. They're proud of the fact that they're still growing in spite of lower marketing and the focus on profitability. All in all, we are delighted with what we've set out to do in the last 3 or 4 years and where they are right now.
I would say we've built a strong team there. Some of the people are those that we acquired, Jonathan Atkinson runs it, who's been there for a long time, run it well. The financial person's been there a long time. We brought in some new claims capabilities because Don pointed out, you grow 2 times in size, you get 2 times as many claim people. Plus, we have more Esurance eyes on Esurance cars as opposed to using third-party. I feel like this is actually a really good time to consolidate our skills and capabilities to go to the next level of growth.
Okay.
Thank you for taking the questions.
Thank you. Our next question comes from the line of Michael Nannizzi from Goldman Sachs.
Thanks. Just wanted to touch on the expense ratio. Can you quantify, Tom, how much of the expense ratio decline was a reduction in advertising?
Mike, that shows up in the Q. There's a table in the Q you can get to the specific number.
Okay.
Mike, I will tell you, I don't think we cut advertising, but we're not doing it in such a fashion that it hurts our long-term brand or market position. We just felt like we shouldn't continue to lean in and advertise and drive quotes when in fact we have tightened underwriting standards.
Got it. Certainly, I mean, the expense ratio improved, the loss ratios is up year-over-year, it's up sequentially. I mean, is that a lever that you feel like you can sort of continue to pull? Is this a manageable level of expenses for the enterprise to run at sort of in the near term?
Well, Matt mentioned what we're working to do. I don't think it'll probably go up some over the course of time. We're really are focused on making sure we do everything possible to get our short-term goals without throwing out our long-term growth or long-term strategy. Whether that's technology, advertising, investing in expanding products like homeowners and Esurance or expanding our benefits business into Canada, we're pushing hard on all of those things. I think you should expect to see us manage between those. Are we always focused on cost? Yes. Did we decide in this quarter because we take our outlooks seriously, and we said, "Let's just not spend the money if we don't need to." We didn't.
Got it. Okay. Thank you.
Thank you. Our next question comes from the line of Paul Newsome from Sandler O'Neill.
Good morning. This is actually meant as a little bit of a softball question. Could you talk about just sort of how the trajectory of improvement would happen for the price increases over time, given the accounting for insurance and the fact that you've got a mix of six and one-month policies? I'm just wondering if maybe some people out there may have got a little ahead of themselves in terms of how quickly the underlying combined ratio improves given the sort of the efforts that you're doing, just sort of the natural lag.
Paul, first, I always think that the questions are thoughtful. There's two lines to look at. If you go back to Pat's slide, there's the blue line and the red line. The blue line is the increase in the earned premiums. Of course, it starts with the rate increases, which we show on the bottom of that slide. Then it goes through to what you actually write. There's usually a little bit of leakage between there because what happens is customers raise their deductibles, which actually improves. It still gives you the same impact as raising price, it just doesn't show up in the blue line, it shows up in the red line, and that the severities go down a little bit. You have it written, and then it takes a while to come through earned.
You can track those two, written and earned, and do some analysis of that and actually determine how it burns through. You could also do the same thing in between the rate taken. You could build a model that would help you on that part. The trickier part of your question is what to do with the red line. The red line is our cost, which is obviously frequency and severity, that bounces around, obviously, by quarter. What we tend to do is smooth it with various portions of time. We look at on latest quarter, we look at latest month, we look at it over six months, 12 months. You'd have to look at the frequency and severity. I think it's unclear yet where, as Matt pointed out, let's just call it the top of frequency is. It's hard to tell.
What we do know is as long as it continues to go up, our blue line will continue to go up. What we tend to do is look at the difference between those, and what we're trying to do is have the earned premium go up faster than the combination of frequency and severity go up. You can see when those lines cross. When they cross, then that's when you get the automatic drop in the combined ratio. I can't give you a specific quarter because none of us really know what will happen with the red line. What we do know, as long as the red line's going up, so is our blue line. And our blue line should go up by more than the red line.
Thank you. Separate question. I personally cover a lot of regional insurers, and they seem to be mostly independent agent channel personal line type businesses there that are similar. They seem to be, frankly, having quite a bit of trouble with a lot of scale issues. Given that sort of environment, given what we've seen with frequency, any thoughts sort of longer term about your independent agent channel, whether or not it needs more scale or not?
It's a good question. First, I would start with the independent agency channel exists because there are some customers who don't really care which insurance company they have, don't necessarily have the highest level of trust in insurance companies, and trust a local advisor to help them select between those companies and give them the right coverage because they don't want to do it themselves. As long as those customers exist, the function and role of the independent agencies will continue to be important. That role, over the last 30 years, has gotten smaller. That used to be a bigger portion of the market, but it kind of leveled out at about 30% of the market or so.
As technology has helped people do more self-serve and that kind of stuff, it could be under a little more pressure, but I think you will continually see people that don't want to do it themselves, and don't necessarily have a trusted brand in mind that they want to pick from. As it relates to scale on the independent agency companies that service that channel, scale does matter. It matters more and more each year, whether that's on data, technology, ability to negotiate lower costs with people. I think you will continue to see a reduction in the share of small carriers, really in all 4 of the channels, which is why we're trying to be in all 4 of those, so we can leverage our scale across all of those.
That said, it is a slow process because many of the companies you're talking about tend to be mutual companies. Over about half of the business is held by mutual companies. They tend to have an extremely high level of capitalization, so they can wait a long time before they have to get out of the business. Maybe we'll take one last question, then we'll wrap up.
Certainly. Our final question comes from the line of Ian Gutterman from Balyasny. Your question, please.
Hi. Thanks. Matt, do you view the ISO Fast Track data as a reasonable proxy for a loss transition?
Reasonable proxy. I'll get in trouble no matter how I answer that. We look at Fast Track data. Ian, it's informative, I think it's directionally correct, I think it confirms for us what we're seeing in the marketplace. Like almost every third-party data source, this is true for all the rate services that are out there, we look at those all the time, there are some data issues with them. There are some reporting issues from certain states. I don't think that they're ever determinative. They're only good indicators.
The reason I asked is the data seems to suggest that there's more of a geographic bias to what's going on than what you guys have suggested the past few quarters. I was curious if you agree with that part of it. Has anything changed in what you've seen over the last quarter to suggest that maybe this isn't as uniform as you thought in the past, that maybe it is a little bit more in certain pockets like California and some other areas?
Yeah. Ian, I don't think I ever said it was uniform. I said it was countrywide, but it varied by geography. Yes, I do agree with that. I think as the Dowling Report, the IBNR Weekly from the middle of October talked about the geographic differences explaining much of the discrepancy in frequency trends across underwriters. I think we do see that. We certainly see pockets, and Tom referred to it on this call already, pockets of geographies where the frequency trends are hitting harder than others. I referred to it on my last call, too. Miles driven is not uniform across the country, and when you look at the Department of Transportation reports, there are geographic differences of the miles driven. Northeast actually not up so much. Other areas of the country up significantly more. Yes, I do think geographic differences explain some of it.
We see geographic differences. When I say it's countrywide, what I'm referring to, it is widespread. It's not as if this is a localized issue where we did something wrong or we failed to implement a rating plan properly in a state. That's why I refer to the countrywide impact. It is everywhere, but it is at different levels in each geography.
Got that.
In a sense, what we're seeing, you see the one movie, which is the movie we report, which has countrywide numbers. We have 50, 100, 200 movies we're watching, and we adjust for each of those movies. Whether it's a particular state, a particular risk class, or a particular product line. We're always adjusting, and that's the way our process and business model works. Thank you all for your questions and insights. I'll try to keep the response respectful of your time. Just know we continue to balance both the short-term and long-term initiatives, so we're creating shareholder value and driving long-term growth. Thank you all, and we'll talk to you next quarter.
Thank you, ladies and gentlemen, for your participation in today's conference. This does conclude the program.