Thank you for joining us for the afternoon program here of the Bank of America U.S. Insurance Conference. We're really pleased here to have Allstate presenting. We have Tom Wilson, the CEO and Chairman of the company. Little known facts that I keep track of, it's Tom's 15th anniversary in the role, if I'm not mistaken, which is going to make him one of the longer-serving CEOs out there. I know Tom may say maybe it's not 15 years. I think it is.
It goes fast, Josh.
It goes fast. He's kind of seen it all. I'm going to give the microphone to Tom for a little bit for some introductory remarks. Then I'll come back to the Q&A. You can always ask me questions by emailing me in the Vericast app that you're watching this in. Let's just give it over to Tom. Thank you very much, Tom, for joining us, and tell us about what's going on at Allstate.
Well, thank you, Josh. I thought what I would do is set some context for the broader group. You obviously know us extremely well. Then we can just jump into questions. Let me, of course, begin. If we go to slide one, this is our Surgeon General warning that we're using forward-looking statements, and we reference some non-GAAP measures. You just need to look at everything we give you. It's in the 10-K, it's on our public documents, it's in our website. Please, as we go forward, feel free to reach out to us. Mark Nogal's on the call here. He can help you get anything you want to help us. If you go to slide two, I'm going to start way up top and then move our way down relatively quickly. This is our strategy.
It's also got some highlights on how we created shareholder value by innovating in both how we execute and create value. We have two parts to our strategy, increase personal property liability market share, then expand our protection solutions, which are the two ovals there on the left. That strategy is working quite well for us. If you look at 2021, you'll see the results. Our property liability market share was up about 1 percentage point, largely through the acquisition of National General, and I'll talk about that in a minute. We also grew Allstate Protection Plans extremely well by expanding its product offerings and our launch at Home Depot this year. Our revenues there were up 20.7% from the prior year.
That reflects that market share growth in our homeowners business, where we're quite strong as well, then Allstate Protection Plans. We also had a great year on investment income. We returned $4.1 billion to shareholders. That reduced common shares outstanding by 7.8% and increased the dividend by 50%. I think the market's starting to move to count on cash more as we move into this higher interest rate environment. That should play well with shareholders. We continue to execute on transformative growth initiative, which is building a low-cost digital insurer with broad distribution. We sold our life and annuities businesses last year for $4.4 billion, then we continue to expand in the protection services business. If you go to slide three, this just shows you the overall numbers.
Property liability premium, it grew to $41.4 billion last year, which is an increase of almost 16%. When it comes to underwriting income, 2021 was really a year of two distinct halves. Josh has followed this closely. In the first half, underwriting income, we benefited from low accident frequency, which reflected the impact of the pandemic. In the second half of the year, frequency started to edge up towards pre-pandemic levels. The cost of claims rapidly accelerated from inflationary impacts. Frequency and severity went up, both of which negatively impacted underwriting income. As you'll see when we get to another slide here, it was a tale really, of making a lot of money and then not making money in the second half. That decline was offset. We had a spectacular year in our performance-based investments.
As a result, adjusted net income was $4 billion, which was below the prior year because of the pandemic had a much bigger impact in 2020. Still had a return on equity of just below 17%, which is at the high end of our target range. Our net income was $1.5 billion, and that reflects that we had a $4.1 billion loss on the sale of our life and annuity businesses. While that loss did reduce net income, the proceeds partially funded our current $5 billion share repurchase program, which contributed to that $4.1 billion in cash to shareholders. Last year, we repurchased 8.7% of our common shares outstanding, and that was up from 5.5% in the prior year.
If you go to slide four, what I'd like to do is discuss auto insurance profitability. I'm sure Josh will have questions on this. I thought I would start to set context with three questions. How good is Allstate? What happened in 2021? When will it improve? If you start with the first question, which is on slide four here, we have a history of generating really industry-leading returns in auto and home insurance. That's the result of excellent operating capabilities and continuing to innovate in the risk and return space. We rigorously allocate our capital, whether that's to geography, risk type, different businesses, and that in auto insurance, we target a mid-nineties combined ratio with the capital that we give to it. In homeowners insurance, we target a mid-sixties underlying combined ratio, and that excludes catastrophe losses.
That's because we need to generate enough underwriting income to offset and manage the loss costs associated with the cats. Our performance is among the best in the industry. You can see that in the graphs at the bottom, where a lower line leads to a higher return. In auto insurance on the left, you can see Allstate, Progressive, and GEICO tend to lead the industry. Our auto insurance combined ratio, represented by the blue line, has remained favorable to the industry by just under seven points from 2017 to 2020. On the right-hand side, you can see homeowners insurance, where we consistently and meaningfully outperform the industry and our peers. That includes State Farm, which is shown in red, and Progressive, which is shown in orange, which both have average higher combined ratios than us over the last five years.
From 2017 to 2020, our homeowners insurance business generated $2.9 billion of underwriting income. Remember, in that business, you don't get a lot of investment income. The industry generated an $18 billion underwriting loss. If you bring in 2020 results, we've averaged about $667 million over the last five years in underwriting income. If you go to slide five, let's discuss what happened to auto insurance then in the last half of 2021. What are you going to do to get back to our target of mid-90s? That pandemic has created a really volatile environment in auto insurance, and it's required us to adapt pretty quickly. Early on, you had the stay at homeowners, you had reduced accident frequency, that led to really strong profitability in 2020 and the first half of 2021.
Loss costs began to accelerate, in the beginning of 2021, really took off as you moved into the second and third quarters of 2021. What happened was, a lot of it was used car prices go up. Used car prices go up, we pay more when a car gets totaled, and it increases the cost of repairing damage. It's like setting your limits go up, but you don't get any more money for it. The cost of settling bodily injury claims, and that's where our customers cause bodily injury to other people, increased. There were a lot more severe accidents, higher medical costs, and increased attorney representation. As you can see from the chart on the bottom left, we generated underwriting insurance income of $1.7 billion during the first half of the year. That's the first story.
Second story, we had a loss of $459 million in the second half of the year. Obviously, on a whole year, it looked okay, but we don't really think about it that way. Second half was not up to our standards. We have a comprehensive and aggressive plan to get back to target levels. That includes raising prices, managing loss costs, and reducing expenses. The chart on the bottom right shows the annualized implemented rate increases in Allstate brand auto insurance over the past three years. You can see in the fourth quarter of 2021, far right-hand side there, we implemented Allstate brand auto rate increases at an annualized rate of $702 million in the fourth quarter.
It's important to remember that those are six-month policies, so that's an annualized premium, and it takes a while for it all to roll in, and I'll give you an example of this on the next slide. In addition to that, we're really looking hard at our claim operating practices to make sure we pay the right amount at the right time in the right way, and that includes strategic partnerships with parts suppliers, repair facilities. What we're really trying to do is mitigate the cost of repair using advanced claim analytics to help us get better in terms of managing both the cost to fix a car and to settle any outstanding liability claims. We're also reducing operating expenses to power transformative growth. Slide six addresses a question of when we will be back to target levels.
I know Josh probably going to ask me for a specific date, because everybody wants to know when the time is. This just shows the lag between implementing price increases, earning premium, and the potential for loss costs to continue to increase. The timing of when your operational actions restore auto is largely dependent on the relative growth of premiums compared to loss costs. Average premium increases show there's a lag because you sell a six-month policy. As a result, it takes six months before all the policies are sold at the higher price, and it takes you 12 months before they're all fully earning at that rate. On the cost side, that severity's increase will be dependent on macroeconomic factors.
Frequency levels are unlikely to keep increasing towards pre-pandemic levels, and we're thinking about that when we're implementing our price increases like we did in the fourth quarter. Although it's possible a portion of that frequency benefit may persist due to the shift in the time of day that the miles are driven. In addition, we obviously need to provide offsets by reducing loss cost pressures, and that's doing the expense reductions. The chart on the bottom of the slide shows our historical trend compared to our earned premium trend. When the blue line is above the red line, that means the combined ratio is under 100. Historically, you see there's a gap between those two lines, and that's the mid-90s target. You can see in 2020, 2021, the big gap, big difference there. That was the pandemic.
That was lack of driving due to the shelter-in-place orders, despite the fact that we gave our customers back $1 billion. In the second half of 2021, you can see they contract. The dashed lines on the right side get to this question of when. It's an illustrative view that shows over time, we expect the lines to return to their historical relationship given the things we're doing to manage the overall margins. Of course, it's the combination, and if it deviates from our expectations that are embedded in our rate filings, we have to delay stuff, we may accelerate stuff, but our goal is to get to our target mid-90s. The darker dashed red line shows a scenario in which the average underlying loss and expenses, while continuing still to increase, kind of begin to flatten out consistent with what's happened historically.
In that illustrative scenario, margins are compressed for a period of time until the accelerated rate increases turn into average earned premiums and restore the margins. The lighter dashed line shows a scenario in which the average underlying loss and expenses go up higher at a faster rate. In that scenario, obviously, you have to pursue higher rate increases to restore margins. However, it takes a little longer to get to the target levels. Of course, those are just examples. I can't give you a timeline on when we will get there, but we did try to give you some metrics, some thumbnail stuff to help you get there. To the extent the written premium growth exceeds loss costs by about 1% for a year, then the improvement on the combined ratio is approximately 0.4 points in year one with an additional 0.5 points for the full year.
Of course, we're doing it by quarter, so we don't really think about this, but we think about it by month really. The simplified example shows that there's a lag impact, right, of when you get that earned premium. Also that you have some variable expenses like commissions and stuff that come out of it. It doesn't factor in things like raising deductibles or limits, which also influence it. It's not a really precise way, but it's a way to give you sort of a thumbnail approach to, are you making progress on what you're trying to do? As you know, what we did is we're starting this year to disclose our rate actions every month. That'll give you and Josh and others a chance to say how you're doing.
I want to shift gears and quickly talk about transformative growth because it's a really important and comprehensive rebuilding of the property liability business, and it's really about creating a flywheel of growth. I think sometimes the focus on auto insurance margins takes away from that bigger and broader story, which has a potential to substantially increase our growth. It's really a multi-year initiative. It's across the whole business, so it's about improving customer value, expanding customer access, increasing the sophistication and investment in customer acquisition, deploying new technology ecosystems, and then enhancing our organizational capabilities. We've made a lot of progress around this. I want to show you the flywheel. If you start at the top there, we reduce expenses. That improves customer value. That enables us to offer a lower, more competitive price while maintaining attractive returns.
That should increase close rates and drive better retention. With that lower price, we need to increase quotes because you want to sell more of it. We accomplish that by both enhancing and expanding distribution and increasing our marketing sophistication and investment. In both of those, we're also trying to reduce expenses on. That leads to higher quotes at lower costs. The new technology platforms further lower costs and enables improved service with greater speed to market, particularly in claims. You see we're doing a lot of virtual stuff in claims. That new technology also gives us the ability to create competitively differentiated products that are affordable, simple, and connected. That increases our pricing power and our customer retention. That's then a flywheel. It just keeps going and drives more and more growth.
That's how we're going to get to increased market share. If you go to slide eight, you can see how transformative growth impacts all the aspects of the customer value chain. We're making a lot of progress on each of these. I won't go through each of them. You can read them faster than I can talk about them. It's just the point is a lot of really substantive things, each of which in and of their own right are a lot of change in an organization. When you put them all together, it's complicated, but it's really, really powerful. As you can see on the bottom of the page, you can see we have 5 phases of transformative growth, and we're making really meaningful progress on it. We're kind of in phase 3 right now and sort of starting to creep into phase 4.
We're starting to see the benefits. The underlying assumptions that we're doing are proving to be true. Go to slide nine. I want to just touch on National General since we did spend $4 billion on it this year. It's been a really successful first year of ownership. We became a top five independent agent carrier. We grew premiums $4.7 billion. We added 4.2 million policies in force. We integrated the people and the processes. We're taking our Encompass IA business and merging it into National General, so it's sort of like a reverse acquisition. The Allstate Health Solutions added over $1 billion of premium to that segment. Arity also expanded its business model in the telematics space by including LeadCloud and Transferly. It was really three really good strategic additions by buying National General.
We're on pace to get our expense synergies out. We also bought SafeAuto late in the year, which further leveraged the platform. This year, we've got three drivers of growth. Product expansion, we're going to sell standard auto and home to the non-standard relationships that they have with independent agents. We're going to sell it to more independent agents and continue to work on our integration. If you move to slide 10, I want to just touch on Allstate Protection Plans. This was really a great buy. We bought it in 2017 for $1.4 billion. It's now about 6x its size in written premium. It had $1.2 billion of revenue last year, which is a 40% compound annual growth rate since the acquisition. We made $142 million last year.
It's really been a terrific addition to the lower oval, which is also a way we create value, which is leveraging our brand, our capabilities to expand the protection we offer to customers. Let me just close on slide 11. It's about why are we an attractive value for shareholders. Obviously, we've talked about sustainable growth. We're increasing our market share and property liability. We're expanding our protection offerings. We're not talking about it, we're doing it. We returned $4.1 billion to shareholders. We have a diversified portfolio of businesses building this low-cost digital insurer transformative growth, expanding in the protection services segment. When you look at value, it's a lot lower than peers in the overall market, despite our growth potential returns, cash to shareholders, and the life and annuity divestiture. Look at the table on the bottom.
In the first three columns, you see performance metrics for Allstate at the top there. Then our P&C peers, S&P life index financials, and S&P 500. You can see we outperformed on cash returns and EPS growth over the past five years. The trailing 12-month price earnings ratio is significantly below anybody on that chart. If you look in the right columns, it shows you what's the magnitude of that rather than just some P/E number. You can see at the current share price, Allstate's trading at 72% of our P&C peers. The price is 88% of the life index, which is lower than the P&C P/E ratio, despite the fact that we're really not have much in the life and annuity businesses anymore. The valuation as a percent of the financials or S&P is even lower.
With that context, Josh, where would you like to go?
Let's talk about transformative growth. Everyone's going to want to talk about rate. Maybe I'll get there, but let's talk about transformative growth instead. Even though Allstate has said many times, we want to sell you insurance products the way you want to buy them, and we're agnostic about how you might want to purchase them, but still, for the most part, you guys are a home and auto company, but with more auto than home, and the vast majority of the business is sold through the captive agency market. If we're looking out five years, what is the new business being sold through captive agency, through Allstate's other channels? Over time, that new to legacy business, how much does transformative growth change the footprint of how Allstate's business looks overall?
That's a good question. I'm going to go back, Josh, to go forward. You're right, we have been talking about multi-channel distribution for a long time. We've done well on it. I think about 4 years ago, I said there's a couple of things we need to do differently. First, we had the Esurance brand in the direct channel, and we weren't taking market share. We were growing all right, but we weren't taking market share. I said, "We need the Allstate brand on that to really grow it." You saw we've done that. It's now, depending on how you want to measure it, 25%-30% of our new business. You would expect over time, as new business turns into renewal business, that the direct channel would be at least 25%-30% of our overall volumes in the company.
At the same time, when we looked at our growth in the direct channel and the agent channel, we came to two conclusions. One is some people want to buy through local people. They just don't want to pay too much for it. We've been cutting our expenses, and we're cutting the costs of the Allstate agents, which if I want to get to, I'll talk about it does create some transition there, so that we offer a better price. We've been reducing and improving our competitive price position, which should drive growth in that local agent channel as well. In the independent agent business, we've been in that business for 20 years. We kind of make some money, then lose money, but we never really grew.
I went to Barry Karfunkel, who was running National General at the time, and I said, "Look, I haven't been able to make this channel work. It's kind of ridiculous that we can't be bigger in the independent agent channel in auto and home insurance," just because to your point, we're good at it. I said, "I either have to decide I'm going to try to fix it again, or I should get rid of it." I decided I'm going to get rid of it, but I'm going to get rid of it, but I'd like to buy you first, and I'd like to give it to you. You and your team have shown you've acquired like 20 companies. You know how to do this stuff. You join us, we'll give you Encompass, and we'll grow in the independent agent channel.
I think all three of those should grow. I think the direct piece will grow because we got the Allstate brand and we're putting more money behind it. All businesses should grow because we've got a better competitive position. I think the Allstate agent business will grow, probably not as fast as the direct business, but I wouldn't count out customers who want to have somebody in a local area buy from them.
If we think about that agency footprint, you've been changing your commission for renewal and new business, it just caused some attrition on the agency side. People might be retiring a little bit earlier. There's the great resignation going on simultaneously, people are retiring earlier. For a while, you paused the appointment of new agents. Have you resumed appointing new agents? How much natural attrition is there annually for agents retiring? When we think about the size of Allstate agent footprint in 2020 or 2021, take your pick, what does it look like in 5 or 10 years?
There's a lot there. The Allstate agents right now, we have a little over 10,000 agents selling for us. There's probably 35,000 people in total selling for us every day through those agencies, and they sell only our stuff. You're right, we've reduced the amount of money they will make on renewals, but we've increased the amount they'll make on new business. To make the same amount of money, they're going to have to sell more. Some people don't want to do that. Most of them want to do it, so a substantial more than a majority. Some are totally into it, and they're charging ahead. If you look at our growth last year, we sold about as much through the existing agents, actually, a little more than the prior year. They're still engaged.
We did stop appointing new people with that business model because it just wasn't economic. We'll always do what's right for cash and what's economics, as opposed to just put growth on the books and not make money on it. We did it even though we didn't have a new model for that Allstate agent local thing set up yet. We've tried 2. We've landed on 1. It's called MSA. They're basically employee agents that either work remotely or live in a joint office. We're starting to expand those, and we'll put as many of those in the market as we can. We still have 90% of the market I'd like to get. I think they'll both go up, but you won't see the existing neighborhood office with 3 people in it grow as much as you will the MSAs.
The MSAs are where we're going to focus growth of Allstate agents.
Let's get on to everyone's favorite topic, the rate story. I don't know what the numbers need to be. The question everyone said to me, or at least they were starting to say it six months ago, is regulators are not going to give Allstate rate increases because you made too much money in 2020. Two, maybe more sophisticated, they said, "Well, it takes a year of book closing before an auto insurer can send data to regulators for them to see. Even if Allstate's losing money right now and the regulators want to give them rate, they're going to have to wait till the 2021 books close, and there's going to be a delay." There's all sorts of reasons why Allstate's not going to be able to get rate.
Allstate announced that every month you're going to show people just how you're getting rate. Is there any truth to the arguments from the skeptics that says the regulators don't want to give you rate?
Well, there's always truth that not everybody wants to do what you want to do. Let me go way up and come down quickly. I once got asked by a shareholder, "What's your strongest muscle?" I'm like, "I don't know. What do you mean by muscle? It's kind of a weird question." They kind of explained to me, and I said, "Our strongest muscle, making money. We know how to make money." Well, we're working on growth, and you can see that over time. I think transformative growth really is the breakthrough for us, which it's all about price. If I get your price right, you will grow. That's what we're working on. We've always been good at it. When you look at what happened in the pandemic, within 10 days of the shutdown in the country, we were giving customers back $1 billion.
We gave $600 million the first pass, another $400 million later. We went proactively to regulators. We called them on Sunday night and said, "Hey, this is a big deal. We're not going to wait for a year to do filings, to collect the numbers for you to come beat us up. We're going to do this because it's the right thing to do." You build up credibility with regulators when you're proactive in doing things that is in your customer's best interest. At this time, it's not a yearly thing for us. Sometimes we file a rate increase, and we don't get everything we want. We immediately file another one shortly thereafter. We're not waiting for annual cycles. The difference here is so much of this, Josh, is in the physical damage coverages, and those settle in 90 days.
It's not like, geez, I think this is going to cost me a lot of money in 3 years, and the regulators. You're like, "Hey, we paid the cash out. See, here it is." You always want to work collaboratively with a regulator. We have a good relationship with them. We give them lots of data. We try to work on transparency. We got $700 million in Q4. We needed $700. We need more. We think we're going to get it from the regulators. Will some states push back? Of course. Will some states make us do it in two bites? Yes. We're confident that we're going to get back to the level we've always operated at, which is better than the industry. When you're low cost, the regulators want to help you win.
Changing gears a little bit. I think Allstate went public in 1993 after Hurricane Andrew. Sears didn't want to recapitalize the company, or maybe not. The public equity markets did. In 2005, you guys lost $4 billion in Hurricane Katrina. In 2009 through 2014, Allstate changed its footprint dramatically, moving away from the coast. Really, the catastrophe volatility at Allstate is lower today than it's ever been, which is a huge source of your profitability. In my mind, it's created two things. One is that Allstate isn't really exposed in places where there is a lot of cat risk, and maybe doesn't want that cat risk. Two, Allstate has a lot of concentration exposure in places that aren't cat prone. Adding more places gives you more concentration on the same streets and everything like that. We look at Allstate's plans to grow.
Can Allstate grow in the same markets where it already has dominating market share? Does Allstate want to increase its catastrophe footprint and go back to places where it moved away from? To what extent does growth require a changing view of catastrophes from what's been a 25-year journey for the company?
Let me start with, I think homeowners is an undervalued business by the market. When you look at our returns, I mean, it's spot on. Man, we make a lot of money in homeowners, and we're good at it. We make a lot more money than everybody else. To your point, some of it was getting smaller in cat markets. Some of it was redoing the way our products work. We have a product called House & Home, where we age rate the roof. If you got a 20-year-old roof, we don't buy you a brand new one because a hailstorm kind of rattled through. Not everybody's in that place. That keeps our costs down. That is fair and appropriate. There's. We got a massive reinsurance program. We have all kinds of different pricing than other people. We're really good at it.
We've got to a place where we're in good position. The weather, of course, has continued to get worse. Some of it is along the coast, as you point out, like hurricanes. Now with wildfires in California, we shut down homeowners in California over a decade ago, and our share is a lot lower there because we wouldn't write homes if there was bushes within 100 feet of it. That turned out to be right, but we're not perfect, we still lost a bunch of money in Paradise. We are in the Camp Fire. We're working on how do we expand it. I think we have expanded in some cat-prone markets, we have what I'm saying would be an infrastructure, whether that's increasing our reinsurance, changing our product, doing our pricing, we can increase.
I think the bigger growth for homeowners will probably come through the independent agents. If you look at those numbers on the rest of the industry, some of our big competitors, you can't lose $0.15 a dollar on homeowners, get no investment income, stay in business very long. It's a capital hog, the people that are doing that are going to have to fix it. You cannot grow your way out of that problem. We'll be right there with good prices, good process, good reinsurance, I think the independent agent channel is going to be a huge growth area. Not so much in direct, Josh, because for whatever reason, we haven't been able to sell as much through direct, I think that's consistent with our competitors, too.
We're running out of time, one question, a little bit of a curveball, I suppose, those things you don't get to ask. I think that you guys don't get enough credit for how ahead of the curve you are on telematics, you're certainly one of the industry leaders. Back about a year ago, Apple launched iOS 14.1 with a lot of security features. What did that mean for Arity and Drivewise? Is there a risk that the edge that Allstate created has been disintermediated by Apple and maybe later Google and whatnot?
I would say it's something we're dealing with. Arity, in and of it, I think it's so valuable in the company. In and of itself, it's a unicorn. That's it. We track 26 million cars a day, every 15 seconds, 700 billion miles. We've got that one. We have expanded it into marketing services. Rather than just we'll collect, we'll track cars for you, we've got a pricing service we're using, helping other people price. People should price by telematics. That's the way to price auto insurance. I think Arity has basically helped us do what we need to do in our auto insurance business, it provides this opportunity to create more shareholder value outside of the insurance business by doing that.
The iOS stuff changes some of it, but we have so much data now that we're working to help auto insurers be able to Arity IQ will enable auto insurers to price a company, a customer, individual customer using telematics, without having to connect to their car. Boy, that's huge. Value creation, lowers the cost of acquisition, lets people target their risk a lot better. While iOS, it's influenced it a little bit, but it's not going to shut us down.
Well, I could talk to you for a lot longer, but I don't have the time. I want to thank you. I see Mark Nogal online. I want to thank him, too, for giving us your time today. We'll be in touch, and I hope you have a great afternoon with some investors, and thank you.
Well, thank you, Josh. We always value your insights. See you soon.