From Torchmark, we've got Gary Coleman, Larry Hutchison, Co-CEOs, and Frank Svoboda, who is the CFO. This team, well, Gary and Larry, have held the CEO office for the past 5 years. Prior to their current roles, Gary served as CFO. Larry served as general counsel. Each of them has 25 years of experience with the company. Actually, more than 25 years. Torchmark's been one of the most consistent companies in terms of earnings, capital distribution, book value growth in the life sector, and I am looking forward to hearing more about it. Guys?
Thanks, Jay. Before we get started talking about Torchmark, I've been advised that I need to read the following. Our presentation today contains forward-looking information and certain non-GAAP measures. Please refer to Torchmark's cautionary statement regarding forward-looking statements and reconciliations contained in the company's 2016 annual report and subsequent Forms 10-Q. Okay, let's talk about Torchmark. The first chart that we have shows Torchmark's total shareholder return as compared to the S&P 500 and the S&P 500 Life & Health Insurance indices. As you can see, Torchmark's the dark blue line, excuse me, we have outperformed those indices over the past 5 years. That outperformance is due in large part to the execution of a business model that we think is quite different than many of the other life insurance companies. The next slide shows you the key components of that model.
Just in summary, Torchmark serves the middle-income market. We sell basic life protection products. We sell them through controlled distribution. Those products, through cost control, are very profitable and generate a great deal of cash flow. Not only enough cash flow to fund future operations, but each year we have excess cash flow, which we endeavor to return to our shareholders. First, let's talk about our market. Industry data shows that 39 million Americans are uninsured and another 34 million are under-insured. The middle-income market that we're in has significant growth potential, but very little competition, and it's the market that Torchmark has focused on since its inception. The products we sell, we focus on basic whole life, term life, and supplemental health products, and these are the products that best meet the needs of the middle-income market. These products have advantage.
They're very simple to understand, both for the agents and the customers. An added benefit to the company is that because these are basic protection products, they produce stable policy liabilities and DAC because they're not subject to fluctuations in interest rates for the equity markets. To sell the products, we use controlled distribution, our exclusive agencies, and Direct Response. This enables us to control cost and to limit competitive pressure. Our agents sell exclusively for us, and they're trained to sell in the middle-income market. Also, our Direct Response operation has over 50 years of experience in that market. The next component of our model is the effective cost control and profitability of our products. Torchmark has long been known for controlling both the cost of acquiring business and administering it.
In our market, we sell low face amount policies, which means a lower annual premium per policy than most life insurers collect. As a result, to be profitable, we have to control the acquisition cost and the administration cost. Due to the cost control, we generate healthy underwriting margins. As a result, we don't have to rely primarily on investment income to generate operating income. As a result of those underwriting margins that we have in the business, the profitability of our products leads to consistent annual cash flows as due to the long-term revenue stream produced by the block. For example, each year, around 90% of our premium revenue comes from policies that were sold in previous years.
Contributing to the long tail of the revenue is the persistency of our business which, due to the products we sell, the market we sell in, the persistency has been very consistent throughout the years, regardless of the macroeconomic conditions. As I mentioned, we generate a great deal of cash for these products. Our first priority for that cash is to use it to fund current operations, but each year we generate $300 million+ excess cash, and Torchmark has long been committed to returning that excess cash to the shareholders. Since 1986, Torchmark has returned over 75% of its net income to shareholders, either through share repurchases or dividends. That's a quick summary of our business model, and while it is very straightforward, the execution of that model is critical to our success.
To continue to be successful, the model stays the same, but the methods by which we execute the model have to change. Like many companies, we're investing in technology. It's critical for us in this marketplace to be able to reach the customers and our agents in a manner that works for them. Also to be nimble and be able to adapt to the changing environment. As such, we're investing in technology that facilitates the digital consumer experience, that helps us use technology to improve the agent experience and make it easier for the agent to do business. We're also replacing the old back office legacy systems and expanding our data analytic capabilities. We think these investments will provide the enhancements we need to continue to be successful with our model. Next, why don't we take a brief look at Torchmark's financial results?
The current slide shows the components of our net operating income. For us, net operating income includes underwriting income, which underwriting income is the premiums less the acquisition costs, less the administrative costs, as well as the policy benefits. Next component, excess investment income, is our investment income less the interest required on the reserves and interest expense. Then we have, of course, income taxes and other expenses. As you can see, our underwriting income is the biggest component of our income, and it really accounts for over 70% of our net operating income. The products that we sell and the long-term nature of them lead to very consistent earnings. In this next chart, we look at the operating earnings of Torchmark over the last 10 years.
During that time period, you can see that we have grown earnings each of the years, and we even grew earnings in the difficult years of 2008 and 2009. Net operating earnings per share has increased at a compound growth rate of about 8.3% over the last 10 years. To me, that's remarkable because during that time period, our yield on our investment portfolios declined each year due to the low interest rate environment. We started out 10 years ago with a portfolio yielding 7%, and it's declined to the current 5.6%. We also show the guidance for 2018. You can see it's considerably higher, and that's due to tax reform. Frank will take some time in a little bit to go over the tax reform and what the impact is on Torchmark.
The next slide shows our GAAP net income, and the main thing here is to show that even our net income follows the same trend as our net operating income. It's growing over the 10-year period. You see a big spike in 2017. That's also due to tax reform. Let's look at the book value as well. I think this chart further demonstrates the consistent growth at Torchmark. In this chart, we're looking at book value excluding the net unrealized gains and losses on securities. Net book value per share has grown at an annual compound growth rate of 10% over the last 10 years. Once again, we have a spike in 2017 due to the impact of the tax reform. Even if you exclude that, the 10-year CAGR for book value has been a healthy 9%.
Finally, we look at the book value per share on a GAAP basis. It shows some fluctuations because it includes the unrealized gains and losses. As you can see, it also shows an increasing trend over time. That's a brief look at our financial results and what comprises them. Now I'd like to turn things over to Larry for a discussion of the operations that lead to that profitability.
Thank you, Gary, and good morning, everybody. I'd like to first start by looking again at the components for net operating income. As you can see, we generated $625 million in underwriting income in 2017. Let's look at the components of our net underwriting income. Most of our underwriting income comes from life insurance. Life insurance is Torchmark's primary focus because there's less competition in this market. It generates investment income. It's less subject to regulatory issues. The chart at the right shows our distribution among underwriting among our three primary life channels. American Income is the leading producer of life insurance underwriting margin, followed by our direct response operation and Liberty National Life Insurance Company. Let's next move to our premium by distribution channel. As you can see, American Income and Direct Response are the biggest life premium producers. United American is our largest health premium producer.
American Income, Liberty National, and Family Heritage are exclusive agencies made up of independent contractors who sell only for us. Direct Response reaches the market through direct mail, insert media, and electronic media channels. United American markets through independent agents and brokers who can also sell for other carriers. Let's next look at each of our distribution channels in a little more detail. American Income is our largest producer of life insurance. It's unique in that its home office employees and all of its agents in the U.S. and Canada are union members. This provides a competitive advantage for American Income. The union relationship has been the backbone of this business over the years. Historically, these union relationships provided union leads to our agency force. However, in the last 10 years, American Income has diversified.
While the union relationships are still very important, a majority of our business today is produced through non-union leads and referrals. American Income has also had a very strong sales growth over the past 10 years. We've grown at an annual compound growth rate of 9.3%. The correlation to agency growth and sales is shown on the next slide. As you can see, American Income's agent count has grown at a compound annual rate of 10.5% over the same period. Our current focus is expanding our producing agent count. In this underserved market, we know we can expand well beyond our current producing agent count. Our second-largest exclusive distribution is Liberty National Life Insurance Company. Liberty National is our oldest company, founded in 1900 as a debit company. Liberty suffered years of decline, and they began a turnaround approximately 10 years ago.
As we converted Liberty National from a fixed cost to a variable cost model, and the turnaround began slowly, but you can see we're gaining traction. Sales have grown at a 7.3% annual compound growth rate since 2012, and we're finally seeing life premium growth. In 2017, we grew life premiums at Liberty National by 2%. Let's again look at our agent count. As you will see that agency sales growth is driven by agent count growth. Liberty National's agent count has grown at a compound annual growth rate of 8.2% since 2012. We're optimistic about the future of Liberty National. Liberty National continue to grow as it expands its current agencies and expands geographically out of its traditional four-state region in the Southeast. Let's next look at our third exclusive agency, Family Heritage Life.
Torchmark acquired Family Heritage in 2012, and Family Heritage sells limited benefit health insurance policies in non-urban areas. Family Heritage has had good sales growth since 2013. Sales have grown at an annual compound rate of 6.7%. Again, let's look at Family Heritage's agent count. You'll see the direct correlation between agent count growth and sales growth. The agent growth has grown at an annual compound rate of 8.9% since we purchased Family Heritage. Family Heritage will continue to grow as it increases the size of its agencies. In addition to our three exclusive agencies, Torchmark has a direct response operation, Globe Life. Our direct response has a history of strong premium growth and sales growth. Sales have declined in the past two years as higher-than-expected claims emerged in 2015 from certain blocks of business.
We made operational changes to maximize margin dollars that by design decreased sales in 2016 and 2017. Sales will grow in the future through our use of analytics and innovation. direct response also contributes to our three exclusive agencies. It provides analytical support and is a source of leads for the three exclusive agencies. Our last distribution is through United American Insurance Company. This is the one distribution where we use independent agents and brokers who can sell for other companies. Torchmark Corporation considers Medicare supplement an opportunistic market. It's very competitive, the cost of entry is low, and it's highly regulated. This makes sales difficult to predict, but in the next slide, you can see we have a stable in-force block of business. In summary, we're very pleased with the position of each of our distribution channels, and we're optimistic about the growth potential for each of these channels.
Frank will now discuss investments, capital management, and tax reform.
Thanks, Larry. Let's once again look at the components of our net operating income. As noted here, we look at the investment function on a standalone basis, as we really consider that to be a better measure of the overall investment performance. Let's take a look at our excess investment income and its components. As Gary indicated earlier, our excess investment income as we define it, is really our net investment income, less the required interest on our net policy liabilities and less the interest that we pay on our debt. The required interest on our policy liabilities is primarily interest assumed in discounting our GAAP reserves and our DAC asset, not interest credited to policies. Let's take a look at our invested assets. At Torchmark Corporation, we follow a conservative investment philosophy and invest primarily in long-dated fixed maturities, which comprise approximately 95% of our invested assets.
We focus on these types of investments because they are a better match for the long, fixed-rate liabilities that are generated by our basic protection products. As seen on this slide, the invested assets have grown from about $10 billion at the end of 2007 to just under $16 billion today, even though we had spent around $3.6 billion over that same period of time to repurchase shares. Just a couple of minutes on our portfolio yield. This chart demonstrates the challenge we have had due to the lower interest rate environment. The downward trend negatively impacts net investment income and therefore our excess investment income. Despite the lower rates, we have grown net investment income in each of the last 10 years, but at a lower growth rate than the percentage growth in our invested assets.
Ultimately, we will benefit from higher interest rates and look forward to them in the form of higher investment income. We're really not concerned about unrealized losses in our fixed maturity portfolio, as we have the intent and, more importantly, the ability to hold those investments to maturity. We have been able to thrive in the current low interest rate environment and believe we can continue to do so as our policy liabilities are not sensitive to interest rate or equity market fluctuations. I'd like to move on to capital management. Our first and foremost priority is to fund our insurance operations and invest in their future growth. Secondarily, our goal is to maintain our capital at the levels necessary to meet regulator and rating agency expectations. We are able to meet these goals each year and still distribute excess capital to the parent company.
As this chart illustrates, it shows the components of our excess cash flow at the parent company. Our excess cash flow is the cash left from dividends received from our insurance subsidiaries, less dividends paid to our shareholders and less interest paid on our debt. The next chart shows the significant excess cash flows that we generate year in and year out at Torchmark. The declines in the past few years have been due primarily to a drain on statutory income and surplus from strong growth in our life businesses, investments in technology and information security, and the elimination of our Part D operations. We always strive to utilize our excess cash as efficiently as possible, and as previously indicated, our first priority is to fund our operations. We evaluate alternative uses.
In the past several years, share buybacks have been the primary use of those excess funds as they have provided the best risk-adjusted return versus other alternatives. This slide really does give a history of some of the repurchases over the past 10 years. As you can see, we have spent over $300 million in repurchases every year since 2011. As we look at the next slide, it really demonstrates the commitment of Torchmark's management and its board of directors to the strategy of returning excess capital to our shareholders. As indicated on the slide, we have returned over 80% of our net income over the past 10 years, excluding the impact of tax reform on our 2017 net income. Share repurchases have been a part of our business model, though much longer than the 10 years depicted on this slide.
In fact, we began our repurchase program in 1986 and have repurchased Torchmark shares every year since, excluding 1994 when we purchased American Income Life. In the future, even though we intend to return all our excess capital to shareholders, absent some more favorable alternatives, this ratio will likely be lower than the historical ratios given that the tax reform is going to have a greater impact on our GAAP income than it will on our excess cash flows. I'd like to discuss the impact of tax reform on Torchmark. At a very high level, the tax reform will provide to Torchmark substantial long-term benefits because our future profits will be taxed at that lower rate. We'll see an increase in GAAP equity, and we'll see an increase in our GAAP income.
Unfortunately, at least initially, the benefits of the lower rates won't translate into materially lower cash taxes due to changes in our tax base. In addition, the admitted statutory deferred assets are going to be reduced substantially, resulting in lower statutory capital. For GAAP, we recorded a one-time adjustment of $874 million in the fourth quarter, largely related to the reduction of our deferred tax liabilities. This adjustment increased net income but had no effect on our net operating income. The adjustment also included around $275 million relating to our unrealized gains on our fixed maturity investments. On the balance sheet, the adjustment increased our book value per share, excluding the effect of unrealized gains on our fixed maturities by about 15%. Looking forward, the effective tax rate on our operating income will decrease around 12 or 13 percentage points and should be in the range of 19%-20%.
This reduction in the effective rate is the primary cause of the increase in the midpoint of our guidance for 2018. The increase would have been approximately $0.08 per share higher if not for lower excess tax benefits that are associated with our stock compensation expense due to the lower tax rate. We expect our return on equity, determined without the effect of unrealized gains on our fixed maturity portfolio, to be in the 14%-14.5% range, which is largely the same range it would have been absent tax reform. Finally, we expect our debt-to-capital ratio to be below 23% in 2018 if there are no further borrowings before the end of the year.
That compares to what would have been about 26% in 2017 absent tax reform and is well below the 30% ratio that is expected by our rating agencies. Now on our statutory financials, unfortunately the favorable impact that we're seeing for GAAP not necessarily translate into our statutory financial reporting. With respect to our statutory income, tax reform will not have a significant impact, at least in the near future. Our cash taxes paid will be largely unaffected as the benefits of the lower rate will be virtually offset by a reduction in the amount of our policy reserves, as well as the acquisition costs that can be deducted. The net impact that we expect is a cash tax saving probably in the range of $5 million-$10 million in 2018.
We do expect, though, to receive some incremental benefit over time as the statutory and taxable income grows. On our statutory balance sheet, we estimate a reduction in our 2017 admitted deferred tax assets in the range of $130 million-$140 million. Now this represents about 50% of the deferred tax assets that were admitted under the old law. This large reduction is attributable to both the lower tax rates as well as changes that are impacting the computation of the deferred tax assets. Ultimately, this reduction will lead to a reduction of about 30 basis points in our consolidated company action level RBC percentage for 2017, and we expect our consolidated RBC percent to be in the 300%-310% range. Should the NAIC adjust the RBC factors in 2018 for tax reform, as we expect, our consolidated RBC percentage should decrease by around another 45 basis points.
At this time, the targeted RBC levels for 2018 are yet to be determined, pending further discussions with our rating agencies and our regulators. Those are our prepared remarks, and we can take any questions at this time.
Let me start off with kind of a naive big picture question. Your business model is relatively straightforward and simple, as you said. You don't make a secret of it. You seem to tell the whole world exactly what you're doing, yet no one's been able to do what you do. There must be something here. Can you unveil the secret sauce? I'm just surprised no one else has been able to kind of achieve what you've been able to achieve in your businesses.
Well, I think the key component is the cost control. Back when I started in the life insurance business back in the 1970s, all the big companies were in the middle income market. There were a lot of companies that are still in the debit side of the market, but there were, I know Prudential, Met, all had large captive agencies. The problem was is that those were very costly agencies, and the cost of the agencies increased at a greater rate than their premium revenue. It ate into their margins. I think that's the reason they moved up to the higher end market. As I mentioned earlier, the policies we sell are very low premiums. They're annual premiums of $400 or less. You've got to be able to control the cost in order to have an underwriting profit to do that.
One way we've attacked it that's different than maybe other companies did in the past is that we don't have a lot of fixed costs in our agencies. All the agents and the head of the agencies are independent contractors. We pay them commissions. They pay all their expenses. We don't pay the expenses of operating an office. We pay their leases and that kind of thing. They pay all of those expenses. We only pay them money when they produce a policy. In the past, a lot of companies made the captive agents or exclusive agent, made them employees, and they had to pay employee benefits and other things. That's what really drove up the cost. Having a variable cost method and then also controlling the fixed costs we do have, I think, has been the key factor for us.
The other question is, the track record obviously is pretty impressive. What could throw it off? What should we be concerned about? A macro environment change, competitive environment change? What are you guys worried about?
There's not so much worry about a macroeconomic change. As I mentioned, we function pretty well in this marketplace. We're selling protection products. We're not selling investment products. People, when they do recognize the need for the insurance and they buy it, they generally keep it. As I said, the persistency is very strong. I think one thing we've been suffering from is a low interest rate environment. For us, we want higher rates. As I mentioned earlier, it doesn't affect our policy liabilities. It just would add more investment income, and there we're kind of hopeful. I think that maybe the bigger threat is the competition. If more companies could find a way to get back in the market, we don't think they'll do it through exclusive agencies. I think it just costs them too much to really gear up and build up to be a threat.
We've seen some try through direct response, and we could see more of that competition. I think Larry and I've talked about a lot, I think we can meet that competition. I think the biggest thing that we have to guard against is that we don't lose our focus. We stay with what we're doing, enhance our methods as needed, but stay focused on our products and our market.
You go over the stat effects on page 37 of tax reform, you also say you typically generate $300 million of excess capital and commit to return that to shareholders. That's the usual plan. What effect, if any, do you think the stat changes are going to have on your capital allocation plans? I have a follow-up.
Okay. Really from a general capital allocation plan, we don't really see any material changes in those as we go forward. I think for 2018, we anticipate our excess, our free cash flow to be still in that $320 million-$330 million range. We have yet to talk with our board with regard to what changes, if any, we'll have with respect to our dividend policy. Really don't see any real material changes in how we think about kind of the mix of how we return our cash to our shareholders. What we know is absent of better alternatives, that we'll look at returning that cash back one way or another.
Okay. I think that answers the follow-up, just to be clear, do you have any plans currently to raise capital externally?
At this point in time, no. From a standpoint that we really need to have those discussions with our rating agencies and our regulators to really see what makes sense for a right target level for Torchmark. We've been historically at around a 325% level. Is a 300% level right? Is 275% right? Is 325% going to be right? Those are the discussions we still need to have. We do believe if in fact we need to replace some of that income or some of that capital to get back to, let's say a 300%, or even the 325% level, that we should be able to fund most of that. We think we can tap the debt markets to be able to do that with the lower debt-to-cap ratio that we're seeing, projecting here for 2018, that we do have capacity to do that.
Something we'll have to work through yet, we think that at least for the most part, we should be able to do that without negatively impacting our excess or the share buyback program.
Okay. I have others, but I'll give others a chance.
I have one other question, and then we'll see if we only have two minutes, but mine I think is a quick one. The lag between agent growth and sales, is there one and how long is it?
There is a lag, and it's usually about 12 months. A new agent is not as productive, has no experience for a veteran agent, so the average premium that they write is usually lower than a veteran agent. We have a needs-based presentation, and as they become more experienced, they better meet those needs of that consumer. The other factor in growth of premium is the activity, and we find that veteran agents are typically more active than a new agent. It takes 6 to 12 months for an agent to figure out the business to have the right level of activity, and to increase that average premium per sale.
Got it. Helpful. We got one minute left. Scott, you want to throw one more in there?
Yeah, sure. You also talk about the RBC levels and your conversations with regulators. I wanted to talk about the NAIC. My understanding is they have until about June to address this for the 2018 filing year. Is that correct? Is the 45 basis points you talk about only due to the investment adjustment, and how would you characterize the direction of the discussions with the NAIC?
The 45 basis points relates to adjustments of the RBC factors for tax reform. If they go back and change it from 35%-21%, that would increase presumably the RBC factor. I'm not sure about the timing as far as that's concerned that they would need in order to have it be effective for 2018. My understanding is that they are working on that and at least intend to make those changes to the RBC factors. As far as the C1 factors on the investment income, from what I understand, those can hit a little bit of a stalemate and the industry is pushing back on the adjustments that they're looking at trying to make with respect to the bond factors. It's anticipated again, that they may make that into 2018.
At least from what I understand, there's at least some risk that they wouldn't.
Well, the main thing I'm driving at is, do you think that they view this as a meaningful capital issue that's going to require more than a passive response from the industry? Or is it the other end of the spectrum, say it's like kind of a shadow capital issue where filers may get forbearance? I mean, which end?
I mean, I really do think that that's yet to be seen. I'm not really sure where they're going to come out on that.
Great.