Our next presenter is Torchmark. We are very happy to have Mark McAndrew, CEO, and Gary Coleman, CFO, with us today to tell us the Torchmark story. As I mentioned this morning, there is a lot of focus in our new primer about cash flow, transparency of cash flow, return of cash to shareholders. On any of those measures, when you look historically, there is one company that is at the top of the list, and that is Torchmark. I am going to pass it to Mark for an update on the TMK story.
Thanks, Ed. First off, I have been told by my general counsel that I have to show this slide so everybody can read that real quick. Good morning, everybody. Torchmark was formed back in 1980 when Liberty National Life in Birmingham, Alabama, acquired Globe Life out of Oklahoma City. For the first 15 years of our existence, we grew rapidly, both internally as well as through acquisition. The company also became much more diverse as we acquired a mutual fund company, started an oil and gas subsidiary, a real estate development company, as well as a major expansion of our property and casualty subsidiary. Over the past 15 or so years, the complexion of Torchmark has changed significantly. We have spun off Waddell & Reed. We also spun off Vesta, which was our property and casualty subsidiary, and we sold Torch Energy.
Other than our administrative offices and our direct response facilities, we have sold off all of our real estate holdings, and a year ago, we also sold off United Investors subsidiary, which basically was our variable life business. We have simplified the company and have focused on growing our core businesses. This change in company direction has also changed the way the company is perceived by analysts and investors. The most common terms used to describe Torchmark today are safe and predictable, particularly in difficult economic climate. I would have to say that those terms are accurate. If you look at this first chart, you will see that our operating earnings per share showed remarkably consistent growth from 1995 through 2011. The compound annual growth rate was 8.7%.
For 2011, we grew our operating earnings per share by 9.6%. If we achieve the midpoint of our current guidance, we will see operating earnings per share growth of 12.2% for 2012, which would be our best year in the past 10 years. Excluding the net unrealized gains or losses on our fixed maturities, our growth in book value per share has also been remarkably consistent. For the period 1995 through 2011, the compound annual growth rate was 10.9%. Our GAAP book value per share has also increased significantly over the same period, although not with the same consistency due to fluctuations in the market value of our portfolio. Over the past 16 years, our compound annual growth rate has been 11.4%. There are three keys to Torchmark's success. First, we have very high underwriting margins.
In fact, I believe our underwriting margins are among the best, if not the best in our industry. How are we able to achieve these high underwriting margins and maintain them? There are really several key factors. First, we sell basic protection products. We focus on life and higher margin supplemental health products. We do not market investment products or high risk volatile primary health care products. Second, we operate in relatively non-competitive markets. Our market is middle income America, a market which we consider to be vastly underserved. Third, we control our distribution. Our three primary distribution systems are all controlled. They're either captive agencies or direct response. Unlike companies who operate through independent distribution channels, we don't have to directly compete for an agent's business through higher compensation, product enhancements, or lower rates. Lastly, we control our expenses, both administrative and acquisition.
As Ed can attest, it has been a part of the culture of Torchmark since its inception. We're constantly finding ways to be more efficient while improving our service. The next chart shows a split of our business between life and health insurance. Over the last 16 years, our life underwriting margins have increased from 56%-78% of our total. This shift was by design. We chose to exit the under age 65 primary healthcare market, and we did that for a number of reasons. The business was high risk, it had very poor persistency, it had low underwriting margins, and it had a much higher potential for litigation and regulation. Our largest and most profitable distribution system is American Income Life, which we acquired in 1995.
As you can see from this graph, it has enjoyed a long history of growth, with life premiums growing at a compound annual growth rate of 9% since 1995. American Income's traditional market has been labor unions. American Income is one of the few all union companies. Both our agents as well as our home office employees are members of OPEIU. We currently have the endorsement of 88 international labor unions. Over the last 10 years, we've made a concerted effort to expand beyond that market. Today, only 28% of our new sales at American Income come from union-endorsed lead cards. In the last 10 years, our producing agents at American Income have grown from 1,768 to 4,381, an increase of 148%, while our net life sales have grown from $66 million-$142 million, which is 115% increase.
We fully expect to continue this growth. In the fourth quarter of 2011, our net life sales at American Income grew by 11%. We had a 12% increase in our producing agents over the prior 12 months. We are currently projecting 12%-14% growth in our life sales in 2012 at American Income. Although I will say that we're seven weeks into the first quarter at American Income, and I'm actually pleasantly surprised with through seven weeks, our sales growth is exceeding 20% at American Income. I'm very encouraged by the growth prospects there going forward. This growth in premium did not result in declining margins. On the contrary, the persistency and mortality of the business at American Income continues to improve.
While our life premiums grew at a compound annual growth rate of 9% over the last 16 years, our life underwriting margins increased at an average rate of 9.5% over the same period. Our second largest distribution system is the direct response operation at Globe Life. It too has a long, consistent history of growth, with life premiums also achieving a compound annual growth rate of 9% over the last 16 years. Globe also operates in a relatively non-competitive market, selling basic life insurance products to middle and lower middle income households. Globe Life also enjoys some competitive advantages which make it very difficult for other companies to enter that marketplace. Our biggest advantage is the ability and experience of our people. We have been in this market for 47 years and have accumulated a wealth of experience in what works and what doesn't work.
We have assembled the most talented group of people in my 32 years with Torchmark. Our second big advantage is our control and cost containment over every aspect of the sales process, from product design, list compilation and modeling, package design, printing, envelope manufacturing, letter shop, all are done in-house. As a result, there's no one who can come anywhere close to matching our costs. While continuing to grow our traditional direct mail and insert media distribution, we are constantly searching for new areas of growth, particularly the internet and social networking sites. While life premiums at Globe grew on average of 9% over the last 16 years, life underwriting margins increased 7% over the same period. However, over the last 10 years, underwriting margins have held at 25% of premium in spite of numerous postage increases and inflation in both our printing and labor costs.
We expect 2012 to be a very good year in direct response. While the midpoint of our guidance assumes only a 6% increase in our life sales, our early results, again, in the first quarter indicate that we should significantly beat that number. In fact, for those of you who follow Torchmark, in our direct response, we always offer an introductory offer of $1 for the first month. We don't treat it as a sale till people pay the first full premium on that introductory offer. If I look at our new sales of people who have paid the $1, who have not yet been billed for the full premium, the last four weeks, those sales are up 31%, which again, is a very good sign. Over the next six to eight weeks, we will see our net sales or reported sales follow that same pattern.
We expect to see a very good first quarter as well as second quarter, and I think we'll significantly beat that 6% that we've included in our guidance. Our third primary distribution system is Liberty National Life, which was the original parent company at the time Torchmark was first incorporated. Liberty National was founded in 1900 and has operated as a home service company in six southeastern states. In fact, the acquisition of Globe Life and the formation of Torchmark in 1980 were a direct result of Liberty National's inability to sustain internal growth. As you can see from this chart, that challenge continues today. Over the past 16 years, life premiums at Liberty National have only grown by $2 million. The challenges at Liberty National are not unique. They've been faced by every home service company.
The biggest challenge has been the cost structure, which is characterized by high fixed acquisition costs. In that distribution system, all of the offices or company offices, up until a few years ago, they were all owned and maintained by the company. All of the sales personnel were employees who were paid a salary as well as benefits in addition to commissions. The high fixed cost structure was just something that could not be sustained, and most companies in that marketplace have either abandoned or de-emphasized this distribution model. Maintaining Liberty National's underwriting margins has also been a challenge. The life underwriting margin in 2011 was actually $5 million less than it was 16 years prior. Beginning in 2003, we began a slow transition away from this high fixed cost structure. Today, our agents and field sales management do not receive salaries. They are compensated by commissions and bonuses only.
Since November 1st of last year, all new hires are independent contractors and are not eligible for employee benefits. In effect of January 1st of this year, all of our sales office expenses are the responsibility of our branch managers. The effect of these changes is a major reduction in our fixed acquisition costs, which will result in a much more consistent and improved underwriting margin on our new business going forward. The challenge to grow remains. In December of last year, we changed the executive management at Liberty National. Roger Smith, who was the CEO of American Income, took on additional responsibilities as CEO of Liberty, and Steve DiChiaro, who was a very successful SGA at American Income, became chief marketing officer.
Their goal for the next six to nine months is to implement similar recruiting, training, and sales processes that have worked at American Income and have worked well at American Income. I'm optimistic that we will begin to see a turnaround at American Income the second half of 2012 and should be in position to see growth at Liberty National in 2013 and beyond. While it's true that Liberty National has been a drag on Torchmark's growth since our inception, the good news is that it's becoming a smaller and smaller drag. As you can see from this chart, in 1995, Liberty National represented 39% of our total life premium. Today, it's only 16%. On the other hand, American Income and Globe, in our direct response, accounted for 43% in 1995. Today, they represent 69% of our total.
In the life underwriting margins, Liberty National shrank from 36% in 1995 to only 13% in 2011, while American Income and Globe grew from 50% to 71% of the total. When we look at new life sales, Liberty National declined from 20% in 1995 to only 7% in 2011. While American Income and Globe increased from 48%, they're now 86% of our total life sales. I've not given up on Liberty National. I believe it will turn around in 2012 and contribute to our future growth. I'm also very encouraged by the recent trends at American Income and Globe and believe both are well-positioned to see the best sales growth they have seen in a number of years. The second key to our success is our conservative long-term investment strategy.
As our life premiums and equity have grown, so too have our invested assets, from $5.5 billion in 1995 to $11.4 billion at the end of last year. Throughout this period, we've invested predominantly in long-term investment grade corporate bonds. We believe this investment strategy is the best match to the long-term duration of our policy liabilities and provides the best risk-adjusted return to our shareholders. On this chart, you can see a history of our below investment-grade bonds. At the end of 2011, they stood at 6.4% of our fixed maturity portfolio, the lowest level since 1999. In June of 2009, our below investment-grade bonds hit an all-time high of $1.2 billion, or 13% of our portfolio, during a time when 30% of the portfolio had been downgraded and over half of those downgrades were two or more notches.
In the last half of 2009, we sold $350 million of below investment-grade bonds and brought that percentage down to 8.1% by the end of the year. As you can see from this chart, the unrealized gains and losses in our investment portfolio have seen some wide swings, particularly in the last five years. At year-end 2011, we had $964 million of unrealized gains, the highest in our history. With most of the same bonds that we held three years earlier and had a $1.8 billion unrealized loss. These fluctuations in the market value of our portfolio normally don't concern us because we intend to hold them and we have the ability to hold them to maturity unless there is a credit problem. Even in 2009, we never had to sell a bond to meet cash needs.
There's been much discussion and concern recently about a prolonged low interest rate environment and the impact it would have on insurance companies. At Torchmark, this has been an issue we've been dealing with for a number of years. At the end of 2011, the average yield on our fixed maturity portfolio was 6.49%. Which is 132 basis points lower than it was in 1995. It's 109 basis points less than it was 10 years ago. It's 53 basis points lower than it was five years ago. While the decline in portfolio yield has been a drag on our growth in earnings per share, it has not stopped it. In 2011, we saw 9.6% growth in our operating earnings per share, even though our portfolio yield dropped by 14 basis points. What impact will continued low interest rates have going forward?
To answer that question, I first need to remind you of the sources of our excess investment income. While 70% of our invested assets cover our policy and debt obligations, only 21% of our excess investment income is derived from the spread between our portfolio yield and the interest required to fund our obligations. The other $3.4 billion of assets represent our equity and contribute 79% of our excess investment income. To project the impact of an extended low interest rate environment, we looked at a scenario where all new investments for the next five years were at a constant yield of 4.75%.
Because of the low volume of our maturities and calls in the next five years, which should only range between 2% and 3% per year, the impact on our overall portfolio yield would be a reduction of about 50 basis points, roughly the same rate of decline that we've seen over the last 10 years. Because our assets and equity will continue to increase over the next five years, the impact is not that great. In 2011, our excess investment income on a per share basis was $2.68, with $0.57 coming from the spread over our obligations and $2.11 coming from the interest earned on our equity. In our stress test, we assumed our policy and debt obligations would increase 4% per year and equity would grow by 3% per year. We also assumed that we would continue to use our free cash to repurchase shares.
If our portfolio yield were to remain flat at 6.49%, our excess investment income per share would grow at a compound annual growth rate of 10% over the next five years. However, if we invest new money at 4.75% over that same period and the overall portfolio yield declines to 6%, our excess investment income per share attributed to the spread on our obligations would drop from $0.57 down to $0.24. The interest on our equity would still grow from $2.11 to $3.17. Our total excess investment income per share would still grow at a compound annual growth rate of 5%. With the growth we expect in our underwriting income and continued use of our free cash in our share repurchase program, we believe we can sustain double-digit growth in our operating earnings per share in spite of a prolonged low interest rate environment.
I would also point out that due to the nature of our products, we can maintain our profitability even with lower interest rate assumptions. For most of our products, lowering our interest rate assumptions, lowering the interest rate we're crediting to our policy reserves by 100 basis points would only require a 1%-3% adjustment in our rates to maintain the same level of profitability on new business written. We actually implemented a 20% increase in our new business rates at Liberty in late 2010 and have made a 5% adjustment in our rates at both American Income and most of the products at Globe Life beginning last month. The last key to Torchmark's success is our strong, consistent free cash flow at the parent company.
We define free cash flow as the dividends received from our subsidiary companies, less the interest expense on Torchmark debt and the dividends paid to Torchmark shareholders. The net amount of cash left over is free cash that can be used by the parent for any corporate purpose. This chart shows the dramatic growth in our free cash flow between 1999 and 2007, from $110 million to $353 million. While it dipped from 2008 to 2010 due to investment impairments and increased capital requirements at the subsidiaries, it rebounded nicely in 2011 to $367 million. We're expecting a small decrease in 2012 free cash as a result of our United Investors sale, but we expect to see renewed growth in our free cash flow in 2013 and beyond. As mentioned earlier, since 1996, we have used our free cash flow to repurchase company shares.
Over that period, we've acquired 128 million shares. During 2011, we spent $788 million to acquire 18.9 million shares, which represented 16% of the diluted outstanding shares at the beginning of the year. For 2012, with the $74 million of assets on hand at the beginning of the year, plus the additional free cash from the subsidiaries, the parent will have between $425 and $435 million of cash available. We do plan to use most, if not all, of this cash to continue our share repurchase program. To summarize, 2011 was a very good year. 2012 is shaping up to be our best year in at least the last 10. We're excited about our prospects for growth and believe we can sustain double-digit growth in our operating earnings per share going forward, even in a low interest rate environment. Those are my prepared comments.
Also, everything I've seen so far this quarter really indicates that 2012 will even be better than our projections. Again, our sales of both American Income and direct response are exceeding our expectations at this point. As I've talked about on some of our calls, our efforts to conserve business that has been lapsing off is running about 30% ahead of our plan at this point through January. Even on the health side, our Part D revenues are exceeding our expectations. We expected that revenue to grow about $100 million this year, and that is doing better than anticipated. We're really excited. We like where we're at right now, and we expect 2012 to be an outstanding year. With that, I'll open it up for any questions. Yes.
Thank you. Two questions. The first one is, if you look at the LIMRA studies that come out, it continues to show the penetration of U.S. life insurance ownership going lower and lower.
Absolutely.
Your recent results, and even beyond the recent results, are very contrary to that. I was just wondering if you'd just give us a little bit more color on how you're able to do that.
Well, that is nice. That's why we like the markets that we're in. We feel like the middle-income marketplace is vastly underserved. I tell people, when I started 36 years ago, I worked for Prudential for a while. They had 35,000 home service agents in this marketplace, and today they have very few. It's been abandoned by most companies. You're right, LIMRA came out with a study a year or so ago that said 74% of the households in this country recognize that they don't have enough life insurance, and it was over 40% of the households had no life insurance. We think it's a vastly underserved market, and we intend to fill that marketplace. The only thing restricting us is not our market, it's just internal. Can we continue to develop people, to recruit more people, and train more people?
Again, if I look at American Income, they've grown from 1,300 agents in 1999 to 4,600. Our goal is to have 10,000 in five years. We think that's very achievable.
The follow-up question is unrelated. It's on the investment portfolio. If I remember correctly, a good portion of your investment portfolio is invested in U.S. financials. Recently, Moody's, I guess it was yesterday or today, announced they're going to be, within the near future, downgrading some of these larger global-type banks two or three notches. I'm curious as to what your initial thoughts are on the impacts on your capital ratios from that.
Well, I'll let Gary comment. I know really since 2008, we have been reducing our exposure in financials and diversifying more into other segments. Gary, you want to comment?
Well, also within the decisions on the bank side, I think we're fairly diversified there, not only the bigger banks, some of the other regional banks. As Mark said, we've reduced that exposure, and we're comfortable with the bonds we have today.
Right now, Gary, banks represent what % of our total portfolio?
It's less than 15% now, it had been closer to 20%.
Ed?
Two questions. The decision to raise prices by 5%.
It seems like that's more than what might've been necessary given new money rates. Is there something that you see that suggests you have more pricing flexibility now than what you thought?
Well, again, raising the rates at American Income by 5%, we see as a non-event. Actually, we're trying to get a little bit ahead of the game. Again, the outlook has been for an extended low interest rate environment. We did bring down the interest rate we credited on new business in 2011. We expect to bring that down further in 2012 on new business written. Really it's just anticipating that we're going to be in a low interest rate environment for several years. We're just trying to stay ahead of the game. It is a little more than what we probably needed to do, but rather than put it through a two-and-a-half % increase, we went ahead and made it 5%.
As you're turning Liberty National into more of an American Income model, is there concern about cannibalizing anything that's going on in American Income in terms of recruiting or sales?
Well, obviously, that was something I had to think hard about. I hated to pull too much talent away from American Income at a time when it was performing very well. We have a lot of good people at American Income, and I'm really not concerned about that. Actually, if I look at first quarter results, our recruiting has been extremely strong. Again, I said, I think on the last call, our agent count in January went from 4,381 to over 4,600. Our life sales at American Income through the first seven weeks of this quarter are running over 20% ahead of a year ago. It's on a very good track, and I feel confident that we'll keep that on track, that we won't lose anything there by having Roger spend some of his time at Liberty National. Any other questions?
Well, I think our time's about up.
We can take one more if there's any.
Sure. I don't see any hands.
Well, I'm going to ask one more then.
Okay.
Of the ability to grow your EPS at the rate that you said long term, how much of that would be buyback in your view?
Well, again, I think we'll see better growth in our underwriting income going forward. A significant amount will still come from share repurchase. I know people have said, well, as our stock price goes up, we'll be able to buy back fewer shares. On the other hand, we also expect our free cash flow to grow going forward. It will continue to be a combination. Gary, you comment?
No, we will see. We're projecting growth in underwriting income going forward at a higher pace than we've been. I still think I agree with Mark, share repurchase is still going to be a big factor there.
Okay. Thank you very much.
Okay. Well, thank you for listening.