Notorious robber Willie Sutton once remarked that he robbed banks because that's where the money is.

These days, the money tends to be in stock. And if you want to make money in stocks, I think it's important to study in the footsteps (and SEC filings) of the great value investors. After all, one of the most important engines of Berkshire Hathaway (NYSE:BRK-A) juggernaut is GEICO -- a company where Warren Buffett's mentor, Ben Graham, once sat as chairman of the board of directors.

Similarly, it's an important exercise to look into Berkshire's SEC filings and try to learn from Buffett's stock prowess. If we look at Berkshire's SEC filings, we can see that on 12/31/99, Buffett owned 662,562 shares of Torchmark (NYSE:TMK). By 3/31/00, Berkshire's ownership in Torchmark more than tripled to 2,073,462 shares. Although we can never know exactly why and under what circumstances Berkshire bought those shares, if we attempt to "reverse-engineer" the investment by reading Torchmark's 1999 annual report and 10-K, perhaps we can learn something valuable.

Here are some of the lessons I learned from studying Berkshires purchases of Torchmark shares:

Invest in low-cost producers
Buffett is a huge fan of companies that are low-cost producers. As previously mentioned, Berkshire's GEICO has a huge cost advantage because it cuts out the middleman (the insurance agent) and directly sells auto insurance to the policyholder. Some other low-cost producers currently in the Berkshire portfolio are Costco (NASDAQ:COST), Wal-Mart (NYSE:WMT), USG (NYSE:USG), and Shaw Industries.

Torchmark, as a life and health insurer, isn't a pure-play direct-to-consumer writer of insurance. However, in 1999, Torchmark's direct-response business accounted for 24.1% (up from 21.7% in 1997) of life insurance premium income. As one of the fastest-growing segments, the direct response was sure to become a more integral part of Torchmark. In 2006, this segment accounted for 30% of life insurance premium income.

Low expenses are a huge competitive advantage for insurers, allowing them to set competitive policy rates (which contributes to growth) while maintaining underwriting margins (which contributes to profitability).

Torchmark's low expenses help the company earn some of the highest underwriting margins in the industry. Administrative expenses as a percentage of premium income fell to 5.6% in 1999, from 5.9% 1998. Their 1999 10-K states, "Torchmark has long been recognized as a low cost administrator of business... Our noncommission expenses related to new sales on a per policy basis are among the lowest in the industry."

Invest in high-quality companies that intelligently allocate capital
Torchmark's return on equity -- a great "quick and dirty" way to judge a company's return on capital and management's capital allocation skill -- was 16.2% in 1999 (calculated using net operating income). Furthermore, the company had unique businesses selling policies to the military and unions. Not only do these businesses have moats built around distribution (historical relationships), but they're also characterized by lower policy obligation ratios (low loss ratios) and low lapse ratios (and thus high persistency). Because it costs much less to renew a policy (because of high acquisition expenses), policies with low lapse ratios are very valuable.

A couple other notes I made while reading Torchmark's 1999 10-K were that the company had an overfunded pension plan, and the company was consistently buying back shares. In 1997, Torchmark's diluted share count was 141.4 million. By 1999, that had fallen to 134 million, and the number has subsequently worked all the way down to 99 million. Many Berkshire investments are big buyers of their own shares, so Buffett must have been pleased with management's penchant for doing this.

Look for companies whose GAAP income understates economic earnings
In 1999, Torchmark posted $274 million in generally accepted accounting principles (GAAP) net income. However, results were dragged down by $72 million in realized losses from the sale of real estate and fixed-income investments, and another $13 million in losses were due to a non-recurring charge from a marketing agreement with Reader's Digest (NYSE:RDA) that didn't work out. A more accurate measure of Torchmark's profitability was its net operating income of $341 million, which is about 25% higher than stated GAAP income.

Invest at a low valuation
We don't know exactly when Berkshire increased its Torchmark stake, but based on SEC filings, I can make a reasonable guess that the purchases were made in the $20 to $24 per-share range, so I'll just guess and say $22 per share. Based on the 134 million diluted shares outstanding at the time, that would give Torchmark a $2.95 billion market cap. That means Torchmark was trading at about 1.35 times book value per share of $16.30 (this number excludes unrealized losses in the investment portfolio) and about 8.7 times trailing net operating income. These are extremely cheap multiples for a company of Torchmark's quality. In comparison, the company now trades at 1.9 times book value and 13 times trailing net income.

The best part about the investment is that in early 2000, when Berkshire was buying, shares went down for no reason. I recently spoke with some of Torchmark's executives and asked them about this. The shares basically went down because they were an "old economy" company during the dot-com boom. When buying a company of Torchmark's pedigree at a cheap price, Berkshire usually had to wait for some sort of scandal or seemingly risky situation, such as the salad oil scandal at American Express (NYSE:AXP) or the cyclical downturn and subsequent chapter 11 filing at USG. In this case, there were no such strings attached, other than the fact that Torchmark's name didn't end with a dot-com. Talk about inefficient markets.  

If we assume Berkshire's purchase of Torchmark was around $22 per share in early 2000, then those shares have nearly tripled since the initial purchase price, providing an extremely attractive 17% annualized return in a time period when the S&P's total return was about 0%.

The pattern seems clear. Mr. Buffett buys shares in extremely high-quality companies when he can get them at a discount. The rewards for his patience often result in years and years of stellar returns.

Related Foolishness:

•  Warren Buffett's Priceless Investment Advice

•  M&T: A Page From Warren Buffett's Playbook

•  Bend It Like Buffett: Dissecting Moody's

Ready to start investing like an adult? Check out a complimentary 30-day trial to Motley Fool Inside Value.

Fool contributor Emil Lee is an analyst and a disciple of value investing. He doesn't own shares in any of the companies mentioned above. Emil appreciates your comments, concerns, and complaints. Berkshire Hathaway, USG, and Wal-Mart are Motley Fool Inside Value recommendations. Costco and Moody's are Stock Advisor recommendations. The Motley Fool has a disclosure policy.