Capital Allocation Is the Job

The Core Insight

A dollar handed to Jack Welch in 1981 became 48 dollars by 2001, and the press crowned him the era's best CEO. A dollar handed to Henry Singleton in 1963 became 180 dollars by 1990, and almost nobody knows the name.

The gap is context. Welch rode the great bull market, an S&P at 14 percent a year, and beat it by 3.3 times. Singleton worked through deep bear markets and beat the same index by more than twelve times. William Thorndike opens The Outsiders with that comparison, then screens decades of records for everyone who cleared both bars. Seven more names came back.

The book reads as a job description. A CEO does two things: runs operations and deploys the cash those operations produce. Schools, boards, and the press grade the first. The eight won on the second. Buffett did the arithmetic on the stakes. A CEO who retains earnings equal to 10 percent of net worth controls more than 60 percent of the company's capital within a decade.

Most boards grade a CEO on growth: revenue, reported earnings, size. Thorndike grades one number, the change in value per share against the market and against peers over a full tenure. On that scale his eight averaged more than twenty times the S&P 500 and more than seven times their peer groups.

The Framework

The tool kit is small and closed. A CEO deploys capital five ways: investment in existing operations, acquisitions, dividends, debt paydown, and share repurchases. Money arrives from three sources: internal cash flow, debt, and issued equity. Identical operating results plus different allocation choices produce two very different outcomes for owners.

With no contact, the eight converged on one pattern.

• Run operations decentralized, and keep every capital decision at the top.
• Ignore reported earnings, and build a cash flow metric the industry lacks.
• Repurchase big blocks of stock when price falls below value.
• Make at least one acquisition worth a quarter or more of enterprise value.
• Pay no meaningful dividend, give no guidance, and stay away from the press.
• Hold the job past twenty years, the group's average tenure.

The temperament shows in the 1974 to 1982 stretch: 18 percent interest rates, two recessions, and a BusinessWeek cover that asked whether equities were dead. All eight bought through that window, shares or whole companies, while peers sat on their hands. Buffett named the force they resisted the institutional imperative, the pull on every CEO to imitate peers.

Key Ideas

The Same Playbook Won in Eight Industries

Tom Murphy compounded Capital Cities at 19.9 percent for twenty-nine years, 16.7 times the S&P and almost four times his media peers. Singleton's Teledyne returned 20.4 percent across twenty-eight years, almost nine times the conglomerates and more than twelve times the index. Bill Anders and his two successors compounded General Dynamics at 23.3 percent, 6.7 times the S&P and 1.8 times defense peers.

John Malone's TCI did 30.3 percent from 1973 to 1998, more than forty times the S&P and five times the public cable companies. Katharine Graham, twenty years out of the workforce when the job arrived, beat the index eighteenfold and her peers more than sixfold. Bill Stiritz did 20.0 percent over nineteen years at Ralston Purina, against 17.7 for peers and 14.7 for the market. Dick Smith ran General Cinema for forty-three years and turned a dollar into 684 dollars, against 43 dollars for the index. Berkshire compounded at 20.7 percent under Buffett from 1965 through 2011, a hundredfold ahead of the S&P.

None of them owned a hot product or a patent wall. All were first-time chief executives, and all but one arrived new to the industry. Murphy's stated goal: arrive at the station first on the least fuel. Thorndike's verdict: the edge was temperament rather than intellect.

The Suction Hose Beats the Straw

In 1972 Singleton called his board member Arthur Rock from a phone booth and said the stock was too cheap to ignore. Between 1972 and 1984 he ran eight tender offers and took in 90 percent of Teledyne's shares, 2.5 billion dollars. The tenders compounded at 42 percent a year. He knew both sides of the trade. His sixties stock issues averaged more than 25 times earnings. His repurchases averaged under 8.

Thorndike splits buybacks into the straw, a small program sipped through the quarters, and the suction hose, a huge tender at a cheap price. Murphy spent 1.8 billion dollars the second way, mostly at single-digit multiples of cash flow, and compounded it at 22.4 percent over nineteen years. Stiritz retired 60 percent of Ralston's shares, second only to Singleton. Malone took in more than 40 percent of TCI and earned above 40 percent compound.

Graham repurchased almost 40 percent of the Washington Post at single-digit earnings multiples, alone among newspaper executives. McKinsey advised her to halt the program, the pause lasted two years, and her son Donald later called it the most expensive consulting assignment ever. Buybacks create value at a discount to intrinsic value and destroy it at a premium. Buffett praised the tool in others and skipped it at Berkshire, a poor fit for its partnership culture. Near the end of his life, Singleton said universal enthusiasm for repurchases meant something was wrong with them.

Cash Flow Outranks Reported Earnings

Wall Street graded on earnings per share. The eight refused the metric. Malone went furthest. Higher net income meant higher taxes, so TCI rarely showed profit and never paid significant taxes while cash flow grew twentyfold. He coined EBITDA to give lenders a number that measured the machine. His mentor Bob Magness said it plainest: better to pay interest than taxes. AT&T repriced the heresy in 1998, twelve times EBITDA and 2,600 dollars per subscriber, paid in stock.

The rest built their own instruments. Singleton's bonus metric, the Teledyne return, averaged cash flow and net income for every unit. The accounting maximized cash, he said, so quarterly earnings jiggle. Dick Smith wrote cash earnings into his first annual report in 1962, likely the term's first appearance in American business. Buffett's instrument is float, money an insurer holds but does not own. Berkshire's float grew from 237 million dollars in 1970 to more than 70 billion by 2011. Munger sums the funding engine in one sentence: generate funds at 3 percent and invest them at 13.

Decentralize Operations, Centralize the Checkbook

Munger calls the shared structure an odd blend, decentralized operations under highly centralized capital allocation. The ratios read like typos. Teledyne kept fewer than fifty people at headquarters over forty thousand employees. TCI ran twelve million subscribers with seventeen people at corporate. Berkshire employs more than 270,000 people and staffs Omaha with 23.

The operating half is real delegation. Capital Cities printed decentralization as the cornerstone of its philosophy in every annual report. Murphy compressed the policy to one rule, hire the best people and leave them alone. The allocating half stayed at the top. Murphy sourced and negotiated every major deal himself. Buffett keeps no investment committee and reads no banker forecasts. Stiritz walked alone into due diligence on a multibillion-dollar deal with a yellow legal pad.

Patience, Then the Oversized Bet

Each of the eight made at least one purchase worth 25 percent or more of enterprise value. Murphy made the largest relative bet in the book. In 1986 he paid 3.5 billion dollars for the ABC network, more than his company's entire value, with financing from Buffett. Within two years the margins at ABC's stations rose from the low thirties past 50 percent. Nine years on, Disney paid 19 billion dollars for the combined company, 13.5 times cash flow.

The bets were rare because the filters were strict. Murphy required a double-digit after-tax return over ten years with no debt in the math. He never won an auction, and his bids landed at 60 to 70 percent of the eventual prices. Graham held acquisitions to an 11 percent cash return over a decade, and very few deals passed her screen. Dick Smith waited a full decade, then put 62 percent of enterprise value into publisher Harcourt Brace Jovanovich. He bought its publishing assets at six times cash flow and sold them later at eleven. The book's checklist flags the word strategic as corporate code for low returns.

Speed came from prepared criteria. When Union Carbide put the Energizer battery division up for sale, Stiritz's team met at one in the afternoon and opened the seller's books. By four they set a bid of 1.4 billion dollars from a back-of-the-envelope model. The final price of 1.5 billion dollars equaled 20 percent of Ralston's enterprise value, and the purchase compounded at 21 percent for fourteen years. Malone bought constantly, 482 acquisitions between 1973 and 1989, and even he priced by one rule. Pay at most five times cash flow after the quantifiable programming and overhead savings.

Selling Is Also Allocation

Bill Anders inherited the worst hand in the book. In January 1991 General Dynamics carried 600 million dollars of debt against negative cash flow. Revenue stood at 10 billion dollars, market value at 1 billion. His read: the Cold War's end left the industry with heavy excess capacity, so a company consolidates or gets consolidated.

In three years he generated 5 billion dollars of cash, half from squeezed operations and half from the sale of most of the company. Anders tried to buy Lockheed's fighter unit, and Lockheed countered with 1.5 billion dollars for his own F-16 business, the largest division and his favorite. The former fighter pilot agreed on the spot. He then sent the cash back to shareholders. Three special dividends, structured as tax-free return of capital, paid out almost half the equity value. A tender of 1 billion dollars retired 30 percent of the shares. Buffett bought 16 percent of the company and handed Anders his proxy. Smith ran the same discipline three times at General Cinema, and sold beverages, publishing, and Neiman Marcus at record prices.

Practical Applications

The epilogue scales the method to a bakery, which is close to seed scale. The owner sets a 20 percent hurdle. The next-door expansion costs 100,000 dollars and returns 20,000 a year, exactly the hurdle. A second store costs 200,000 dollars and models at 25 to 37.5 percent, with wider error bars.

Set your own hurdle before the next spend. Thorndike's checklist puts the floor above the blended cost of capital, mid-teens or higher, and ExxonMobil holds every project to 20 percent. Then write the founder's version of the five options: product work, a hire, a channel experiment, banked runway, retired debt or equity. Rank them by expected return and fund from the top. Kahneman's slow system needs a trigger, and a one-page return sheet is the cheapest trigger the book found.

Use the buyback benchmark even without buybacks. Ralston's hurdle for any acquisition was its own share repurchase return, and a deal had to beat it with some certainty. Your version is the return on your best existing motion, and any new bet must beat it.

Price your equity the way Singleton priced his. Chabraja ran that trade at General Dynamics in 1999. His stock traded at 23 times next year's earnings against a historic 16, so he paid for Gulfstream in stock. The translation: raise when the market prices you ahead of the business, and book dilution as a cost of capital.

Pre-commit the crisis stance. The eight bought through 1974 to 1982, and after Lehman, Buffett put more than 15 billion dollars to work in twenty-five days. Decide now what you buy if your market panics, and write the prices.

Who This Is For

Founders with revenue and a surplus to steer get the most. Every chapter is a decision they already face: fund the product, acquire, distribute, retire equity, or wait. Investors get a grading system for CEOs that starts from allocation instead of charisma. Readers who want sales machinery or product method leave empty-handed, because the book holds one argument and eight proofs.

The evidence deserves suspicion. Thorndike screened decades of records for market-beating outcomes, so the sample is selected on the result it explains, and no failed iconoclast gets a chapter. Part of the edge was tax law. Dividends were taxed twice for most of the period, so buybacks carried a structural advantage. That advantage is now consensus and therefore thinner. The figures come from the author's student research teams and admiring interviews rather than an audit. The one printed failure, the bankruptcy of Smith's legacy theater unit, gets a paragraph.

Skip it until real money moves through your company. The tools need a surplus to work on.

The Decision

The appendix holds the test, and it fits one line. Buffett asks whether a CEO created at least a dollar of value for every dollar of earnings retained. All eight passed. Most executives never face the question.

The founder version runs this week. Write your hurdle rate at the top of one page. Under it, list every planned use of next quarter's cash with a conservative return beside each. Fund what clears the bar, bank what does not, and date the page for the review.