Truly great businesses, earning huge returns on tangible assets, can’t for any extended period reinvest a large portion of their earnings internally at high rates of return.
— Warren Buffett, 2007
Download the Slides Deck: The_Capital_Cap
Listen to the deep-dive discussion – The Price of Free Growth
Visa earns a return on invested capital of thirty-eight percent, and the figure is rising. It spends less than three percent of its revenue on capital expenditure. It generates thirty-four billion dollars of free cash flow, and it returns nearly all of it to shareholders through dividends and repurchases. The most profitable business a portfolio is likely to hold reinvests almost nothing in its own base.
This is not a failure of ambition. It is a structural fact, and it is the defining trait of a particular kind of business—one whose growth does not pass through its balance sheet. The world digitizes its payments, and Visa’s toll widens in the world’s wake, without a dollar deployed to widen it. The same property that produces the extraordinary return produces an equally extraordinary limit: a company that need not fund its growth has nowhere to put its cash.
These two facts are one fact, seen from two sides. The return is high because almost no capital is tied up in the business. The reinvestment is nil because there is no base to feed. The overflow—the dividends, the buybacks—is not generosity. It is the only exit for cash the company cannot absorb. This is the price of free growth, and it is worth stating precisely, because it explains both why such businesses are treasured and why they are quietly constrained.
Visa: Why the Return Cannot Be Kept
Consider what a shareholder actually receives. Visa’s return on its own capital is thirty-eight percent. But a shareholder does not earn thirty-eight percent, because the company cannot reinvest at thirty-eight percent—it has no base large enough to absorb the attempt. It earns what it earns on a sliver of capital, and returns the rest.
The rest is recycled at a far lower rate. The dividend yields under one percent; buybacks add roughly three. The cash that overflows is returned to the shareholder, who must find somewhere else to put it—and that somewhere else will not earn thirty-eight percent, because businesses that earn thirty-eight percent on large quantities of capital do not exist. The gap between what the business earns and what the owner can compound is the price of free growth, paid in full.
None of this makes Visa a lesser business. Its earnings per share have grown at seventeen percent for nine years; its moat is wide and its uncertainty low; it offers return, permanence, and a tranquility that comes from having nothing to decide and nothing to fund. Among the three components of compounding, it holds all three in a rare balance—a high rate, a long duration, and a modest quantity, none of them sacrificed to the others. It is the closest thing to an equilateral figure the market provides. But the balance has a cost, and the cost is that the quantity cannot grow.
MSCI: The Overflow at Its Limit
MSCI shows what happens when the overflow has nowhere to go for long enough. Its capital expenditure is four percent of revenue; its free cash flow grew from 0.31 billion to 1.48 billion; its operating margin expanded from forty-two percent to fifty-five. It is a free-growth business as pure as Visa. And it has returned so much capital, for so long, that its shareholders’ equity has gone negative. It has repurchased its own book value out of existence—ninety-four million shares reduced to seventy-five, funded until the equity line itself turned red.
This is not distress. It is the logical endpoint of a business that cannot absorb its own cash. When there is no base to build, and the cash keeps arriving, the company returns everything, and then returns more than everything. MSCI is the overflow carried to its conclusion.
S&P Global: The Business That Bought Its Quantity
S&P Global began where Visa remains: a free-growth business of the purest kind. Its ratings franchise threw off cash at a return on invested capital that reached fifty-seven percent in 2016 and held above fifty for five of the next six years. It was an equilateral figure, and by the structural logic just described, it could not enlarge its base. Its growth passed outside its balance sheet, and the balance sheet stayed small.
In February 2022, it did what Visa cannot. It acquired IHS Markit for forty-four billion dollars, paid in stock, and in a single stroke enlarged its base by more than a third. Shares outstanding rose from 241 million to 325 million. Revenue rose from 8.3 billion to 11.2 billion. The company took the overflow of a free-growth business and converted it into a larger base—a base it had to buy.
The price appeared immediately, and at exactly the rate the structure predicts. Return on invested capital fell from 51.97 percent in 2021 to 12.69 percent in 2022, and then to 6.13 percent in 2023. A base purchased at market price earns, at best, the market’s return, and forty-four billion dollars of goodwill and intangibles entered the denominator at market price. The return did not collapse because the business deteriorated. It collapsed because free growth had been exchanged for purchased growth, and purchased growth does not come free.
This is the trade Visa never makes, because Visa cannot. A free-growth business is capped: its base grows only as fast as the world grows it, and no faster. S&P refused the cap. It reached beyond what its own structure would provide, and it paid the difference in return—fifty-seven percent surrendered for a base that a third larger.
The Purification That Follows
What S&P did next is as revealing as the purchase. It began to divest—Engineering Solutions to KKR for 975 million, then OSTTRA, and in 2026 the spin-off of Mobility, the automotive-data business inherited from IHS Markit. Mobility generated 1.6 billion in revenue, a fraction of what IHS Markit had brought. S&P was not undoing the acquisition; forty-four billion cannot be undone by shedding one and a half. It was purifying the base it had bought—expelling the pieces that did not compound like a free-growth business, keeping the recurring financial data that did.
The return has begun to recover: 6.13 percent, then 8.95, then 10.45. Whether it climbs back toward the rates that once defined the company, or settles into the low teens of an ordinary compounder, is the open question. If it recovers fully, S&P will have accomplished what no pure free-growth business can—enlarged its base and restored its rate, holding two things its structure says cannot be held together. If it settles, it will have diluted a rare business into a common one. The years ahead will decide which, and the decision is not yet made.
Alphabet: The Same Move, Built Instead of Bought
Alphabet is executing the identical maneuver by a different means. It takes the free cash flow of Search—a free-growth toll if one ever existed—and pours it into cloud and artificial-intelligence infrastructure, a base measured in the tens of billions of capital expenditure. It is deforming its equilateral figure toward quantity, reaching past the cap that free growth imposes.
The difference from S&P is that Alphabet builds its base rather than buying it. S&P’s purchased base entered the denominator at market price, and the return fell at once. Alphabet’s constructed base will earn whatever it earns, and the market cannot yet price what that will be. The question is no longer what was paid, but what the base will return—the same question this framework asks of every heavy toll: will it absorb capital productively, or dilute it. That Berkshire Hathaway has taken a position in Alphabet is the judgment of the most demanding of allocators that this costly new base will be absorbed rather than wasted.
The Cap and the Price of Leaving It
A free-growth business receives everything and can enlarge nothing. Its return is high because its base is small, and its base stays small because its growth never passes through its balance sheet. Visa cannot leave this condition; it can only overflow, as MSCI overflows to the point of negative equity.
Two others left it. S&P bought its way to a larger base and paid in return—fifty-seven percent reduced to ten, now slowly climbing as the purchased base is purified. Alphabet is building its way to a larger base, and the bill has not yet come. The cap is the price of free growth: the rate is maximal precisely because the quantity is fixed. And the price of exceeding the cap is the rate itself—surrendered at the moment the base is bought, or wagered at the moment it is built.
Watch the video