« 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
“Over the long term, it’s hard for a stock to earn a much better return than the business which underlies it earns. If the business earns 18% on capital over 20 or 30 years, even if you pay an expensive looking price, you’ll end up with a fine result.” — Charlie Munger
Download the Slide Deck (PDF) : The_Compounding_Spread
Listen to the deep-dive discussion – Why Visa’s Thirty Percent Becomes Twelve Percent
I. The Paradox
Visa earns over 30% on its invested capital. It has done so for years, and the architecture of the business—a toll on every electronic transaction, growing at zero marginal cost—suggests it will continue to do so for decades. Thirty percent, compounded over twenty years, turns every dollar into twenty-three.
And yet the shareholder of Visa does not earn 30%. The long-term return to the equity holder has been closer to 12 to 15% annually. Extraordinary by any standard—but not 30%. Not even close.
Where does the difference go? The answer lies in a mechanism the market rarely names and most investors never decompose: the spread between the rate at which a business generates wealth and the rate at which that wealth can be recycled back into compounding.
II. The Generation Rate
The generation rate is what the business earns on its existing operations. For a Freesurfer, this rate is exceptionally high—because the growth is free. Revenue grows as the external base expands (more transactions, more bonds issued, more assets indexed), while costs remain essentially fixed. The result is an ever-widening margin on an ever-growing revenue stream, producing returns on invested capital of 30% or more.
This rate is real. It is not an accounting illusion. Visa genuinely converts thirty cents of every dollar of capital into profit, year after year. The ROIC that analysts cite and investors celebrate is the generation rate—and it is magnificent.
III. The Recycling Problem
But the business cannot consume its own output. Visa generates roughly $20 billion in free cash flow every year, and it has almost nowhere to put it. The payment network is already built. Processing the next billion transactions requires no additional capital. The infrastructure that produces the 30% return does not need—and cannot absorb—the cash that the 30% return produces.
So the cash must leave. It exits through two channels: dividends, which return cash directly to the shareholder, and buybacks, which retire shares and increase the per-share value of what remains. Both are forms of recycling—the business is converting its operational output into shareholder return. The buyback is, precisely, the purge of the unabsorbable—the mechanism by which the business expels the cash it cannot compound internally.
The dividend yield is roughly 1%. The buyback yield is roughly 2 to 3%. Together, perhaps 3 to 4%. This is the recycling rate on the returned capital—not 30%, but 3 to 4%.
IV. The Spread
Here is the mechanism that explains everything.
The generation rate is 30%. The recycling rate on returned capital is 3 to 4%. The spread between the two—roughly 26 percentage points—is the invisible cost of being capital-light. It is the price the Freesurfer pays for needing no capital: the cash it produces cannot compound at the rate that produced it.
The shareholder’s actual return is not the generation rate. It is a blend: the organic growth of the business (driven by the generation rate acting on the existing base) plus the shareholder yield (driven by the recycling rate on the returned cash). For Visa, this blend lands at approximately 12 to 15%: organic growth of 10 to 12% from the expanding external base, plus 2 to 3% from buybacks and dividends.
Twelve to fifteen percent, compounded over twenty years, is wealth-building by any measure. But it is not thirty percent. The spread between generation and recycling is where the difference goes.
V. Munger’s Convergence, Refined
Munger taught that over time, the return on a stock converges to the return the business earns on its own capital. If the business earns 18%, the shareholder will eventually earn something close to 18%. This principle is one of the foundations of the framework.
But the principle requires a refinement. The convergence is not to the headline ROIC. It is to the effective compounding rate—the blend of the generation rate and the recycling rate, weighted by how much capital the business retains versus returns.
A business that earns 30% but returns most of its cash at 3 to 4% does not converge to 30%. It converges to the blend—12 to 15%. A business that earns 18% and reinvests all of it at 18% converges to 18%. The business with the lower ROIC but the higher recycling rate can deliver a higher return to the shareholder than the business with the spectacular ROIC that cannot reinvest.
This is why ROIC, taken alone, is an incomplete measure of compounding quality. The generation rate tells you what the engine produces. The recycling rate tells you how much of that production compounds forward. The spread between the two tells you what the shareholder actually receives.
VI. The Spectrum of Spreads
Place the great businesses on this axis and a map emerges.
The Freesurfer (Visa, SPGI, MSCI): Generation rate 25 to 35%. Recycling rate 3 to 4% (via buybacks and dividends). Spread: very wide. Shareholder return: 12 to 15%. The engine is magnificent. The recycling is modest. The spread is the price of needing no capital.
The capital absorber (Google, if AI works): Generation rate 20 to 25% on Search. Recycling rate 12 to 20% if AI/Cloud deployment earns sustained returns. Spread: narrower. Shareholder return: potentially 15 to 20%. The engine is slightly less magnificent, but the recycling is more efficient—the cash stays inside and compounds at a rate closer to the rate that produced it.
The dream business (Buffett’s ideal): Generation rate equals recycling rate. Spread: zero. Every dollar earned is redeployed at the same rate. The compounding is complete—no leakage, no degradation, no reinvestment problem. This business may not exist in its pure form, but it is the theoretical ceiling of compounding efficiency.
The mediocre business: Generation rate 8 to 10%. Recycling rate 8 to 10% (because it must reinvest everything just to maintain operations). Spread: zero, but at a low rate. The compounding is complete but unimpressive. No spread, no excess, no wealth created.
VII. Why 12% on a Freesurfer Is Still Extraordinary
The spread sounds like an indictment. It is not. Twelve to fifteen percent compounded over twenty years, on a business with a near-certain architecture and a gap below the floor, is one of the finest outcomes available in public markets. The 30% ROIC is not wasted—it drives the organic growth that produces the 10 to 12% revenue expansion at zero cost. The spread is simply the portion that cannot compound internally and must exit the building.
What makes the Freesurfer exceptional is not that the spread is zero. It is that the generation rate is so high that even after the spread, the shareholder’s blend is 12 to 15%—a rate that most businesses cannot achieve even as their headline ROIC, let alone as a delivered return to shareholders.
VIII. The Deep Mechanics
Compounding is not one rate. It is two rates and the distance between them.
The generation rate is what the business produces. The recycling rate is what the shareholder keeps compounding. The spread is what escapes—returned to the shareholder as cash that must find a new home at a lower rate.
Every investment decision is, at its deepest level, a judgment about these two rates and the spread between them. The investor who buys a Freesurfer at a gap accepts a wide spread in exchange for near-certainty on the generation rate. The investor who buys a capital absorber accepts uncertainty on the recycling rate in exchange for a narrower spread. The investor who buys the index accepts mediocre rates on both in exchange for diversification.
IX. The Spread Creates the Dual Duration
There is one final connection, and it closes the loop on the entire framework.
The recycling spread does not merely explain the shareholder’s return. It explains the Freesurfer’s dual duration—the coexistence of short-term cash and long-term growth inside the same business. The spread is the mechanism. The dual duration is the result.
The spread splits the shareholder’s return into two distinct components with two distinct time signatures.
The short-duration component is the recycled cash—dividends and buybacks arriving continuously, quarter after quarter. This is the purge of the unabsorbable. It returns to the shareholder quickly and must be redeployed. It is cash in hand, short duration by nature, representing 3 to 4% of the annual return.
The long-duration component is the organic growth—the external wave compounding the intrinsic value of the business over decades, captured not as cash but as appreciation in the stock price. This is the free wave riding inside the moat. It compounds silently, long duration by nature, representing 10 to 12% of the annual return.
Three to four percent short. Ten to twelve percent long. Together, 12 to 15%. The dual duration is not a design choice. It is the structural consequence of a business that generates at 30% but can only recycle at 3 to 4%. The gap between the two rates forces the return into two time signatures—one fast, one slow—and the shareholder receives both at once.
If the spread were zero—if the business could absorb everything it generates at the same rate—nothing would exit and the return would be pure long duration. All value would compound internally. This is the dream business.
If the wave did not exist—if there were no organic growth, only cash distributions—the return would be pure short duration. All value would exit immediately. This is the cash cow without a future.
The Freesurfer is neither. It lives between the two because it has both: a wave that creates the long component and a spread that creates the short component. The dual duration exists because of the spread, not despite it.
The ROIC tells you what the engine can do. The spread tells you what arrives in your pocket. The dual duration tells you when it arrives. They are three views of the same architecture, and together they describe the deepest mechanics of how a Freesurfer compounds wealth.
Now you can name all three.
Watch the Video
Related posts on averagingup.com:
Heavy Tolls and Light Tolls · Why Berkshire Is Right About Google · $190 Billion to Grow · The Buffett Equity Coupon and the Munger Satisfactory Result · Toll Position
Based on the framework from The Infinite Investor, available at averagingup.com.