Listen to the deep-dive discussion – Why Spreadsheets Misprice Return on Infinity (21:04 min)
“Bearing a gift beyond price — almost free.”
— Neil Peart, “The Spirit of Radio,” 1980
* * *
I. The Toll and the Wave
Every business can be decomposed into two forces operating on different time horizons. The A is the toll — cash now. What the business collects today. Visa processes a payment and takes a fraction of a cent. Moody’s rates a bond and charges a fee. MSCI licenses an index and collects a royalty. The A is always short duration. What varies is the moat that protects it — which determines whether the A will repeat tomorrow.
The B is the wave — cash later. Growth. What the business will collect tomorrow that it does not collect today. In most businesses, this growth must be built. R&D spending, acquisitions, new store buildouts, capital allocation. Every dollar of future revenue requires a present dollar of investment and a human being deciding correctly. This is mechanical growth — the B that costs money and demands flawless execution.
But a rare configuration exists. In certain businesses, the growth arrives without a decision, without an investment, without an effort. Visa does not spend a dollar to digitalize the world’s payments. Governments push cashless economies. Banks install terminals. Consumers adopt mobile wallets. The growth arrives at the doorstep — uninvited, unpaid for, unannounced. This is natural growth — the B that costs nothing.
When a business contains both — a toll that collects cash now AND a wave that delivers growth for free — it is a Dual Duration company. The A finances the present. The B builds the future. But because the B is free, the relationship between the two is not additive. It is multiplicative: A × B. The free B adds volume. More volume generates more revenue. More revenue funds more buybacks. More buybacks reduce the share count. Higher cash flow on fewer shares restarts the cycle on a higher base. The B feeds the A. The A accelerates. This is the Freesurfer — a business that rides a structural wave at zero cost.
And when the moat protecting the toll is auto-reinforcing — when each transaction strengthens the network, when each user deepens the lock-in, when the moat does not need to be maintained because it maintains itself — the Freesurfer reaches its highest form. The toll is permanent. The wave is free. The moat is self-sustaining. The A, the B, and the moat fuse into a single structure where nothing needs to be decided, nothing needs to be spent, and nothing needs to be monitored. We call this the A paramount. Five businesses on earth qualify: Visa, Mastercard, S&P Global, Moody’s, and MSCI.
* * *
II. The Surfer Who Never Falls
An ordinary surfer falls. The wave passes. The cycle turns. The surfer paddles back out and waits for the next wave. Each ride is temporary. Each wave is an effort.
The Freesurfer rides a structural wave — digitalization, indexation, the expansion of global credit. These waves last decades. But even a structural wave can weaken at the firm level. FICO rode the wave of financialization with a regulatory moat — Fannie Mae and Freddie Mac mandated FICO scores for every mortgage. The wave is still there. But the FHFA authorized VantageScore as an alternative. The board cracked. The surfer fell while the wave continued without it.
The A paramount does not fall. Because the moat is not separate from the wave. The auto-reinforcing moat IS the board, and the board repairs itself. Each transaction reinforces the duopoly. Each new market that digitalizes adds volume to the rail. The board becomes stronger as the wave advances. The wave strengthens the board. The board channels the wave. The loop is closed.
This is the permanent wave. Not a wave that lasts a long time. A wave that cannot stop because the system that produces it is the same system that produces the board. The digitalization produces the transactions. The transactions reinforce the duopoly. The duopoly channels the digitalization. The wave and the board are the same thing.
* * *
III. Growth at No Cost
The standard discounted cash flow model has a numerator and a denominator. The numerator is the projected free cash flow — and the model is blind to the source. A dollar of cash flow is a dollar, whether it was earned by reinvestment or received for free.
But the denominator — the discount rate — is not blind. It embeds risk. Execution risk: will management allocate capital correctly? Competitive risk: will the moat hold? Growth risk: will the projected revenue materialize? For most businesses, these risks are real and material. ASML must spend $15 billion per lithography cycle. Berkshire must deploy capital through a single allocator’s judgment. Amazon must build warehouses for every dollar of growth. The growth carries risk because the growth requires decisions, and decisions can be wrong.
For the A paramount, every component of that risk is absent. The growth does not require a decision. The moat does not require maintenance. The competitive landscape does not require monitoring. The CEO of Visa does not decide that the world will make more digital transactions tomorrow. The world decides for itself.
In “The Freesurfer,” we called this the Free Growth Premium — what earlier posts in this series called the Gratuity Premium and what we now name more precisely — the observation that the discount rate for a Freesurfer should be lower than for a Growth Engine because execution risk is absent. We estimated a reduction of roughly one percentage point. That estimate was conservative. Far too conservative.
The Free Growth Premium is not a small adjustment to the WACC. It is the entire WACC. Every risk component that justifies the standard 8–10% discount rate — execution, competition, reinvestment, management — is absent or inverted in the A paramount. The premium is not 100 basis points. It is 400 to 500.
* * *
IV. What Buffett Does Without Saying
Warren Buffett discounts at the risk-free rate. Few people know this. The method sounds reckless — why would you apply the Treasury rate to an equity? But Buffett’s logic is precise: he only buys businesses whose cash flows are as certain as a Treasury’s coupon. If the business meets his standard of certainty, the risk-free rate is the correct discount rate. Everything above it is growth for free — the growing equity coupon that Graham described but never quantified.
This is the practical application. You take the free cash flow per share. You multiply it by the inverse of the risk-free rate — approximately 20× at current rates. That gives you the value of the A — the toll, priced as a perpetual bond. Any growth on top is gravy. Free gravy.
Applied to Visa in March 2026: free cash flow per share of approximately $19.39. Multiplied by 20: a fair value of approximately $388. Market price: $299. Under-valued by 23%.
Applied to Mastercard: free cash flow per share of approximately $19.78. Multiplied by 20: approximately $396. Market price: $498. Over-valued by 26%.
Same rail. Same wave. Same business. Radically different prices. The method tells you which side of the rail to stand on. And it tells you something deeper: at $299, you are paying for the toll alone. The permanent wave — the growth, the digitalization, the structural expansion of electronic payments across 85% of the world’s transactions that are still in cash — you receive for free.
The FCF × risk-free rate valorizes the A. The natural B is the gravy. And the gravy of the A paramount is the only gravy in finance that costs nothing to anyone.
But here is the problem. In the standard DCF framework, you cannot add a separate growth cell on top of the risk-free discount rate. The model forces you to choose: either you use a low discount rate (which inflates the present value of existing cash flows) or you model growth explicitly (which requires a higher discount rate to compensate for the uncertainty of that growth). The two levers work against each other. The spreadsheet cannot express a business where the discount rate should be low AND the growth is high AND the growth carries no risk. The tool was not built for this configuration.
This is why the simple formula — FCF × 20 — works where the spreadsheet fails. It values the toll at the risk-free rate and accepts the growth as a permanent, structural, free gift on top. No growth cell needed. The gift does not need to be modeled because it does not need to be predicted. It is the structure of the world.
* * *
V. The So-Called Risk-Free Rate
So far, we have argued that the A paramount deserves the risk-free rate as its discount rate — which already doubles its fair value compared to the standard WACC approach. But the argument does not stop there. Because the risk-free rate is not, in fact, risk-free.
The 10-Year U.S. Treasury carries real risks that no one calls “risks” because the instrument is defined as risk-free by convention, not by reality.
Inflation risk. The Treasury pays 4.2% nominal. Inflation runs at 3.5%. The real yield is 0.7%. If inflation rises to 5%, the real yield turns negative. The Treasury destroys purchasing power. Silently. Certainly.
Reinvestment risk. The 4.2% coupon must be reinvested. At what rate? No one knows. The Treasury does not compound naturally. Each coupon is a reinvestment decision — a mechanical B inside a “risk-free” instrument.
Sovereign risk. The United States carries $35 trillion in debt. The risk-free rate is guaranteed by the capacity of a government to tax and to print. This is not zero risk. It is risk that no one dares to name.
Duration risk. The 10-Year at 4.2% loses 15–20% of its value if rates rise to 6%. The “risk-free” instrument is volatile.
* * *
VI. The Negative Risk Premium
Now hold the Treasury’s risks in one hand. And hold the A paramount in the other.
Inflation? Visa benefits from it. Every transaction is worth more in nominal dollars. The toll is a percentage — inflation inflates the numerator automatically. Inflation is a tailwind, not a risk.
Reinvestment? The natural B reinvests itself. Digitalization pushes volume without a decision. No coupon to replace. No reinvestment risk. Compounding is automatic and structural.
Sovereign risk? Visa operates in 200 countries. If one government collapses, the other 199 continue to digitalize. The sovereign diversification of the A paramount is total — incomparably broader than a single government’s promise.
Duration? The A paramount has no maturity. It does not repay the principal in 10 years. It compounds indefinitely. Duration is infinite — and that is an advantage, not a risk.
The Treasury carries inflation risk, reinvestment risk, sovereign risk, and duration risk — and the world calls it risk-free.
The A paramount carries none of these risks — and the world applies a premium of 400–500 basis points above the so-called risk-free rate.
The world is upside down.
The true discount rate for the A paramount is not the risk-free rate. It is below it. The risk premium is not positive. It is negative. Not as a rhetorical provocation. Because the risks that define the risk-free instrument — inflation, reinvestment, sovereign concentration, duration — are absent or inverted in the A paramount. The A paramount is safer than the Treasury. It compounds faster. It diversifies better. And it does not depend on a single government’s ability to honor its debt.
* * *
VII. The Collision
The standard financial model rests on a fundamental assumption: risk and growth are correlated. More growth requires more reinvestment, more decisions, more execution risk. Therefore more growth demands a higher discount rate. Less risk means less growth means a lower discount rate. The two variables move together. The Treasury has zero growth and low risk. The startup has 50% growth and extreme risk. The spreadsheet trades one against the other. Always.
The A paramount decorrelates them. The risk is below the risk-free rate. The growth is 12–15% per year. And the growth is free. Two variables that the model assumes move together are moving in opposite directions. Low risk AND high growth. Simultaneously. The model has no cell for this configuration.
Consider the effects separately. The negative risk premium alone — a discount rate of 3% instead of 5% — increases the multiple from 20× to 33×. The model still functions. The cell accepts 3%. The valuation is generous but computable. If the Freesurfer had no growth, the sub-risk-free rate would simply produce a higher multiple. The spreadsheet would grind its teeth but survive.
The free growth premium alone — 12% perpetual growth at zero cost — with a standard discount rate of 8% already breaks the Gordon model: FCF ÷ (8% − 12%) = FCF ÷ (−4%). The denominator is negative. The value is undefined. But the analyst never lets this happen. He caps terminal growth at 3–5% precisely to prevent the model from breaking. He assumes the growth will slow because the model requires it to slow. The model constrains the reality instead of describing it.
Now combine both. A discount rate of 3% and a growth rate of 12%. FCF ÷ (3% − 12%) = FCF ÷ (−9%). The denominator is not just negative. It is deeply negative. The two errors do not add. They multiply. The negative risk premium pushes the discount rate down. The free growth premium pushes the growth rate up. Both forces widen the gap between the denominator’s two components — in the same direction. The model does not miscalculate slightly. It cannot calculate at all.
This is the collision of the two premiums. In every other asset class, the discount rate and the growth rate partially offset each other. Higher growth comes with higher risk. The denominator stays positive. The model functions. In the Freesurfer, the two variables are decorrelated. The risk falls while the growth rises. The denominator crosses zero and goes negative. The model was built for a world where risk and growth are married. The Freesurfer is the divorce. And the spreadsheet has no cell for divorce.
The Free Growth Premium is not the negative risk premium plus the free growth. It is the negative risk premium times the free growth. The error is multiplicative. And it compounds every year that the growth persists without cost while the model continues to apply a discount rate designed for a world where growth always costs something.
* * *
VIII. The Root
Why is the risk negative? Not merely low. Not merely absent. Negative. The answer lies in a concept introduced in the first chapter of The Infinite Investor: ergodicity.
A system is ergodic when what one person experiences over time matches what the group experiences at one moment. Most of finance is non-ergodic: your path matters, ruin is possible, and the average lies. A stock that falls 50% and then rises 50% has an average return of zero but a real return of minus 25%. The path destroyed wealth that the average concealed. This is the volatility tax. This is why survival comes first. This is why the book begins where it begins.
The mechanical B is non-ergodic. Every human decision is a point on the path where the path can collapse irreversibly. Kraft Heinz: $49 billion acquisition, $15 billion writedown, value destroyed permanently. A CEO who misallocates. An R&D cycle that produces nothing. An acquisition that destroys instead of creates. Each mechanical decision is a fork where the path can diverge catastrophically from the average. And the divergence does not self-correct. The writedown does not come back. The dead tree does not regrow. The path matters. The path is irreversible. This is non-ergodicity in its purest form.
The natural B is ergodic. The digitalization of tomorrow does not depend on a decision made today. The wave cannot make a mistake because the wave does not decide anything. There is no fork. There is no point of failure. There is no bad acquisition, no misallocation, no CEO who gets it wrong. The path is the path. The average is the path. The path is the average. The current flows in one direction — always, irreversibly — but the irreversibility works in favor of the holder, not against him. The digitalization does not reverse. The indexation does not reverse. The expansion of global credit does not reverse. The irreversibility of the natural B is the mirror image of the irreversibility of the mechanical B. One destroys. The other builds. Both are irreversible. But the direction is opposite.
This is why the risk premium is negative. The standard discount rate embeds a premium for path risk — the probability that the path will diverge from the average. In the A paramount, the path cannot diverge from the average because there are no decisions to create divergence. The path risk is not zero. It is negative — because the irreversibility of the natural B guarantees that the future is better than the present. The wave does not stop. The wave does not reverse. The wave does not make a mistake. The path is predetermined by the structure of the world. And a predetermined path that only goes up does not carry risk. It carries certainty. And certainty, in a world of non-ergodic instruments, deserves a premium — not a discount.
The book opens on non-ergodicity as the threat. The permanent wave closes on ergodicity as the destination. The vehicle that carries the investor from one to the other is the natural B. The Freesurfer. The reef. The permanent wave. The only asset whose path cannot betray the holder.
* * *
IX. The Spreadsheet That Cannot See
A discounted cash flow model has a cell for the discount rate. The cell accepts a positive number. It does not accept zero. It certainly does not accept a negative number.
This is not a flaw in the analyst. It is a flaw in the instrument. The DCF model was built for businesses whose growth carries risk. It was not built for businesses whose growth is a structural force of the world — free, permanent, and more certain than the government bond used as the baseline.
The model assumes that future cash flows are uncertain and must be discounted. But the future cash flows of Visa are more certain than those of the U.S. Treasury. There will be more digital transactions in ten years than today. That is more certain than the capacity of the U.S. government to honor its obligations in ten years.
The implication is vertiginous. A discount rate of 0% means each dollar of future free cash flow is worth a dollar today. A negative discount rate means each future dollar is worth more than a dollar today — because the future is more certain than the present. For the A paramount, this is not an abstraction. It is the structural reality that the spreadsheet cannot express.
And because the tool cannot express the reality, the market cannot price the reality. The margin of expansion persists. Not temporarily. Permanently. The undervaluation of the A paramount is embedded in the methodology of valuation itself.
A clarification is necessary here, because the blindness of the spreadsheet is often confused with a more familiar concept. Behavioral finance teaches that markets suffer from myopia — the tendency to undervalue distant cash flows because they are far away in time. This is real but it is not the problem. The discount rate exists precisely to correct for distance. Bringing the future to the present is the DCF’s job. It does that job, imperfectly but functionally. The lointain — the distant — is a problem of time. The spreadsheet manages time.
The invisible is different. The growth of Visa is not undervalued because it is far away. It is undervalued because it does not appear anywhere. The toll is visible — $19 of free cash flow per share, published quarterly. The margins are visible. The revenue is visible. Everything is visible except one thing: the cost of the growth is zero. And zero does not occupy a line in the spreadsheet. There is no row labeled “reinvestment for growth = $0.” There is an absence of a row. The spreadsheet does not undervalue something distant. It misses something present. The invisible is not a problem of time. It is a problem of structure. And no discount rate corrects for structural absence.
And the infinite is different still. Return ON infinity does not mean the return arrives infinitely far in the future. It means the return arrives now, every quarter, on an investment of zero. The word “infinite” in this framework is a word of architecture, not of time. MSCI does not compound at infinity in the future. MSCI compounds at infinity in the present — because the cost of the B is zero and division by zero gives infinity now. The dividend by zero is not a promise. It is a fact that the spreadsheet cannot express.
This is why the Free Growth Premium is permanent. The myopia bias closes when the future becomes the present — the distant cash flow arrives and the market reprices. The Free Growth Premium never closes because the invisible does not become visible with the passage of time. The zero-cost growth remains invisible next quarter, next year, next decade — because the spreadsheet will never add a row for something that costs nothing. The cause is structural. The effect is permanent.
* * *
X. The Embedded Margin of Expansion
Benjamin Graham invented the margin of safety — buy below intrinsic value, and the discount protects you if you are wrong. It is the foundational concept of value investing. But it is temporary. You buy at a discount. The market corrects. The discount closes. You must find another. Each margin of safety is an event — a momentary gap between price and value that must be replaced by the next.
The A paramount does not need a margin of expansion. It has one — embedded, structural, permanent.
The margin of expansion does not close because the tool that prices the asset cannot capture what the asset is. The DCF will continue to apply 8–10%. The WACC cell will not accept a negative number. The analyst will continue to embed execution risk in a business that carries none. The methodology is permanent. Therefore the undervaluation is permanent.
This is why Visa compounds at 15–18% per year indefinitely. It is not outperformance. It is the market perpetually catching up to an asset it cannot correctly price. And each time the market catches up, the natural B has already advanced. The market runs behind a train that accelerates. The PE rises from 25 to 30. The earnings grow. The PE falls back to 28. The earnings grow again. The PE rises to 31. The market is always late. Structurally. Perpetually.
Graham protects through price. The A paramount protects through physics. Graham’s margin is temporary. The A paramount’s margin is permanent. Graham demands that you find the next discount. The A paramount asks nothing. It compounds in the gap between what it is and what the spreadsheet says it is. And that gap never closes.
The negative risk premium and the margin of expansion are the same thing seen from two sides. The negative risk premium is the cause — the market applies 8–10% to an asset that deserves less than the risk-free rate. The margin of expansion is the effect — the gap between the price and the reality, stored in every share, dormant, invisible, structural.
Call it the latent margin of expansion. It is latent because it is always present, always waiting to be partially released, and always renewing itself after each release. When the market re-rates MSCI from PE 30 to PE 40, the latent margin is partially realized. But it immediately reforms at PE 40 — because the negative risk premium has not changed. The spreadsheet is still broken. The cell still cannot accept a negative number. The analyst still embeds execution risk in a business that carries none. The cause is permanent, therefore the effect is permanent.
This is why Visa compounds at 15–18% per year indefinitely. It is not outperformance. It is the latent margin releasing, quarter after quarter, year after year, as the market perpetually catches up to an asset it cannot correctly price. And each time it catches up, the natural B has already advanced. The latent margin reforms. The cycle restarts. The PE rises, the earnings grow into it, the PE appears to normalize, and the latent margin is already there again — waiting for the next release.
The latent margin of expansion is the negative risk premium stored in the price. And the stock never depletes.
* * *
XI. The Hierarchy of Certainty
A new hierarchy emerges — not of return, but of certainty:
| Asset | Discount Rate | Growth Cost | Inflation | Compounds |
| Treasury | 4.2% | Zero (no growth) | Erodes | No |
| Stalwart | 6–7% | $4B/yr (ads, R&D) | Neutral | Slowly |
| A + B Mechanical | 8–10% | Heavy (capex, M&A) | Neutral | If executed |
| A Paramount | ≤4.5% (negative premium) | Zero | Benefits | Always |
The Treasury is the floor of certainty by convention. The A paramount is the floor of certainty by physics. Convention can be revised — sovereign defaults happen, currencies collapse, governments change. Physics cannot be revised. The digitalization of the world’s transactions is not a policy decision. It is the direction of civilization.
* * *
XII. Why the Wave Is Permanent
Visa was built. Bank of America in 1958. Dee Hock restructuring the network in the 1970s. Decades of mechanical decisions — negotiating with banks, installing terminals, convincing merchants, regulating interchange fees. Every mile of the rail was laid by hand. For decades, Visa was a mechanical business. A toll with a mechanical growth engine.
And then — the tipping point. The network reached critical mass. Governments adopted cashless as policy. Banks had no choice. Merchants could not refuse. Consumers stopped carrying cash. The digitalization became a force of the universe, not a strategy of Visa. The mechanical dissolved into the natural. The builder set down the vase, and the vase began to grow on its own.
This is irreversible. The world does not return to cash. The rail does not get dismantled. The terminals do not get uninstalled. Even if Visa disappeared tomorrow, the world would continue to digitalize payments. The wave does not depend on the surfer. The surfer rides because the wave exists independently of it.
The analyst’s spreadsheet cannot see this. The discount rate cell cannot express this. The WACC model cannot contain this. And so the margin of expansion persists — structural, permanent, embedded in the gap between the tool and the reality.
The permanent wave does not ask to be priced correctly. It does not need to be. It compounds in the gap between what it is and what the spreadsheet says it is. And that gap — that embedded margin of expansion — is the gift that never closes.
* * *
Watch the Video
This post is part of the framework described in The Infinite Investor, available at averagingup.com.
Related posts: “The Freesurfer” — “The Hierarchy of Moats” — “The Attention Cost” — “Position Sizing: The Final Act of Conviction”