This essay turns from the architecture of government finance to the structure of taxation. The same pattern operates here. A provision that appears to be a neutral feature of the tax code — the absence of capital gains tax on inherited assets — is in fact a deliberate policy choice, worth roughly $72 billion in foregone federal revenue each year, that flows almost entirely to those who need it least. It has a name in the tax policy literature: the Angel of Death loophole. That name captures the mechanism but misses the structure. A loophole is something narrow and accidental. This is a benefit — deliberately awarded, architecturally embedded, and protected for a century by a label that has successfully named it out of existence.

The correct name is the Death Benefit.

I. The Label Does the Work

The term “death tax” is a piece of smart political framing. It implies that death triggers a new burden — that the government, at the moment of a family’s greatest loss, reaches in and takes a portion of what would otherwise pass intact to the next generation. The framing is powerful precisely because it is visceral: death as the occasion for government intrusion, the bereaved family as the victim. This framing has worked remarkably well. For decades it has suppressed public understanding of a provision that costs the Treasury roughly $72 billion annually — a provision that, properly understood, gives rather than takes.¹

Under current American law, death is simply not a taxable event. It is the occasion on which the government delivers its greatest financial gift to wealthy estates: the permanent forgiveness of a lifetime of accumulated, untaxed appreciation. The correct label is not the death tax. It is the Death Benefit.

The collision with life insurance terminology is deliberate. A life insurance death benefit is a payment earned by paying premiums over a lifetime — a contracted sum, modest relative to the wealth it protects, that provides a family some security after a loss. The tax code’s Death Benefit is something else: a permanent forgiveness of capital gains that may run to millions of dollars — or, for the largest estates, to hundreds of millions or more — earned not by either paying premiums or taxes but simply by dying before selling. Both are triggered by death. Both flow to heirs. Only one has a name in public discourse. The life insurance industry named its product. The tax code’s Death Benefit has successfully avoided being named at all — and a benefit without a name is a benefit that cannot be discussed, debated, or reformed in ordinary political conversation. We are left able to talk only about the thing that is named but does not exist — the death tax — while the thing that exists but has no name remains invisible.²

The academic and professional tax policy literature uses a different term: the Angel of Death loophole. That name circulates among tax lawyers, economists, and policy analysts — those who already understand the tax code’s intricacy and are debating its reform in specialized venues. It has never entered ordinary political discourse, and not only because it is too technical. “Loophole” is the wrong word. A loophole is something narrow and accidental, the product of inattention, something that a careful lawyer might find and exploit. The Death Benefit is none of those things. It is a $72 billion annual policy choice, embedded in the Internal Revenue Code since 1921, that has survived every serious reform effort for a century. It is not a loophole. It is an architecture. Calling it a loophole understates what it is. Calling it a death tax inverts its effect. The accurate name is the one that has been successfully suppressed: it is a benefit, awarded at death, to those fortunate enough to have been able to spend a lifetime accumulating what they did not need to sell.

II. How the Mechanism Works

The Death Benefit operates through a provision called step-up in basis. Understanding it requires understanding what “basis” means — which is less technical than it sounds.

When you buy an asset — a share of stock, a piece of real estate, an interest in a business — the price you paid is your basis. When you eventually sell, you owe capital gains tax on the difference between what you received and what you paid: the gain above your basis. If you bought stock for $100,000 and sold it for $400,000, your basis is $100,000, your gain is $300,000, and you owe capital gains tax on $300,000. This is how the tax is supposed to work for everyone. You pay once, on the gain, when you realize it.

Step-up in basis changes this calculation at death. When an asset passes to an heir, the heir’s basis is not the benefactor’s cost of the asset — it is the asset’s fair market value on the date of death. The heir “steps up” to the current value. If that same stock, purchased for $100,000, is worth $400,000 when the owner dies, the heir inherits it with a basis of $400,000. If the heir sells it the next morning for $400,000, the gain is zero. The $300,000 in appreciation — the entire economic gain the asset produced during the owner’s lifetime — has permanently vanished from the tax base. It will never be taxed. Not as capital gains. Not as income. Not as anything.

Three things happen simultaneously at death: the basis resets to fair market value, the lifetime gain permanently escapes taxation, and the heir can sell immediately without owing a dollar in capital gains tax. The government asks for nothing. The appreciation disappears.

The heart attack example

Consider two investors with identical investment portfolios — each holds stock purchased years ago for $1 million, now worth $10 million. Alice sells on Monday morning. She owes capital gains tax on $9 million in appreciation — roughly $1.8 million at current rates. Her daughter inherits $8.2 million.

Bob planned to sell on Monday, but has a heart attack Sunday night, before selling. His daughter inherits the entire $10 million with a stepped-up basis of $10 million. She sells Monday morning and owes zero capital gains tax.

The $1.8 million difference between these two families was not due to any difference in their economic circumstances, their investment skill, their contribution to society, or their need. It was due entirely to which investor died before executing the trade. The tax code does not treat these families differently because their situations differ. It treats them differently because one of them died at the right moment.

For the properly positioned estate, dying is not a cost. It is a windfall.³

The history of the provision

Step-up in basis is not an ancient feature of American tax law. It was introduced in the Revenue Act of 1921 — only five years after the modern estate tax became law in 1916. Congress’s purpose was explicit and, in its original context, defensible: an asset that had just been subjected to estate tax at its full fair market value should not also face capital gains tax on the previously taxed appreciation when the heir sells. The basis was reset to fair market value to prevent double taxation. Step-up was not a gift. It was a receipt — the accounting acknowledgment that the appreciation had been taxed once, at the estate level, and should not be taxed again.

That logic was sound for any estate actually subject to estate tax — step-up prevented the appreciation from being taxed again after the estate had already paid tax on it. The problem is that step-up was written as a universal rule, applying to every estate regardless of whether estate tax was paid. In 1916, the estate tax reached 0.454% of the population — a small but real fraction of wealthy estates for whom the double-taxation concern was genuine. At its peak coverage in 1976, it applied to roughly 7% of all deaths. As long as the estate tax covered a substantial fraction of estates, the mismatch between the universal rule and the conditional rationale was modest. The subsequent decades brought a series of legislative increases to the exemption threshold: from $600,000 in the 1980s through the Bush-era escalations of the early 2000s, to the Tax Cuts and Jobs Act of 2017, which doubled the exemption to roughly $11 million per individual. That increase was originally scheduled to sunset after 2025; the One Big Beautiful Bill Act made it permanent. Today the exemption stands at $13.99 million per individual — and the estate tax reaches fewer than one estate in a thousand, 0.08% of adult deaths. The universal rule has become a near-universal gift: the original rationale applying to almost no one, the benefit flowing to almost everyone.

The provision Congress designed as a receipt for taxes paid has become, for 999 estates in a thousand, a pure tax forgiveness — the Death Benefit — awarded where no estate tax was paid and no complementary tax exists to justify it.

Preventing double taxation has become no taxation.⁴

The dynastic engine

For most estates, the Death Benefit is a one-time, windfall event: a lifetime of appreciated gains forgiven at a single death, the heir receiving a fresh basis, the cycle ending there. But for estates large enough to sustain a specific financial strategy, the Death Benefit becomes something more: the mechanism that powers a dynastic engine that can run without interruption across generations.

The strategy is straightforward, and it requires no offshore accounts, no complex trust structures, no aggressive legal position. It requires only a portfolio large enough to sustain low-cost collateralized borrowing, and an appreciation rate that exceeds the cost of borrowing against it.

A family holding a large diversified portfolio — equities, real estate, private investments — borrows against those holdings rather than selling them. The borrowed money funds consumption. Interest on the loans is serviced from the portfolio’s dividend and interest income, from modest additional borrowing against growing collateral, or from minimal asset sales. Lenders are comfortable with this arrangement because the portfolio’s appreciation rate keeps the loan-to-value ratio well within safe limits regardless of how long the loan runs — the collateral grows faster than the debt. The portfolio continues to appreciate. No large realization event is ever required.

At death, the accumulated appreciation is permanently forgiven through step-up. The portfolio passes to the next generation with a fresh basis — every dollar of lifetime appreciation extinguished from the tax base. The heirs retire any outstanding loans from sales of stepped-up assets, paying zero capital gains tax on those sales because the basis equals the sale price. They inherit the remaining portfolio at its new, reset basis and restart the cycle.

The government is not entirely shut out. It collects income tax on the dividends and interest the portfolio generates — the modest cash yield that even a growth-oriented family cannot fully eliminate. What it will not collect is tax on the appreciation, typically the larger component of total return by a substantial margin. For a portfolio constructed to minimize yield — low-dividend equities, real estate held for appreciation rather than income — the taxable fraction of total economic return may be a small portion of actual wealth accumulation. The tax code taxes the trickle. The flood is tax free.

This is the complete borrow-spend-die strategy: borrow against appreciating assets, service the interest from portfolio income, retire the principal at death from stepped-up sales that generate no taxable gain, and repeat across generations. It is not exotic. It is the straightforward rational response to the incentive structure the Death Benefit creates. The strategy has a simpler name in the parlance of old money: never touch the principal. Wealthy families have passed this maxim down for generations — live off the income, borrow against the growth if needed, and preserve the corpus intact for your heirs. The Death Benefit is what transforms that maxim from prudent advice into a tax-optimized dynasty engine: without step-up, the preserved principal would eventually face capital gains tax; with it, the corpus resets at each generational transfer, permanently untaxed, and the next generation begins the cycle anew with a clean slate. The tax code has made dying, for those with enough wealth to hold until it happens, among the most tax-advantaged wealth-transfer mechanisms available under American law.⁵

Step-up and the dynamics of concentration

Piketty’s central observation in Capital in the Twenty-First Century is that when the return on capital persistently exceeds the economy’s growth rate — r > g — wealth concentration compounds across generations without self-correcting. Dynastic strategies operate within this environment and amplify it in a specific way. A large diversified portfolio tends to earn above-average returns — scale provides access to asset classes, co-investment opportunities, and terms unavailable to smaller investors, a pattern Fagereng and colleagues document empirically across wealth levels. Meanwhile, the dynasty’s borrowing cost sits near the risk-free floor, because its large liquid collateral makes default risk negligible for lenders. The dynasty therefore captures a spread between what it earns and what it pays to borrow that is wider than what ordinary investors can access — compressed in its favor from both sides simultaneously. Step-up then ensures that spread compounds without fiscal drag across every generational transfer, converting a structural pre-tax advantage into a permanent post-tax advantage by extinguishing the accumulated appreciation on which every other investor would eventually pay capital gains tax.⁶

III. Who Gets the Benefit

The Death Benefit is not distributed neutrally across the population of asset holders. It flows overwhelmingly to those who already have the most — and within that group, it flows most clearly to those whose estates fall below the threshold at which the estate tax begins to provide even a partial justification for it.

The basic distributional facts

Close to half of all estate value at death consists of unrealized capital gains — gains that were never taxed during the decedent’s lifetime and, under step-up, will never be taxed at all. According to the Congressional Budget Office, in 2019 the top 20% of decedents’ estates received 56% of the total step-up benefit, with the top 1% alone receiving 18% of the total. Among households receiving an inheritance, those with economic income already over $1 million — before the inheritance — expected to inherit an average of $3 million; those with economic income under $50,000 expected to inherit an average of $62,000.⁷ The 48-to-1 ratio between those figures is not a description of different levels of need. It is a description of how the Death Benefit distributes: to those who already have the most, compounding advantage upon advantage.

The provision labeled a burden — the estate tax, the so-called death tax — touches almost no one. Taxable estate tax returns represented 0.08% of adult deaths in 2019, down from a peak of roughly 7% in 1976. The provision that awards a benefit — step-up in basis, the Death Benefit — flows to a far larger universe of estates carrying appreciated assets below the exemption threshold. The label that dominates public discourse describes a tax that almost no one pays. The benefit that has no label in public discourse flows to estates at every level of the upper wealth distribution, concentrated most heavily at the top.

Two tiers of benefit

The Death Benefit operates differently at different wealth levels, and the difference matters for understanding who it truly serves.

For the merely wealthy — families with $2 million to $10 million in appreciated assets, solidly above the thresholds at which most Americans live but well below the ultra-wealthy — the Death Benefit is usually a single-generation event. A lifetime of appreciated stock, a home purchased decades ago for a fraction of its current value, a small business built and never sold: these pass to heirs with a stepped-up basis, the appreciation permanently forgiven, the heir free to sell without paying the capital gains tax that their benefactor would have owed. The benefit is real and substantial. But it is usually a one-time gift at their benefactor’s death. The merely wealthy do not generally have portfolios large enough to sustain or create the dynastic engine described in the previous section. They receive the Death Benefit once, per generation.

For the ultra-wealthy — families with portfolios large enough to execute the borrow-spend-die strategy across generations — the Death Benefit is not a single-generation event. It is a perpetual wealth-compounding machine. Each generation borrows against appreciated assets, services the debt from portfolio income, dies with the appreciation intact, and passes a fresh-basis portfolio to the next generation to restart the cycle. The benefit is not merely forgiven once. It is forgiven at every generational transfer, compounding the advantage across time without limit and without any fiscal interruption.

The concentration dynamic and its trajectory

The Saez-Zucman wealth distribution data illuminate why this distinction matters. The share of total wealth held by the top 0.1% has returned to levels last seen in 1929 — nearly 20% of all household wealth in the hands of roughly 160,000 families. Meanwhile, the share held by the “merely rich” — families in the top 10% but below the top 1% — has actually declined slightly over the past four decades, even as their absolute wealth has grown. The ultra-wealthy are pulling away not merely from the middle class but from the merely wealthy as well.⁸

This divergence will deepen. The tax policy changes of recent decades — steadily raising the estate tax exemption, making step-up permanent, and most recently cementing both under OBBBA — may have been intended primarily to confer immediate tax advantages on wealthy donors and constituents. But their long-term structural consequence may prove far more profound than their authors anticipated. The dynastic engine the Death Benefit enables does not merely preserve existing wealth. It compounds it, generation after generation, immune to the market competition and the institutional checks that prevent accumulation from hardening into permanent dynasty.

Today’s extreme fortunes — unprecedented in scale, the product of a specific era of technological and financial concentration — are beginning their multigenerational transmission through a tax arrangement specifically designed, whether intentionally or not, to make that transmission permanent. Step-up is not the only mechanism producing this outcome — housing appreciation, financial concentration, network monopolies, and educational assortativity all play roles — but it is the one examined here, and it is the one that has successfully hidden behind a name that describes something else entirely.

There was a time when a billionaire was considered a financial curiosity — a figure so remote from ordinary experience as to seem almost fictional. Today billionaires number in the thousands, and the world’s first trillionaire may arrive within this decade. The scale is worth pausing on. A single billionaire controls wealth equivalent to one thousand millionaires. A single trillionaire controls wealth equivalent to one million millionaires. When a trillionaire’s wealth grows by $100 billion — as has happened repeatedly in recent years — that single person’s gain equals the entire accumulated wealth of 100,000 millionaires. No corresponding number of millionaires was created. The wealth did not distribute. It concentrated further.

The income alone from such a portfolio is staggering. At a modest 5% annual yield, a trillion-dollar portfolio generates $50 billion each year — enough, at $1 million annually, to support 50,000 families in comfort without selling a single asset. With borrowing against the appreciated principal, even more could be supported — and still without triggering a taxable event. That income stream, taxed as received, is the trickle. The appreciation that step-up permanently forgives at each generational transfer is the flood.

At each generational transfer, the Death Benefit ensures that the concentrated appreciation at the apex will escape taxation entirely, while the merely wealthy below the apex receive a one-time forgiveness that does not carry the compounding dynastic advantage available only at the top.

Yet even for the largest estates — those that do owe estate tax on the value above the exemption threshold — step-up applies in full to the entire appreciated portfolio. The estate tax is a partial, one-time charge on value above the threshold. The Death Benefit is a permanent, recurring exclusion of appreciation that operates regardless of estate size. The effective estate tax rate on large estates averages well below the 40% statutory rate, given the valuation discounts, family limited partnerships, and charitable vehicles available at that scale. And at every generational transfer, whatever the estate’s size, the first $13.99 million in appreciated assets passes with a fully stepped-up basis and no capital gains tax. No estate tax is owed on that tranche. No capital gains tax will be owed when the heir sells. The appreciation accumulated during the previous generation on that exempt portion is permanently excluded from taxation — not deferred, not reduced, permanently excluded — at every death, in every generation.

For married couples using both spouses’ exemptions, as standard estate planning provides, the combined threshold rises to $27.98 million. Consider a $55 million estate divided equally between two married couples — $27.5 million to each. Each couple’s combined exemption covers their entire share. No estate tax is owed. No capital gains tax will be owed when the heirs sell. The entire $55 million passes as a pure Death Benefit, all appreciation permanently excluded from taxation. An estate that most readers would assume comfortably triggers estate tax escapes it entirely — and each heir couple restarts the borrow-spend-die cycle with a fresh, stepped-up basis.⁹

The result, projected forward across generations, is not merely a wealthy elite but a dynastically entrenched one — an economic class whose position is not merely the product of market success but of a legal arrangement that makes that position self-perpetuating. The United States has seen great fortunes before. What it has not seen — what the combination of extreme concentration and this self-perpetuating wealth machine may now be creating — is a class of families whose economic dominance is structurally secured against the forces, including taxation, that have historically interrupted the transmission of dynastic wealth.¹⁰

IV. The “Taxed Twice” Objection Fails

The most common objection to reforming step-up in basis is that the money was already taxed. The argument runs: the decedent earned income, paid income tax on it, invested the after-tax proceeds, built wealth over a lifetime — and now the government wants to tax it again at death. Double taxation. The family is being penalized twice for the same dollars.

This objection has genuine emotional force. It also fails on its own terms — not as a matter of policy preference but as a matter of arithmetic.

What was actually taxed

Consider the numbers. A person invests $1 million of after-tax income — income on which they have already paid income tax, as the objection correctly notes. That investment appreciates to $10 million during their lifetime. The estate, well below the $13.99 million exemption threshold, owes no estate tax. It passes to the heir with a stepped-up basis of $10 million. The heir sells and pays zero capital gains tax.

What was taxed? Not the original $1 million investment — the after-tax income the decedent earned and invested — it was indeed taxed once, as income, when it was earned.

What was never taxed? The $9 million in appreciation. Not once. Not ever. It accumulated tax-free throughout the decedent’s lifetime — unrealized, untouched, compounding without fiscal interruption — and was permanently excluded from capital gains taxation at death through step-up. The $9 million is not money that was already taxed. It is income that was never taxed at all, but would have been taxed if realized before the benefactor’s death.

The double-taxation objection could apply, at most, to the original $1 million. It has no application whatsoever to the $9 million in appreciation, which is the entire subject of the Death Benefit. Eliminating step-up and replacing it with deemed disposition — taxing the gain at death as if the decedent had sold — would not double tax the original $1 million. It would tax the $9 million once, for the first time, which is all anyone else’s income is taxed.

The objection’s narrow legitimate application

The double-taxation concern does have a narrow legitimate application. For estates large enough to owe estate tax — those above the $13.99 million exemption per individual — the appreciated value of the estate is subject to estate tax at rates up to 40%, only slightly above the 37% ordinary income rate that high-earning wage earners pay on their last dollar of labor income. The estate tax threshold is a gross fair market value test, not a net gain test. An asset purchased for $1 million and now worth $14 million is valued at $14 million for estate tax purposes, with the $13 million in appreciation fully included regardless of what the decedent originally paid. If death were also treated as a realization event for capital gains, the same appreciation would face both estate tax and capital gains tax. That is a genuine design question about the coordination of two taxes on the same economic gain.

The concern is real but narrow. Those subject to potential double-taxation constitute a vanishingly small population — fewer than one estate in a thousand, 0.08% of adult deaths — and for the remaining 99.92%, no estate tax is paid, no coordination problem arises, and the Death Benefit is a pure exclusion of appreciation from taxation with no offsetting tax of any kind. Moreover, the coordination problem can and has been solved: Canada’s deemed disposition regime credits the decedent’s original cost basis against the deemed capital gain, taxes only the appreciation above that basis, and coordinates with any estate-level taxes — taxing the gain once, not twice, at the higher of the applicable rates. A deemed disposition regime in the United States could do the same, in effect restoring step-up to its original 1921 function as a receipt for taxes paid rather than a freestanding reward for dying. The design question merits a solution; it is not a reason to preserve a $72 billion annual benefit for our nation’s wealthiest heirs.

Most importantly, the double-taxation objection as deployed in public debate is almost never about this narrow and rare coordination problem. It is deployed as a general argument against any taxation of inherited wealth, applied indiscriminately to estates of all sizes, the vast majority of which pay no estate tax and face no coordination problem at all. The objection is invoked for the many to protect the few, obscuring the fact that for most estates the question is not double taxation but no taxation.¹¹

The asymmetry with wage income

The appreciation that generates the Death Benefit was also income — economic gain accruing to the asset holder over time. It was simply not realized as taxable cash income. Under a comprehensive income tax, the form of the income should not determine its taxability: a dollar of wage income and a dollar of appreciation are both increases in the holder’s economic position, and both should be taxed only once. Step-up ensures that one of them is taxed once and the other is never taxed at all. A proper concern for double taxation, applied consistently, would require that no single economic gain be taxed more than once as a capital gain — which is precisely what deemed disposition achieves. Applied selectively, as it is in practice, it serves as a justification for taxing neither.¹²

V. The Wage Earner Has No Equivalent

The Death Benefit is sometimes described as a provision that affects only the very wealthy — a narrow elite whose tax treatment is remote from the experience of ordinary Americans. The distributional data in Section III confirms that the benefit is heavily concentrated at the top. But the structural argument cuts deeper than distribution. The Death Benefit does not merely give wealthy estates a better outcome than wage earners. It creates a parallel tax universe — one in which the fundamental rules governing how income is taxed operate entirely differently depending on how that income was earned and how it is held.

The one-seventh ratio

The aggregate measure of this disparity is striking. Inherited income — income received through bequest and inheritance — is taxed at less than one-seventh the average effective tax rate on income from work and savings.¹³ That ratio is not the product of a single provision or a single policy choice. It is the cumulative result of a tax architecture that systematically advantages inherited wealth over earned income at every point: preferential capital gains rates during life, the deferral of tax on unrealized appreciation, and the permanent forgiveness of that appreciation at death through step-up. The Death Benefit is the final and most consequential layer of that architecture — the provision that converts a tax preference into a permanent exemption.

For a family executing the dynastic strategy across generations, the effective rate is not one-seventh. It is zero on the appreciated portion of wealth — the component that typically dominates total return. The one-seventh figure is a system-wide average that includes less sophisticated estates and assets that generate taxable income along the way. The full strategy — holding appreciating assets, borrowing to consume, dying with a stepped-up portfolio, restarting the cycle — produces an effective capital gains rate of zero, in perpetuity, across every generation willing and able to execute it. The one-seventh is the average. Zero is the floor, available to those with sufficient wealth to reach it.

What the wage earner cannot do

The contrast with wage income is not merely quantitative. It is structural. A wage earner has no equivalent mechanism. They cannot defer their income until a more favorable moment. They cannot arrange for a lifetime of earnings to escape taxation by the simple act of not spending them. They cannot borrow against their future labor income at favorable rates, service the interest from the income their labor generates, and arrange for the entire principal to be forgiven at death. Every dollar of labor income is taxed in the year it is received, at ordinary rates, without exception.

The appreciation that generates the Death Benefit is economically equivalent to wage income in one fundamental respect: it represents an increase in the holder’s economic position. A worker who earns $100,000 in wages and an investor whose portfolio appreciates by $100,000 are both $100,000 wealthier at the end of the year. The worker pays income tax on their $100,000. The wealthy investor pays nothing — not this year, not next year, and under step-up, not ever. The tax code treats these two identical increases in economic position as categorically different events: one taxable, one permanently exempt. The only relevant difference between them is that one was earned through labor and the other through the passage of time and the ownership of capital.

VI. Other Objections

The family farm and forced sale objection

No discussion of step-up reform is complete without addressing the family farm — the most emotionally powerful argument against taxing gains at death. The argument runs: a family holds farm land that has appreciated enormously over generations. If death triggers a capital gains tax, the heirs will be forced to sell the farm to pay the bill. The family loses the land. The community loses the farm. Reform destroys what generations built.

This argument has genuine human weight behind it. It deserves a direct answer, in two parts.

The first part is empirical. There appears to be no documented evidence that any family farm in the United States has ever been sold to fund federal estate taxes.¹⁴ The forced-sale story is politically powerful precisely because it is vivid and sympathetic — but it describes something that does not appear to happen in practice, because the farms it invokes are typically below the estate tax threshold, qualify for existing installment payment provisions, or are transferred within families through mechanisms that defer or reduce the tax liability regardless of reform. The family farm argument has been the most durable objection to estate and inheritance tax reform for decades. It has not been substantiated by a single documented case.

The second part is structural. The forced-sale concern attaches primarily to the estate tax — a tax on the total value of the estate, payable in cash shortly after death. A deemed disposition regime is different in a critical way: it is a capital gains tax on the appreciation, not a tax on the total value, and it can be structured to avoid any forced sale entirely. Canada’s system — the most directly comparable deemed disposition regime — handles farm transfers through a rollover provision: farm property transferred to a child carries over the decedent’s original cost basis, deferring any capital gains tax until the heir eventually sells. No tax is due at death. No sale is required. The gain is deferred, not forgiven — the heir will eventually pay when they sell — but the farm stays in the family for as long as the family chooses to keep it.¹⁵

The existing US tax code already contains an analogous provision: installment payment of estate taxes over fifteen years for qualifying farm and closely held business interests. A deemed disposition regime could extend similar treatment to capital gains triggered at death — payable over time, not immediately, with no forced sale required. The administrative tools exist. The forced-sale concern is a design question, not a fundamental objection to taxing gains at death.

The administrative objection

A related objection concerns administrative complexity. The one serious US attempt to move away from step-up — the carryover basis reform enacted in 1976 — was repealed in 1980 before it took effect, after the Joint Committee on Taxation found that compliance required reconstructing the original purchase price of assets potentially held for decades, often without contemporaneous records. The administrative burden was real and the reform was genuinely unworkable as designed.

But this history argues against carryover basis specifically, not against taxing gains at death generally. The 1976 reform failed because it required heirs to reconstruct a dead person’s historical records — an archaeological exercise that was often impossible and always burdensome. Deemed disposition avoids this problem entirely. Under deemed disposition, the taxable gain is the difference between the fair market value at death — a current appraisal, not a historical reconstruction — and the decedent’s adjusted cost basis, which the decedent maintained on their own tax records throughout their lifetime as a living taxpayer. The administrative burden falls on the living taxpayer maintaining their own records — the normal expectation of any taxpayer — rather than on heirs reconstructing records that may not exist. For financial assets, the problem is further diminished by the mandatory cost-basis reporting rules that have applied to brokerages since 2011: the living taxpayer’s basis is already tracked electronically in the brokerage system. The primary remaining appraisal challenge is real estate and closely held business interests — significant, but a far narrower problem than the wholesale basis reconstruction that defeated the 1976 reform.

Canada introduced deemed disposition in 1972 and has operated it for over fifty years without the administrative collapse US opponents predict. The 1976 failure is the history of a specific, poorly designed reform. It is not the history of the approach the essay is describing.¹⁶

VII. Tax Neutrality, Not a New Tax

The often-heard suggestion that step-up reform would impose a new tax on inheritance frames such reform as an inequitable “death tax” in a slightly different form. The preceding sections have established that this framing is precisely backwards. The current system is not neutral — and death is not a non-event for taxes. It actively advantages dying over selling, holding over deploying, and inheriting over earning. Reform would not impose a new tax. It would remove an existing subsidy.

The minimal policy argument is this: death should be tax-neutral; neither tax-favored nor tax-punished. The wage earner who receives $1 million in labor income pays income tax on $1 million. The investor who realizes $1 million in capital gains pays capital gains tax on $1 million. The investor who dies with $1 million in unrealized gains pays nothing. Tax neutrality requires that the third outcome be treated like the second — taxed once, at the applicable rate, like everyone else’s income.

The lock-in objection

Step-up is sometimes defended on the ground that it corrects for a distortion the capital gains tax itself creates: the lock-in effect. The capital gains tax gives investors an incentive to hold rather than sell appreciated assets, because selling triggers an immediate tax liability while holding defers it indefinitely. This distortion has real economic costs — capital that would be more productively deployed elsewhere stays locked in existing positions to avoid the tax. Step-up, on this argument, corrects for lock-in by ensuring that at least at death the slate is wiped clean, freeing heirs to redeploy inherited assets without the tax friction that would otherwise deter them.

This argument has a surface plausibility that dissolves on inspection. Step-up does not reduce lock-in. It maximizes it — in a specific and perverse way that the lock-in defense never acknowledges.

Under current law, the rational strategy for any investor holding appreciated assets is unambiguous: never sell. Selling during life triggers capital gains tax immediately, reducing the compounding base by the tax paid. Holding until death triggers the Death Benefit, permanently excluding the gain from taxation. The optimal exit is not any sale during life — it is death. Step-up does not merely fail to solve the lock-in problem. It converts lock-in from a reluctant response to a tax friction into an affirmative wealth-maximizing strategy. The investor who holds rather than sells is not being irrational or sticky. They are executing the dominant strategy the tax code offers.

Eliminating step-up removes this incentive entirely. An investor facing deemed disposition at death — the same capital gains tax whether they sell today or hold until death — has no tax reason to prefer death as the realization event. They can make portfolio decisions on the merits: sell when the capital is better deployed elsewhere, hold when it is not. The lock-in step-up produces is not an incidental byproduct of a well-intentioned provision. It is the predictable equilibrium response to a system that makes dying the optimal financial transaction. Removing the Death Benefit removes the distortion.¹⁷

The international comparison

The United States is not alone in having step-up in basis — it is in fact the most common approach among OECD countries that also levy estate or inheritance taxes. But the US combination of unlimited step-up and a near-universal estate tax exemption is distinctive, and the OECD has said so directly. In its 2021 report on inheritance taxation and its 2025 working paper on capital gains, the OECD specifically identified step-up as a provision that should be reconsidered where estate tax exemption thresholds are very high — a description that fits the United States more precisely than any other member country.¹⁸

In countries that retain step-up alongside a functioning inheritance or estate tax, the provision has a coherent rationale: the appreciation was taxed at the estate level, and step-up prevents it from being taxed again at the capital gains level when the heir sells. The step-up is a receipt for taxes paid, not a freestanding benefit. In the United States, where the estate tax exemption has been raised to the point that fewer than one estate in a thousand pays any estate tax, step-up has been severed from the tax it was designed to complement. It is no longer a receipt. It is a gift.

Other countries have taken a different path. Canada, when it introduced capital gains taxation in 1972, deliberately chose deemed disposition over step-up — treating death as a constructive sale at fair market value, taxing the gain once, and giving the heir a fresh basis equal to the value on which tax was paid. The Carter Commission, whose report formed the basis for the Canadian reform, expressed the governing principle in a phrase widely attributed to its chair, Kenneth Carter, that has become foundational in Canadian tax policy: “a buck is a buck is a buck.” Income is income regardless of its form, and the tax system should treat it accordingly. Death is a realization event, not an occasion for permanent forgiveness.¹⁹

Australia, when it introduced capital gains taxation in 1985, chose carryover basis rather than step-up — deferring the tax until the heir sells rather than forgiving it at death. Denmark and Hungary treat death as a full realization event, consistent with a comprehensive capital gains tax system. New Zealand’s Tax Working Group, when it considered introducing a capital gains tax in 2019, specifically examined and rejected the US step-up model as inconsistent with the principle of taxing economic gains comprehensively.

The international consensus among countries that have thought carefully about this question is that step-up is the wrong answer. Countries that have chosen differently — Canada most prominently — have done so on the principled ground that income should be taxed once regardless of how it is realized, and that the tax system should not make dying the most tax-advantaged financial transaction available.

The revenue picture

The scale of the Death Benefit is not incidental to the policy argument. According to the Congressional Budget Office’s December 2024 budget options report, replacing step-up with carryover basis — deferring the tax until the heir sells rather than collecting it at death — would reduce the deficit by approximately $197 billion over ten years. Replacing step-up with deemed disposition — taxing the accrued gain at death as if the decedent had sold — would reduce the deficit by approximately $536 billion over the same period.²⁰

These are not small numbers. They represent tax that other people — wage earners, savers, investors who sell during their lifetimes, small business owners who realize their gains rather than holding until death — already pay on equivalent economic gains. The Death Benefit does not merely forgive the appreciation in wealthy estates. It shifts the tax burden onto those who cannot or do not execute the holding strategy: those without sufficient wealth to sustain the cycle, those who need to sell to fund retirement or education or a business, those who simply did not have the good fortune to die at the right moment. Every dollar of appreciation permanently excluded at death is a dollar that the existing capital gains rate must recover from a narrower base of realizations. The Death Benefit is not simply a subsidy to wealthy estates — it is a subsidy paid in part by every investor who sells during their lifetime.²¹

Both reforms represent a fundamental improvement over the current system, and both are worth supporting. Deemed disposition is the more equitable approach: it taxes each generation’s accumulated appreciation at each generational transfer, ensuring the obligation is collected while the gain is still economically present. Carryover basis is less immediately equitable — it permits further deferral — but it ends the permanent forgiveness that is the Death Benefit’s defining feature. The tax obligation follows the asset rather than being extinguished at death.

That distinction matters, however, because dynasties will exploit carryover basis as effectively as they exploit step-up. A family executing the borrow-spend-die strategy never sells. Under carryover basis, the basis carries forward from decedent to heir, then to grandheir, then to great-grandheir — across as many generations as the family continues to hold. The embedded tax liability grows on paper but is never triggered. And crucially, even if some distant generation eventually sells, the time value of money has steadily eroded the real value of what is collected. A capital gains tax on appreciation that accrued decades ago, collected three generations later, may be worth only a small fraction of its face value in present terms. The obligation survives; its economic substance diminishes with each passing generation.

Carryover basis is therefore not a consolation prize but a serious reform — one that would end the dynasty strategy’s most essential advantage for ordinary estates and create at least the prospect of eventual collection even for the largest. If the administrative and political challenges of deemed disposition prove prohibitive, a well-designed carryover basis regime, combined with targeted relief for small estates and illiquid assets, would represent genuine progress. An objection to deemed disposition that converts a critic into a strong supporter of carryover basis is a good outcome for the reform project. The enemy of equity here is not carryover — it is the permanent forgiveness the current system provides.

The pro-capitalist case for reform

This is the point at which the standard framing of step-up reform — as a tax increase on the wealthy, a burden on bereaved families, an attack on the right to pass wealth to one’s children — is most completely inverted.

A tax code that systematically taxes the act of accumulation more heavily than the act of inheritance is not neutral with respect to capitalism’s promise. It is anti-capitalist in a specific and precise sense: it advantages hereditary wealth over productive accumulation, dynastic transmission over the creation of new value, the accident of birth over the exercise of talent and effort. Capitalism’s distinguishing commitment — the universal right to accumulate, open to anyone regardless of birth — is undermined by a provision that makes the return on inherited capital higher, after tax, than the return on newly created wealth. The Death Benefit does not merely distribute unfairly. It tilts the playing field against the very behavior capitalism is supposed to reward.

Reforming it — replacing the Death Benefit with a system that taxes appreciation once, at death or at subsequent realization, like everyone else’s income — is not an attack on capitalism. It is a defense of capitalism. It removes a subsidy to hereditary transmission that the system’s own logic cannot justify. It restores the principle that income is income regardless of its form. And it eliminates the incentive structure through which today’s extreme fortunes are converting themselves into a dynastically entrenched economic class — a class whose position is secured not by continued productive contribution but by the compounding of an inherited advantage that the tax code makes permanent.²²

VIII. The Label Inversion

The essays in this series share a structure. Each identifies a fiscal fiction or framing — a legal or accounting artifact that presents a governance choice as an economic necessity, a policy decision as a natural feature of the landscape. The Social Security Trust Fund is not a savings account; it is a political commitment dressed in the language of pre-funding. The debt ceiling is not a fiscal constraint; it is a political tool that was repurposed into a weapon. The government’s ability to create money is not what the standard story describes; the constraints are legal and political choices, not laws of nature.

The Death Benefit follows the same pattern, with one additional feature: it is protected not merely by a fiction about what it is but by a label that names something else entirely. The Trust Fund fiction tells a false story about a real provision. The death tax label tells a true story about a provision that does not exist — and in doing so, makes the provision that does exist invisible.

The death tax does not exist for 999 estates in a thousand. The Death Benefit exists for all of them. But only one has a name in ordinary political discourse, and therefore only one can be retrieved, argued about, defended, or reformed in ordinary political conversation. The label does not merely misdescribe the provision — it occupies the cognitive space where the correct description should be, making it nearly impossible to have the right conversation because the right conversation has no linguistic handle. We are left able to talk only about the thing that is named but does not exist, while the thing that exists and costs the Treasury $72 billion annually remains, for most citizens, simply invisible.²³

Naming the benefit is the first act of reform. Not the legislation, not the revenue estimate, not the international comparison — the name. The Death Benefit. A benefit, awarded at death, to those fortunate enough to have been able to spend a lifetime accumulating what they did not need to sell. A benefit with no counterpart in the tax treatment of wage income, no justification in the original logic of the provision that created it, and no defense in the principles of a tax system that claims to treat income equally regardless of its form. A benefit that has survived every serious reform effort for a century because its defenders have successfully prevented the merits from being examined — by ensuring that the benefit has no name in ordinary political discourse.

What the system actually does

Step back from the details — the basis calculations, the deemed disposition mechanics, the estate tax coordination questions — and consider what the current system actually does, stated plainly.

It taxes the wages of a nurse at ordinary income rates, in the year they are earned, without deferral or forgiveness.

It taxes the capital gains of a small investor who sells appreciated stock to fund their retirement, in the year of the sale, at preferential but real rates.

It permanently exempts the lifetime appreciation of a large investment portfolio, accumulated tax-free across decades of compounding, from any capital gains tax — provided only that the holder dies before selling.

It then enables the heirs of that portfolio to borrow against their fresh-basis inheritance at rates unavailable to ordinary borrowers, service the borrowing from portfolio income, and repeat the entire cycle — generation after generation, without any capital gains tax at any generational transfer, in perpetuity.

This is not a neutral system. It is a system that taxes labor and penalizes realization while subsidizing accumulation and rewarding dynasty. It does not merely fail to prevent the emergence of a permanently entrenched economic elite. It is one of the primary mechanisms through which that elite perpetuates itself — through which the fortunes being assembled in this era of extreme concentration will be transmitted, compounded, and dynastically secured across generations that have not yet been born.

A tax system that calls the elimination of this subsidy a “death tax” while calling the subsidy itself nothing at all has successfully named its own loophole out of existence. The fiction here is not about money creation or government borrowing. It is about whose income gets taxed and whose does not — and about a political vocabulary so thoroughly captured by those who benefit from the answer that the question itself has become difficult to ask.

The anti-capitalist core

Capitalism’s distinguishing promise is the universal right to accumulate — the proposition that anyone, regardless of birth, has the right to build wealth through productive activity, and that the system’s institutions will keep that promise open rather than reserving it to those elites who were born into advantage. That promise is what distinguishes capitalism from the aristocratic systems it replaced, in which accumulation rights were the hereditary property of dynastic elites whose position was secured by birth rather than earned by contribution.

The Death Benefit violates this principle not as a side effect but as its central operation. Generation after generation, it ensures that the largest fortunes compound faster than the economy grows, that they pass between generations without capital gains taxation, and that the heirs who receive them begin each cycle with advantages — in borrowing cost, in asset access, in scale — that no newly accumulating investor can replicate. The playing field does not merely tilt. It tilts further at each generational transfer, locked in that direction by a tax provision that rewards the fact of inheritance over the act of creation.

Reforming the Death Benefit is not an attack on capitalism. It is a defense of capitalism — the restoration of a principle that the current system has spent a century quietly abandoning. The principle is simple. It is the principle widely attributed to Canada’s Carter Commission fifty years ago and that every comprehensive income tax system claims to uphold. “A buck is a buck is a buck.” Income is income regardless of its form. And a tax system that exempts one category of income — the appreciation of wealth held until death — from the principle it applies to every other category has not achieved neutrality. It has achieved the opposite: a permanent, compounding, dynastically self-sustaining exception to the rule that everyone pays once, on what they earn, like everyone else.

The death tax does not exist. The Death Benefit does. It is time to call it what it is.²⁴

“By taxing estates heavily at death the State marks its condemnation of the selfish millionaire’s unworthy life.”

— Andrew Carnegie, “The Gospel of Wealth,” North American Review, June 1889. Text via Carnegie Corporation of New York: https://www.carnegie.org/about/our-history/gospelofwealth/

Notes

1. Joint Committee on Taxation, Estimates of Federal Tax Expenditures for Fiscal Years 2025–2029, JCX-45-25 (2025). The JCT lists the provision as “Exclusion of capital gains at death” and estimates its revenue cost at $66.3 billion in FY2025, $72.5 billion in FY2026, $76.4 billion in FY2027, $79.9 billion in FY2028, and $84.2 billion in FY2029 — a five-year total of $379.3 billion. The provision applies entirely to individuals; the corporate column shows no revenue cost. The essay uses “roughly $72 billion annually” as a shorthand reflecting the FY2026 estimate, but the figure is growing: from $66 billion to $84 billion over the five-year window. The JCT estimate is a static revenue cost — it does not account for behavioral responses, meaning the actual revenue gain from eliminating step-up would differ as investors adjusted their realization behavior. The JCT’s terminology (“Exclusion of capital gains at death”) is more precise than the colloquial “step-up in basis” and directly describes what the provision does: it excludes from taxation the capital gains that accrued during a decedent’s lifetime.↩︎

2. The cognitive mechanism here is what George Lakoff describes as framing: mental structures activated by language that shape how we perceive and discuss political reality. A concept without a name cannot be activated as a frame and therefore cannot function as an object of political thought or debate. The “death tax” label is a near-perfect instance of what Lakoff calls framing the issue before the debate begins — by the time anyone argues about step-up reform, the frame is already set: taxation at death, burden on families, government taking. The Death Benefit cannot be argued against, reformed, or even defended within that frame because the frame contains no space for it. The provision that exists and does the damage has no linguistic handle in ordinary discourse; the provision that does not exist has one of the most politically effective labels in modern tax policy history. See George Lakoff, Don’t Think of an Elephant: Know Your Values and Frame the Debate (Chelsea Green, 2004).↩︎

3. The timing precision here is not hypothetical. Under IRC §1014, basis is set at the asset’s fair market value on the date of death. An asset sold the day before death triggers capital gains tax on the full appreciation. The same asset, unsold at death, receives a full step-up. The difference between these outcomes — potentially millions of dollars — is determined by the order of events within a single day. The provision does not merely advantage those who plan to hold until death. It creates a tax cliff at the precise moment of death that can turn an identical economic position into a dramatically different tax outcome depending on timing alone.↩︎

4. For estates above the exemption threshold — those paying estate tax on the value above $13.99 million per individual — step-up retains a partial coherence as the historical complement to estate tax paid. But even for those estates, the exemption amount itself receives the Death Benefit unconditionally: every estate, regardless of size, receives step-up as a pure benefit on its first $13.99 million in appreciated assets. The portion above the threshold is taxed, imperfectly — the effective estate tax rate averages less than 17%, well below the capital gains rate the decedent would have paid on realized gains during life. Step-up on the taxed portion is at least the historical complement to some tax paid. Step-up on the exempt portion — which covers the entirety of 999 estates in a thousand, and the first $13.99 million of every estate without exception — is the Death Benefit, unqualified and unconditional. The provision designed as a receipt for taxes paid now functions, in the overwhelming majority of its applications, as a substitution for taxation rather than a complement to it.↩︎

5. The scale advantage compounds this dynamic further. A dynasty borrowing against a $500 million diversified portfolio faces default risk so negligible that lenders charge rates near the risk-free floor — substantially below the rates available to smaller borrowers pledging less diversified collateral at higher loan-to-value ratios. The net spread between portfolio appreciation and borrowing cost is therefore wider for larger wealth holders than for smaller ones, not because their assets are better but because their scale reduces their cost of capital. The strategy is not merely self-sustaining — it is self-amplifying. Each generation inherits a larger portfolio at a fresh basis, can borrow at a lower rate relative to appreciation, and compounds the advantage further. The Death Benefit is the mechanism that converts this pre-tax advantage into a permanent post-tax advantage by ensuring that the surplus the spread generates is never subject to capital gains tax at any generational transfer. Switzerland’s mortgage system illustrates what happens when a structurally similar incentive — maintaining debt against appreciating assets rather than repaying it — is embedded in a tax architecture accessible to ordinary households: the behavior follows at scale, not as sophisticated tax planning but as the rational response to the incentives the tax code creates. The American dynasty strategy requires a wealth level most households cannot reach; the Swiss system demonstrates that the behavior itself, given the right incentives, is entirely ordinary.↩︎

6. Piketty’s r > g thesis has generated substantial critical literature. The debate centers on three questions: whether r has in fact persistently exceeded g historically (Rognlie argues the apparent rise in capital’s share reflects housing returns rather than broad productive capital); whether r > g is a theoretical necessity or a contingent historical pattern that may not continue; and whether r > g is sufficient to produce increasing concentration absent other institutional factors (Acemoglu and Robinson argue institutions matter more than the r-g spread). What is not seriously disputed is that capital has historically earned returns exceeding economic growth rates in many periods and places, creating a tendency toward concentration absent countervailing forces. The essay uses Piketty’s framework as background context for that tendency — not as a prediction of inevitability. The household-level dynasty argument does not depend on these macro disputes in any case: the borrow-spend-die strategy works whenever a specific large portfolio earns returns exceeding its borrowing cost, which is demonstrably true for large diversified portfolios over long periods regardless of the aggregate r-g relationship. The step-up amplification point is purely mathematical, following from the structure of IRC §1014 and the arithmetic of compound returns. On heterogeneous returns by wealth level, see Fagereng, Guiso, Malacrino, and Pistaferri, “Heterogeneity and Persistence in Returns to Wealth,” Econometrica 88(1) (2020). On the Piketty-to-dynasty connection in the legal literature, see Kades, Eric A., “Of Piketty and Perpetuities: Dynastic Wealth in the Twenty-First Century (and Beyond),” 60 Boston College Law Review 145 (2019). On the borrow-spend-die strategy and step-up, see Yale Budget Lab, “Buy-Borrow-Die: Options for Reforming Tax Treatment of Borrowing Against Appreciated Assets” (2024). The three-layer synthesis presented here does not appear in any of these sources in combined form. For empirical confirmation of the scale of unrealized appreciation in estates, see Gordon, Robert, David Joulfaian, and James Poterba, “Revenue and Incentive Effects of Basis Step-Up at Death: Lessons from the 2010 ‘Voluntary’ Estate Tax Regime,” American Economic Review 106(5) (2016): 662–67.↩︎

7. Batchelder, Lily L., “Leveling the Playing Field between Inherited Income and Income from Work through an Inheritance Tax,” NYU Law and Economics Research Paper No. 20-11 (2020). The income figures are pre-inheritance economic income, meaning the data describes households that are already high-income receiving the largest inheritances — the Death Benefit flowing toward existing advantage rather than toward need. The 48-to-1 ratio reflects not merely differences in asset accumulation but differences in asset type: high-income households disproportionately hold appreciated financial assets and business interests — precisely the assets for which step-up provides the greatest benefit — while lower-income households hold assets whose value lies primarily in their use rather than in accumulated unrealized appreciation.↩︎

8. Saez, Emmanuel, and Gabriel Zucman, “Wealth Inequality in the United States since 1913: Evidence from Capitalized Income Tax Data,” Quarterly Journal of Economics 131(2) (2016). The finding that the merely rich have lost relative ground to the ultra-wealthy — with fortunes of $20 million or more growing much faster than those of only a few million — is consistent with the scale-advantage mechanism described in footnote 5: larger portfolios sustain lower-cost borrowing, earn persistently higher after-tax returns, and benefit most powerfully from a step-up provision that ensures their appreciation is never subject to capital gains tax at any generational transfer. The Death Benefit does not merely reflect existing wealth concentration; it contributes to its acceleration.↩︎

9. The $55 million example illustrates the principle at a recognizable scale. Taken to its logical extreme, the same arithmetic applies to any size estate: a trillion-dollar estate divided among approximately 35,730 married couples — each receiving $27.98 million or less — would owe no estate tax and no capital gains tax on any of it. The Death Benefit has no cap. There is no upper limit on the total appreciation that can be permanently excluded from taxation through step-up — only the number of recipients and their combined exemptions. The provision that Congress introduced in 1921 to prevent double taxation of assets that had already borne estate tax contains, in its current form, a mathematical mechanism by which unlimited wealth can pass entirely untaxed across generations, provided it is divided among enough heirs. Rather than dispersing dynastic wealth, such a division would replicate it — seeding 35,730 new households each positioned above the threshold required to execute the borrow-spend-die strategy, each capable of compounding tax-free into the next generation.↩︎

10. The distributional asymmetry extends to the loss side of the ledger as well. Under current law, an heir who inherits a highly appreciated asset receives it at its stepped-up date-of-death value as their cost basis. If the asset subsequently declines in value, the heir may claim a capital loss — a deduction against taxable income — measured from that stepped-up basis. The decedent’s lifetime gain was forgiven at death. The heir’s post-death decline is deductible. The Treasury collected nothing on the appreciation and subsidizes the subsequent decline. The asymmetry runs in the same direction at every point: gains accumulated during life escape at death; losses incurred after death are deductible. This is not an incidental feature. It is the logical consequence of a system structured so that the estate is never on the wrong side of a tax transaction. Heads the dynasty wins; tails the Treasury loses.↩︎

11. The rhetorical deployment of the double-taxation objection follows the pattern identified throughout this series: a concern that has genuine application in a narrow, specific case is generalized into a universal principle that protects a far broader population than the concern actually covers. The Trust Fund argument — Social Security is pre-funded — applies at most to current retirees who paid into the system at specific contribution rates; it is deployed to foreclose any discussion of fiscal policy. The double-taxation argument applies at most to the tiny fraction of estates above the exemption threshold; it is deployed to foreclose any discussion of step-up reform. In both cases, the specific concern provides cover for a general position whose beneficiaries extend far beyond those for whom the concern is legitimate.↩︎

12. The Haig-Simons definition of income — the standard academic formulation underlying most comprehensive income tax theory — defines income as consumption plus change in net worth over a period. Under this definition, unrealized appreciation is income as it accrues, regardless of whether it is realized as cash. The US income tax departs from the Haig-Simons ideal by taxing only realized gains, a concession to administrative practicality and the liquidity constraints that would arise from taxing paper gains annually. Step-up converts this administrative concession — deferral of tax until realization — into a permanent exemption by making death a non-realization event. The concession to practicality becomes a gift to dynasty. Deemed disposition restores the realization requirement without requiring annual mark-to-market taxation: the gain is deferred during life, as under current law, but taxed once at death rather than permanently forgiven.↩︎

13. Batchelder (2020), cited in footnote 7. The one-seventh figure reflects the combined effect of preferential capital gains rates, the deferral of tax on unrealized appreciation during life, and the permanent forgiveness of that appreciation at death through step-up. It is a system-wide average across all inherited income; for estates executing the full dynastic strategy described in Section II, the effective rate on appreciated capital is zero.↩︎

14. Thomas, Kathleen DeLaney, “Tax and the Myth of the Family Farm,” 110 Iowa Law Review 1811 (2025), published May 15, 2025, https://ilr.law.uiowa.edu/volume-110-issue-4-1/2025/05/tax-and-myth-family-farm. Thomas conducts a comprehensive review of available evidence and finds no documented case of a family farm sold to fund federal estate taxes — language drawn directly from the paper’s abstract. The finding is consistent with the structure of the estate tax: the $13.99 million per-individual exemption excludes virtually all working family farms from estate tax liability, and those few farms above the threshold qualify for installment payment provisions under IRC §6166 that allow estate taxes to be paid over fifteen years. The forced-sale narrative has persisted for decades in political debate without empirical foundation.↩︎

15. Under the Canadian Income Tax Act, farm or fishing property transferred to a child may be elected at any amount between the property’s adjusted cost base and its fair market value, allowing the capital gain to be reduced to zero at the executor’s discretion. The child takes over the parent’s cost basis and pays capital gains tax only when they eventually dispose of the property. A lifetime capital gains exemption of $1.25 million (indexed to inflation as of 2024) further reduces or eliminates the tax on qualifying farm property when it is eventually sold. The combination — rollover deferral plus a substantial exemption — means a genuine working family farm transferring to the next generation faces either no immediate tax or a substantially reduced one, while the provision that produces the American Death Benefit — permanent forgiveness of unlimited appreciation for large investment portfolios — does not exist in the Canadian system.↩︎

16. The 1976 Tax Reform Act replaced step-up with carryover basis effective for deaths after December 31, 1976. Implementation was immediately delayed, then repealed retroactively by the Crude Oil Windfall Profit Tax Act of 1980 before it had taken effect for a single estate. The Joint Committee on Taxation’s analysis identified two primary problems: the difficulty of establishing the decedent’s original basis for assets held for many years without adequate records, and the administrative burden on executors required to gather that historical information across potentially large and diverse estates. Neither problem applies to deemed disposition, which requires only a current appraisal of fair market value at death — a standard element of estate administration regardless of the tax treatment of gains — and the decedent’s adjusted cost basis as maintained on their own tax records during life. The basis-reconstruction problem is further diminished for financial assets by the mandatory cost-basis reporting rules that have applied to brokerages since 2011 under Section 403 of the Energy Improvement and Extension Act of 2008, Division B of Public Law 110-343, 110th Congress (enacted October 3, 2008), which added IRC §§ 6045(g), 6045A, and 6045B requiring brokers to report customers’ adjusted basis in securities transactions: covered securities purchased after the relevant effective dates have electronically tracked bases already in the brokerage system. The primary remaining challenge is real estate and closely held business interests, both of which require appraisal regardless of the tax regime. The United States has one brief empirical data point on carryover basis in practice: the 2010 estate tax lapse. Under the Economic Growth and Tax Relief Reconciliation Act of 2001, the estate tax was eliminated entirely for deaths occurring in calendar year 2010, with a modified carryover basis regime substituted — heirs received the decedent’s original basis rather than a stepped-up basis, subject to a $1.3 million general basis increase allowance and an additional $3 million for assets passing to a surviving spouse. Gordon, Joulfaian, and Poterba’s analysis of estates that filed under the 2010 regime found that unrealized capital gains accounted for roughly 44% of the fair market value of non-cash assets, with many of the largest gains on assets held for at least two decades — empirically confirming the scale of appreciation that step-up forgives in typical large estates. The 2010 experience also documented that carryover basis was administratively manageable for financial assets, where brokerage records preserved original purchase prices, while creating genuine difficulties for real estate, closely held businesses, and assets acquired through prior inheritances — consistent with the essay’s argument that the administrative challenge is concentrated in illiquid assets rather than the broad asset universe. The 2010 lapse was an imperfect experiment: widely anticipated years in advance, giving wealthy families time to plan around it, and lasting only one year. But it is the only real-world test of carryover basis on actual US estates, and its results support rather than undermine the case for deemed disposition. Two further design questions deserve acknowledgment. First, if asset values fall between the date of death and the date of distribution to heirs — a genuine risk for illiquid assets where probate can take months or years — heirs could face a tax bill calculated on a value that no longer exists. A valuation election allowing heirs to use the lower of date-of-death value or distribution-date value, or installment payment provisions analogous to IRC §6166, would address this without requiring the gain to be forgiven entirely. Canada’s spousal rollover and farm property rollover elections provide working models. Second, for heirs whose primary inherited asset is a modestly appreciated home or small investment account, a deemed capital gains tax could consume a meaningful fraction of a modest inheritance. A specific exemption for the first tranche of inherited appreciation per recipient — analogous to Canada’s $1.25 million lifetime capital gains exemption for qualifying farm and small business property — would protect ordinary heirs while preserving the deemed disposition tax for the large appreciated positions that are the primary policy target. Both problems are genuine and soluble; neither constitutes a fundamental objection to deemed disposition, only a calibration challenge that other countries have addressed successfully.↩︎

17. The lock-in effect of the capital gains tax is a genuine and well-documented distortion. The correct remedy, however, is not a permanent exemption concentrated at death and at the top of the wealth distribution, but a more neutral realization regime — one that does not privilege any particular moment of realization over others. If capital gains rates are thought to be too high for moderate-income investors, the appropriate remedy is a progressive rate structure that reduces the rate for smaller gains while maintaining it for the large unrealized positions that account for most of the step-up benefit. Capital gains preferences are defensible as an incentive for ordinary household investment and risk-taking; they are not defensible as a permanent subsidy for dynastic wealth accumulation. The current system does not achieve a sensible balance between these objectives. It eliminates the tax entirely for the largest and most concentrated gains — those held until death — while taxing smaller realizations during life at rates that create genuine lock-in for ordinary investors. A reform that replaced step-up with deemed disposition and used the resulting revenue to lower capital gains rates on lifetime realizations would reduce lock-in at the bottom of the distribution while eliminating the dynastic exemption at the top.↩︎

18. OECD, Inheritance Taxation in OECD Countries, OECD Tax Policy Studies No. 28 (2021), https://doi.org/10.1787/e2879a7d-en; Hourani, Diana, and Sarah Perret, “Taxing Capital Gains: Country Experiences and Challenges,” OECD Taxation Working Papers, No. 72, OECD Publishing, Paris, February 25, 2025, https://doi.org/10.1787/9e33bd2b-en. The 2021 inheritance report states in Section 4.2.9 (“Tax treatment of unrealised capital gains at death”) that “the step-up in basis should be reconsidered, particularly where inheritance or estate taxes are not levied, or where inheritance or estate tax exemption thresholds are very high” — language that describes the post-OBBBA United States with precision. Note that the OECD uses “should,” a stronger recommendation than the “could” sometimes cited in secondary sources. The 2025 Hourani paper further identifies step-up as a provision that “significantly adds to lock-in effects in countries where it applies” (Section 4.2, p. 25) — directly contradicting the standard US defense that step-up reduces lock-in.↩︎

19. Royal Commission on Taxation (Carter Commission), Report of the Royal Commission on Taxation, 6 vols. (Ottawa: Privy Council Office, 1966). The Commission was chaired by Kenneth LeM. Carter, a Toronto tax lawyer and accountant. The phrase “a buck is a buck is a buck” is attributed to Carter in Canadian official discourse — including Canada’s 2024 federal budget and the Deputy Prime Minister’s remarks on capital gains reform — and is treated as the foundational expression of the Commission’s comprehensive income tax framework. The Canadian Tax Foundation describes it as a summary of the Commission’s views rather than a verbatim quotation; the specific location within the six-volume report has not been pinpointed in secondary literature, which cites Macdonald, Les, “Royal Commission on Taxation,” Canadian Encyclopedia, February 7, 2006, as the source for the attribution. The phrase is a riff on Gertrude Stein’s “a rose is a rose is a rose.” Whether Carter’s own words or an apt summary of his position, the principle is unambiguous: a dollar of wage income and a dollar of capital appreciation are both increases in economic capacity and should be treated identically for tax purposes. The deemed disposition rule at death is the direct application of this principle to the realization question: if gain is income, and income should be taxed once, then the permanent exclusion of gain from taxation at death is a departure from the principle the tax system should not permit.↩︎

20. Congressional Budget Office, “Change the Taxation of Assets Transferred at Death,” in Options for Reducing the Deficit: 2025 to 2034 (December 12, 2024), https://www.cbo.gov/budget-options/60943. The CBO presents two alternatives: carryover basis (heirs inherit the decedent’s adjusted basis and pay capital gains tax when they sell), estimated to reduce the deficit by $196.9 billion over 2025–2034; and deemed disposition (gains taxed as if the decedent had sold at death, included in the decedent’s final income tax return, with a deduction from estate taxes to avoid double taxation), estimated to reduce the deficit by $536.1 billion over the same period. Note that the underlying revenue estimates were produced by staff of the Joint Committee on Taxation. Both figures are static estimates that do not account for behavioral responses — the additional realizations that would occur during life once the incentive to hold until death is removed. The carryover basis alternative phases in slowly because it only affects assets acquired after the effective date; the deemed disposition alternative raises substantially more revenue by taxing the full stock of accumulated unrealized gains at each death.↩︎

21. The tax-shifting argument here is distinct from a distributional argument about who pays. The point is not merely that wealthy estates pay less — that is established by the one-seventh figure in Section V. The point is that the Death Benefit’s exclusion of appreciation from taxation does not eliminate that tax burden; it relocates it onto those who do realize gains during life. A separate objection to the dynastic engine framing is that dynasties sometimes fail — through heir incompetence, family fragmentation, overconsumption, or adverse markets. That is true. But heir incompetence is not a policy. A tax system designed for a republic should not rely on the Vanderbilts squandering their fortune as the mechanism for preserving economic mobility. The question is what the tax code does when dynasties succeed — and the answer, under current law, is that it actively assists them.↩︎

22. The anti-capitalist character of the Death Benefit connects to a broader argument about the institutional prerequisites of capitalism developed elsewhere in this series. Capitalism’s promise of universal accumulation — the right of anyone, regardless of birth, to build wealth through productive activity — requires institutions that keep that promise open: competitive markets, enforceable property rights, and a tax system that does not systematically advantage one form of wealth transmission over another. The Death Benefit violates this institutional requirement by making the return on inherited capital higher, after tax, than the return on newly created wealth — penalizing the capitalist virtue of productive accumulation relative to the aristocratic virtue of simply being born to the right family. The reform of step-up is, from this perspective, not a departure from capitalism’s logic but its restoration.↩︎

23. See footnote 2 on the Lakoff framing mechanism. The specific consequence here is that reform requires naming what is to be reformed — and the “death tax” label has occupied that naming function for decades, ensuring that the only thing retrievable in ordinary political discourse is a tax that does not exist for 99.92% of estates. The benefit that exists for all of them remains, in ordinary political discourse, nameless. That is not a rhetorical inconvenience. It is the primary political protection the Death Benefit has enjoyed for a century.↩︎

24. The international evidence surveyed in Section VII suggests that the United States is not merely an outlier in its treatment of capital gains at death — it is an outlier in a specific and compounding way: unlimited step-up combined with a near-universal estate tax exemption, producing a Death Benefit untethered from any tax that could justify it as a double-taxation remedy. The countries that have chosen differently — Canada, Denmark, Hungary, Australia — did so on the ground that income should be taxed once regardless of its form, and that death should be a tax-neutral event rather than the most tax-advantaged financial transaction available under law. That ground is available to the United States as well. The design tools exist. The international precedents exist. The administrative solutions to every objection that has been raised exist and have been demonstrated to work. What does not yet exist is a political vocabulary adequate to the task — a vocabulary that can name the benefit, describe its structure, and make the case for neutrality in terms that ordinary citizens can retrieve and use. This essay is an attempt to contribute one word to that vocabulary: the Death Benefit. It is enough to start with.↩︎

Bibliography

Batchelder, Lily L. “Taxing Privilege More Effectively: Replacing the Estate Tax with an Inheritance Tax.” Hamilton Project Discussion Paper 2007-07, 2007.

Batchelder, Lily L. “Estate Tax Reform: Issues and Options.” Tax Notes, January 2, 2009. SSRN Working Paper No. 1320304.

Batchelder, Lily L. “What Should Society Expect from Heirs? A Proposal for a Comprehensive Inheritance Tax.” Tax Law Review 63(1), 2017. NYU Law and Economics Research Paper No. 08-42. https://its.law.nyu.edu/faculty/profiles/representiveFiles/batchelder%20-whatshouldsociety_DB0B2FA7-1B21-6206-60DCCFEE17ED4009.pdf

Batchelder, Lily L. “Leveling the Playing Field between Inherited Income and Income from Work through an Inheritance Tax.” In Tackling the Tax Code, Brookings/Hamilton Project, 2020. NYU Law and Economics Research Paper No. 20-11.

Carnegie, Andrew. “The Gospel of Wealth.” North American Review, June 1889. Text via Carnegie Corporation of New York: https://www.carnegie.org/about/our-history/gospelofwealth/

Congressional Budget Office. “Change the Taxation of Assets Transferred at Death.” In Options for Reducing the Deficit: 2025 to 2034. December 12, 2024. https://www.cbo.gov/budget-options/60943

Fagereng, Andreas, Luigi Guiso, Davide Malacrino, and Luigi Pistaferri. “Heterogeneity and Persistence in Returns to Wealth.” Econometrica 88(1), 2020.

Gale, William G., and Joel B. Slemrod. “Rethinking the Estate and Gift Tax: Overview.” NBER Working Paper 8205, 2001. Published in Rethinking Estate and Gift Taxation, Brookings Institution Press, 2001.

Gordon, Robert, David Joulfaian, and James Poterba. “Revenue and Incentive Effects of Basis Step-Up at Death: Lessons from the 2010 ‘Voluntary’ Estate Tax Regime.” American Economic Review 106(5), 2016: 662–67.

Gordon, Robert, David Joulfaian, and James Poterba. “Revenue and Incentive Effects of Basis Step-Up at Death: Lessons from the 2010 ‘Voluntary’ Estate Tax Regime.” American Economic Review 106(5), 2016: 662–67. https://doi.org/10.1257/aer.p20161037

Hourani, Diana, and Sarah Perret. “Taxing Capital Gains: Country Experiences and Challenges.” OECD Taxation Working Papers, No. 72. OECD Publishing, Paris, February 25, 2025. https://doi.org/10.1787/9e33bd2b-en

Jacobson, Darien B., Brian G. Raub, and Barry W. Johnson. “The Estate Tax: Ninety Years and Counting.” IRS Statistics of Income. Washington: Internal Revenue Service, n.d.

Joint Committee on Taxation. Estimates of Federal Tax Expenditures for Fiscal Years 2025–2029. JCX-45-25. Washington: Government Publishing Office, 2025. https://www.jct.gov/getattachment/8c830c45-1680-4f7e-a649-2a0106f6b6e3/x-45-25.pdf

Kades, Eric A. “Of Piketty and Perpetuities: Dynastic Wealth in the Twenty-First Century (and Beyond).” Boston College Law Review 60(1), 2019: 145. https://bclawreview.bc.edu/articles/274

Kades, Eric A. “Of Piketty and Perpetuities: Dynastic Wealth in the Twenty-First Century (and Beyond).” Boston College Law Review 60(1), 2019: 145.

Kopczuk, Wojciech. “Bequest and Tax Planning: Evidence from Estate Tax Returns.” Quarterly Journal of Economics 122(4), 2007.

Kopczuk, Wojciech, and Joel B. Slemrod. “The Impact of the Estate Tax on Wealth Accumulation and Avoidance Behavior.” In Rethinking Estate and Gift Taxation, Brookings Institution Press, 2001.

Lakoff, George. Don’t Think of an Elephant: Know Your Values and Frame the Debate. White River Junction, VT: Chelsea Green, 2004.

Macdonald, Les. “Royal Commission on Taxation.” The Canadian Encyclopedia, February 7, 2006. https://web.archive.org/web/20120115170836/http://www.thecanadianencyclopedia.com/articles/royal-commission-on-taxation

Moretti, Enrico, and Daniel J. Wilson. “Taxing Billionaires: Estate Taxes and the Geographical Location of the Ultra-Wealthy.” American Economic Journal: Economic Policy 15(2), 2023.

OECD. Inheritance Taxation in OECD Countries. OECD Tax Policy Studies No. 28. Paris: OECD Publishing, 2021.

Piketty, Thomas. Capital in the Twenty-First Century. Cambridge: Harvard University Press, 2014.

Royal Commission on Taxation (Carter Commission). Report of the Royal Commission on Taxation. 6 vols. Ottawa: Privy Council Office, 1966. Digitized 2008 by the Privy Council Office of Canada; available via Government of Canada Publications: - Vol. 1 (Introduction): https://publications.gc.ca/site/eng/471998/publication.html - Vol. 2 (Tax system objectives): https://publications.gc.ca/site/eng/472002/publication.html - Vol. 3 (Taxation of individuals and families): https://publications.gc.ca/site/eng/9.699804/publication.html

Saez, Emmanuel, and Gabriel Zucman. “Wealth Inequality in the United States since 1913: Evidence from Capitalized Income Tax Data.” Quarterly Journal of Economics 131(2), 2016.

Tax Policy Center. “Taxable Estate Tax Returns as a Percentage of Adult Deaths, Selected Years of Death, 1934–2019.” Washington: Urban Institute and Brookings Institution, 2024.

Thomas, Kathleen DeLaney. “Tax and the Myth of the Family Farm.” Iowa Law Review 110, no. 4 (2025): 1811. https://ilr.law.uiowa.edu/volume-110-issue-4-1/2025/05/tax-and-myth-family-farm (PDF: https://ilr.law.uiowa.edu/sites/ilr.law.uiowa.edu/files/2025-05/ILR-110-Thomas.pdf)

Yale Budget Lab. “Buy-Borrow-Die: Options for Reforming Tax Treatment of Borrowing Against Appreciated Assets.” 2024. https://budgetlab.yale.edu/research/buy-borrow-die-options-reforming-tax-treatment-borrowing-against-appreciated-assets

Yale Budget Lab. “‘Buy-Borrow-Die’: Options for Reforming Tax Treatment of Borrowing Against Appreciated Assets.” 2025. https://budgetlab.yale.edu/research/buy-borrow-die-options-reforming-tax-treatment-borrowing-against-appreciated-assets

Zelenak, Lawrence. “Taxing Gains at Death.” Vanderbilt Law Review 46 (1993): 361–397. https://scholarship.law.vanderbilt.edu/cgi/viewcontent.cgi?article=2321&context=vlr (Zelenak’s foundational argument for deemed realization at death over carryover basis; the law review article that established his position in the literature. A shorter companion piece appeared simultaneously: “Taxing Gains at Death,” 59 Tax Notes 287–289 (1993).)

The first essay, “The Trust Fund Is a Comfortable Lie — and We’re Paying for It in Confused Citizens,” showed how the Social Security Trust Fund encodes a gold-standard fiction into a fiat-currency world — making a political commitment look like a savings account, and making “the money is already there” a reassurance that conceals a choice. The second essay, “The Debt Ceiling Was Built to Help Treasury Borrow — Then Someone Found the Gun,” showed how a tool of congressional humility became a weapon of extortion — and how the constraint it appears to impose has no economic content, only political utility.

This essay continues the series. The first two examined fictions about what the government must do — that it must pre-fund Social Security, that it must borrow before it can spend. This essay examines a fiction about what the government cannot do: create money. The prohibition turns out to be a legal choice rather than an economic law. The practice turns out to be already occurring, at multi-trillion dollar scale, under a different name. And the choice of covertness over transparency turns out to serve specific interests that have nothing to do with the inflation risk invoked to justify it.

More essays will follow. The fiscal fictions examined here — about savings, about borrowing, about money creation — are the foundation on which a larger set of convenient fictions about taxation has been built. Those arguments are easier to make once this foundation is visible.

The Received Wisdom

Everyone knows the government can’t print money. Printing money causes inflation. The Weimar Republic printed money and destroyed its currency. Zimbabwe printed money and produced billion-dollar loaves of bread. The United States government, by contrast, is constrained to obtain funds the responsible way: through taxation, which raises money from citizens, or through borrowing, which raises money from investors who expect to be repaid with interest. These are the two options. They are the whole menu.

Politicians across the spectrum repeat this as settled fact. Budget debates proceed entirely within it. When a spending proposal is challenged as unaffordable, “we can’t print money” is offered as the terminal argument — the appeal to a law of nature that ends the discussion. It is the fiscal equivalent of gravity: not a policy choice, not a legislative artifact, but a feature of economic reality that no government can repeal.

This essay is about that claim. Not about whether printing money is wise — it often isn’t. Not about whether inflation is a real constraint — it is, and an important one. Whether “the government cannot create money” describes an economic law or a legal choice is the question. Whether the prohibition is what it appears to be.

The Trust Fund essay showed that the government’s fiscal commitments are political rather than actuarial. The debt ceiling essay showed that the constraint on borrowing is nominal rather than real — that a ceiling unadjusted for inflation or economic growth has no coherent relationship to real fiscal capacity. This essay shows that the constraint on money creation is legal rather than economic. It also shows something else: the prohibition on printing money does not prevent the government from printing money. It requires the government to pay Wall Street for the privilege.

The argument has three layers: what the law actually says, what the government actually does, and what a better system would look like. None of these is the story most citizens have been told.

What the Law Actually Says

Before reaching the statutes, the basic operational constraint deserves a clear statement, because it is not a law at all. Treasury, like any bank customer, can only spend from balances in its bank account — the Treasury General Account (TGA) at the Federal Reserve — and can only increase its balances through the same mechanisms available to any bank depositor: receiving payments such as tax receipts, or receiving the proceeds of borrowing. Banks do not provide their customers, public or private, with the ability to create money in their own accounts. Only banks, and the Federal Reserve, can create money because only banks can issue the deposit liabilities that function as the primary medium of exchange in our economy. This is not a statutory requirement. It is a consequence of how the modern payments system works.

It was not always so. During the Civil War, the Legal Tender Act of 1862 authorized the Treasury to print $450 million in paper fiat currency — United States Notes, known as Greenbacks — and declare them legal tender. The government could act as its own bank because it was printing physical currency and handing it directly to soldiers, suppliers, and contractors. That avenue closed during the late twentieth century, with the final issuance of United States Notes wrapping up in 1971 and Congress permanently repealing the Treasury’s note-reissuance requirements in 1994 (Pub. L. 103–325, title VI, §602(f)(4)(B)). In the modern digital economy, nearly all money exists as electronic ledger balances clearing through Federal Reserve Banks. Treasury cannot hand out paper scrip to pay its bills. It must credit digital accounts, and those accounts are controlled entirely by the Federal Reserve and the commercial banking network. The operational constraint on government money creation is therefore real — but it is largely a consequence of the architecture of the modern payments system, not of any specific statute prohibiting money creation.¹

Yet physical currency issuance is not the only way Treasury could create money. If Treasury could sell securities directly to the Fed, the Fed would credit the Treasury General Account in exchange, effectively creating money for government use. The natural instrument for such a transaction is a zero-interest perpetual bond — a consol — that carries no coupon payment and no maturity date. It would cost Treasury nothing to service and would impose no future repayment obligation. A consol issued directly to the Fed in exchange for a TGA credit would be, in economic substance, government money creation: new purchasing power appearing in Treasury’s account, backed by an obligation that generates no cash drain.

For the first two decades of the Federal Reserve System, the authority to purchase debt directly from the Treasury was highly ambiguous even though it was frequently utilized during wartime and tax seasons. While the Emergency Banking Act of 1933 explicitly authorized direct purchases, the arrangement was abruptly halted by the Banking Act of 1935. A Senate committee amendment prohibited direct purchases and required that all Fed purchases of Treasury obligations occur only in the “open market” — through secondary market transactions with private dealers, not from Treasury itself. The committee report stated only that the provision “has been modified so as to provide that direct obligations of the United States may be purchased only in the open market.” One sentence. No reasoning. No floor debate for or against it.

When insiders were later asked why the prohibition was inserted, they gave three different answers. Marriner Eccles suggested it was inserted at the request of government securities dealers, who stood to lose business if Treasury could bypass them and deal directly with the Fed. W. Randolph Burgess thought it was meant to limit Federal Reserve balance sheet expansion. Others believed it was aimed at discouraging deficit financing by forcing Treasury into the market. These explanations, offered as much as twelve years after the fact by people who were present, are irreconcilable. The prohibition’s rationale was never publicly argued, never authoritatively settled, and has remained contested ever since. Nonetheless, Congress recognized the usefulness of direct purchases and temporarily restored a limited direct purchase authority during World War II, capping it at $5 billion and renewing it periodically through the postwar years until the exemption finally lapsed on June 1, 1981. The constraint the public treats as a law of nature is younger than many voters and has been reversed before. But the timing is suggestive. Most Depression-era reforms moved revenue away from Wall Street; this one moved revenue toward it, at the moment when dealer transaction volumes had collapsed and the provision’s beneficiaries needed it most — inserted without public debate into a bill whose controversy was focused entirely elsewhere.²

Whatever the reason for the open-market requirement, it is much more than a mere procedural inconvenience — it imposes a genuine market discipline that direct purchase would not. Treasury must issue debt that private buyers will accept at the offered terms before the Fed can acquire it from the open market. A zero-interest consol issued to raise funds directly from the Fed would fail entirely in the open market: with no coupon and no maturity, it offers the buyer no economic return. One might imagine selling such a consol at a discount — but discounting an instrument with no future cash flows at all means discounting it to zero, which defeats the purpose of issuing it. The open-market requirement therefore means that the natural instrument for transparent monetary financing — the zero-interest consol — is precisely the instrument that private markets will never buy and thus that the Federal Reserve will never have the opportunity to purchase. Treasury must attach interest sufficient to attract private buyers, creating future fiscal obligations and financing costs. The cheapest and simplest monetary financing option is permanently out of reach.

Layered atop this is the debt ceiling, examined in the previous essay: even the permitted route of borrowing is capped by a nominal dollar figure that adjusts for nothing. The ceiling does not prevent money creation directly; it constrains the one financing channel that the other structures have left open.

The governance structures surrounding these mechanics — statutory prohibitions, central bank independence, market discipline, debt limits — are real and consequential. The result is not a coherent system designed to enforce a sound economic principle. It is a stack of historical accidents, unexplained operational constraints, and institutional sediment, each layer serving its own purpose at its own moment, producing in combination a result that has been naturalized into apparent economic law. But together they describe how humans have chosen to regulate the monetary system, not what the monetary system or economics requires. The public narrative conflates these two things, presenting governance constraints as operational impossibilities, as if the law of the land were also a law of physics.

What Actually Happens: The One-Way Mirror

Here is what the law actually prevents: the Federal Reserve purchasing Treasury securities directly from Treasury at the moment of issuance, crediting the Treasury General Account with newly created reserves.

Here is what the law does not prevent: the Federal Reserve purchasing those same Treasury securities from Goldman Sachs two days after Goldman Sachs purchased them from Treasury.

In both cases, Treasury has issued a bond and received money. In both cases, the Fed holds the bond and has created new reserves. The only difference is that Goldman Sachs held the bond for forty-eight hours and collected a spread for its trouble.

The Federal Reserve’s balance sheet expanded from roughly $900 billion in 2008 to $9 trillion by 2022, an increase of more than $8 trillion, through four rounds of quantitative easing. In each round, the Fed purchased Treasury securities and other assets from financial institutions, crediting those institutions’ reserve accounts with newly created money. Treasury, having sold those bonds to the same institutions days or weeks earlier, had already received the proceeds and spent them. The circuit runs: Treasury issues debt, the public buys it, the Fed buys it from the public, the public has cash instead of bonds, Treasury has spent the money. Those reserves did not come from some pre-existing pool of Fed savings. They were created at the moment of purchase by a bookkeeping entry that expanded both sides of the Fed’s balance sheet simultaneously — assets up by the value of the bond, liabilities up by the value of the newly credited reserves. This is what “printing money” means in a modern payments system: not a physical press, but a keyboard entry that brings new money into existence.

The same mechanism applies when the Fed buys commercial debt — mortgage-backed securities, agency bonds, corporate paper. All of it is purchased with newly created reserves. All of it generates income that the Fed is required by statute to remit to Treasury after covering its expenses, making the federal government the effective beneficiary of private-sector interest streams that the Fed has monetized on its behalf.

This is monetary financing. It is called quantitative easing, open market operations, asset purchase programs — anything but what it is. The IMF’s own 2022 research paper acknowledged exactly why the labeling matters: “central banks do not employ monetary financing in a transparent manner and following theory prescriptions,” which is why “the analysis cannot easily pinpoint historical episodes of monetary financing.” The practice is occurring. It is not being labeled as what it is.³

Paul Sheard, at the time S&P’s chief global economist, put the mechanics plainly in a 2014 paper: QE “involves the central bank creating bank reserves to acquire assets,” and this “does sound at least like the electronic printing of money.” He went on to argue that QE differs from true monetary financing — that argument comes below — but his concession of the mechanics is the starting point.⁴ Adair Turner, former chairman of the UK Financial Services Authority, presenting at the IMF’s own annual research conference in 2015, went further: “the technical feasibility and desirability in some circumstances of monetary finance is not in doubt,” and “all the really important issues are political.”⁵ Among those who have actually conducted the technical debate, the question is no longer whether the government can create money. The question is whether it should do so openly.

The monetary sovereign proxy. Considered as a whole, the Federal Reserve performs operationally everything that a monetary sovereign does. It creates money from nothing, acquires assets with those created reserves, earns income on those assets, and remits that income to Treasury. When its interest expenses exceeded its income during 2022 to 2025 — rates having risen above the yields on its QE-era bond portfolio — it continued paying its obligations by crediting reserve accounts, because it cannot be insolvent in the currency it issues. The “deferred asset” that appeared on its balance sheet during that period was an accounting convention. The Fed was never at risk of being unable to pay. It created the reserves needed and recorded the excess as a balance sheet entry to be recovered from future earnings. This is precisely what a monetary sovereign does.¹¹

The United States government has built a legal architecture that prohibits it from acting openly as a monetary sovereign. Simultaneously it has created an institution — the Federal Reserve — that acts as its monetary sovereign proxy, performing all the same operations under a different label. The cost of conducting monetary sovereignty through a proxy rather than directly is the broker tax: every dollar of effective monetary financing must pass through the primary dealer oligopoly, which collects a spread at auction and another on the resale to the Fed. Billions of dollars annually flow to a licensed group of financial institutions as the price of maintaining the fiction that the government is not doing what it is doing.⁶

The private toll. The primary dealer system creates a structural oligopoly within the Treasury market. A select group of roughly two dozen designated financial institutions — Goldman Sachs, JPMorgan, Citigroup, and their peers — are the only participants required to bid in every Treasury auction, making them the guaranteed backstop for every issuance. More importantly for this essay’s argument, they are the exclusive counterparties for Federal Reserve open market operations: when the Fed buys Treasury securities, it buys only from primary dealers, not from the broader universe of auction participants. Foreign governments, pension funds, and other direct bidders who purchased at auction must sell through primary dealers to reach the Fed. The dealers collect a spread on both legs of that round trip.

During QE periods when the Fed was purchasing $80 to $120 billion of Treasuries each month, the aggregate dealer spread on those transactions represented billions of dollars annually — a private fee on what is, in economic substance, a government financing operation.¹⁴ The open-market requirement does not prevent monetary financing. It requires that monetary financing be conducted through a channel that makes it profitable for a specific, small, licensed group of financial institutions. The prohibition is not a wall between the government and a dangerous practice. It is a one-way mirror. The practice occurs on one side; the public is told it does not exist on the other. And a private toll is collected at the glass.

Is QE Really Monetary Financing? The Permanent/Reversible Question

The most sophisticated objection to QE as monetary financing is that QE is categorically different from true monetary financing because QE is explicitly designed to be reversed. The Fed can sell back its Treasury holdings at any time, extinguishing the reserves it created. True monetary financing — such as helicopter money or direct central bank fiscal transfers — is intended to be permanent. In this view, a central bank that can and will unwind its asset purchases is merely managing liquidity, not financing the state.

Certainly, a central bank that creates reserves with the genuine intention to reverse the operation when conditions normalize is doing something conceptually different from one providing a permanent fiscal transfer. The credibility of that reversal commitment is what separates monetary policy from monetary financing, and the distinction is not cosmetic.

The reality, however, is a complete inversion of these theoretical labels: “permanent” fiscal injections are routinely destroyed through taxation, while “temporary” monetary balances show a stubborn tendency to become permanent.

Consider helicopter money — the textbook definition of a permanent fiscal transfer. The moment those dollars are spent into the economy, they immediately enter the regular stream of national commerce, where they are relentlessly drawn down and extinguished by federal taxation. Helicopter money is only as permanent as the tax code allows it to be. Conversely, Federal Reserve QE balances — explicitly advertised as temporary liquidity injections — have exhibited a stark ratchet effect over the last two decades, with each crisis establishing a higher baseline floor that is never fully dismantled, as the chart below makes visible.

This empirical breakdown reveals that permanence is the wrong criterion for evaluating monetary financing instruments altogether. Helicopter money, zero-interest consols, and QE reserves are all subject to the same macroeconomic withdrawal mechanism: federal taxation, which extinguishes currency regardless of the accounting pipeline that birthed it. The relevant criterion is never “will this instrument exist forever?” but always “what do current economic conditions require?” — the fundamental question Abba Lerner frames in his Functional Finance writing, and one whose answer remains identical across every instrument.⁷ The permanent-versus-reversible distinction is not a boundary line of economic law. It is a minor difference in transmission mechanics. And unlike the reversibility of QE — an informal commitment with no legal mechanism to enforce it against a future Fed that chooses otherwise — an explicit statutory instrument governed by clear conditions is both more honest about what it is and more governable in practice.

What monetary financing actually is. The permanence test — distinguishing QE from helicopter money by whether reversal is intended — imports the commodity-currency mental model through the back door. It treats “real” financing as something that gets fully repaid and monetary financing as something that doesn’t. Under Lerner’s Functional Finance, repayment is secondary to economic conditions. Whether a monetary instrument is retired next year or in fifty years depends on what conditions warrant, not on what label was applied at issuance.

The consolidated balance sheet provides the right criterion. All Fed holdings of Treasury securities constitute monetary financing, because all share the same economic character: the Fed is a government institution that remits its earnings to Treasury, so debt the Fed holds generates interest that flows in a circle. The government owes money to an institution that returns the money. Whether this arrangement will be wound down in five years or fifty or never is a question about future monetary policy, not about current economic substance. The Fed’s $4.3 trillion in Treasury holdings represents $4.3 trillion of government self-financing — the figure that the standard public debt measures systematically obscure.

The Weimar problem. The objection that terminates most public discussions of monetary financing must be met directly: the experience of Weimar and Zimbabwe. Both cases are real. Both involved money creation that produced catastrophic hyperinflation.

The examples are real, but the inference is wrong. What Weimar Germany and Zimbabwe shared was not money creation per se — it was money creation in economies that had already lost the productive capacity to absorb additional demand. Germany in 1923 was paying war reparations in foreign currency while its industrial base was under foreign occupation. Zimbabwe in 2008 had experienced a collapse in agricultural output following land confiscation and the flight of skilled workers. In both cases, money was being created to purchase goods and services the domestic economy could not supply. More money chased fewer goods: the definitional mechanism of inflation.

The United States in 2009, after a year of QE1, with unemployment above 10 percent and the economy operating far below capacity, was not comparable on any economically relevant dimension. The economy had vast unused productive capacity — workers, factories, office space — capable of absorbing additional demand without bidding up prices. The Federal Reserve created roughly $3.5 trillion in reserves through QE1 to QE3. Inflation averaged below 2 percent throughout the period. The AIER cites Venezuela and Zimbabwe as the inevitable destination of any monetary expansion but does not explain why $3.5 trillion in QE produced a decade of below-target inflation rather than hyperinflation. The answer, which Abba Lerner articulated in his 1943 paper “Functional Finance and the Federal Debt,” is that the relevant variable is not the money supply but productive capacity. Create money when the economy has slack to absorb it; do not when it does not. The constraint is real. It just is not the constraint that “printing money causes inflation” implies.⁷

What Transparent Instruments Would Look Like — And How to Govern Them

Given that monetary financing already occurs at multi-trillion dollar scale, the question is not whether to permit a dangerous new practice. The question is whether to conduct an existing practice transparently.

Two instruments have been proposed in the serious literature as vehicles for transparent monetary financing. Both are simpler than what the government currently does.

The first is the platinum coin. 31 U.S.C. §5112(k) authorizes the Secretary of the Treasury to mint platinum coins in accordance with “such specifications, designs, varieties, quantities, denominations, and inscriptions as the Secretary, in the Secretary’s discretion, may prescribe.” No denomination limit is specified. A Treasury secretary could mint a coin with a face value of $1 trillion, deposit it with the Federal Reserve, and the Fed would credit the Treasury General Account accordingly — no bond issued, no interest obligation, no maturity date. Whether this particular authority extends to this particular use is contested. It is also suggested by some that the Federal Reserve would refuse to accept the coin for deposit. The point is that the debate is about legal interpretation, not about economic possibility.

The second and more serious candidate instrument is the zero-interest perpetual consol — a bond with no coupon and no maturity date, issued by Treasury to the Fed in exchange for a credit to the Treasury General Account. Paul Sheard, writing in 2014, described exactly this instrument as the natural implementation vehicle for monetary financing, calling it “nonmarketable zero-coupon perpetuities to the central bank.”⁴ Adair Turner, presenting at the IMF in 2015, described the same thing as “a non-interest-bearing non-redeemable due from government receivable.”⁵ Two senior mainstream economists, writing independently, arrived at the same instrument as the obvious description. Making it a legal instrument of Treasury financing would formalize what the serious literature already treats as standard. The zero-interest consol is also economically superior to the current interest-bearing arrangement for the consolidated government in every rate environment — the objections to it are institutional and political rather than economic.

The coin and the consol are economically identical. Both create purchasing power for the government at zero net cost, require no future interest payments and no principal repayment, and could be retired — the coin melted, the consol cancelled — when economic conditions warrant.

If Treasury issued these instruments when conditions warranted and retired them when conditions required, the outstanding stock at any moment would be a public ledger of accumulated monetary financing decisions. This number currently exists nowhere in honest form. The IMF acknowledged that opacity prevents empirical analysis of monetary financing episodes, because the practice cannot be studied when it is not labeled as what it is.³ An explicit stock, published and tracked, would give citizens something currently absent: a visible, auditable account of how much money the government has created for fiscal purposes.

The most common objection to transparent monetary financing is the credible-commitment problem: to be stimulative, monetary financing must be credibly permanent, but to be safe from inflation, the central bank must retain the ability to reverse it. Critics argue these requirements are irreconcilable.

The framing asks the wrong question. Whether the government can make a credible commitment to permanence is genuinely uncertain. But whether the government can make a credible commitment to rationality — to doing what economic conditions require — is not. Central banks make exactly this kind of commitment routinely, with genuine market credibility. The Fed’s forward guidance ties policy to observable economic indicators. No one argues that interest rate policy faces an insoluble credibility problem because the Fed might someday change rates.

Monetary financing can be governed by the same structure. Given three economic states, there are three responses: when the economy is below capacity and inflation is below target, monetary financing is appropriate and the commitment not to reverse is credible because withdrawal would be contractionary and nobody wants that; when the economy is approaching capacity and inflation is rising, new monetary financing stops and the stock held stable; when the economy is at or above capacity and inflation is at or above target, withdrawal of monetary financing is appropriate and benign — the economy can absorb monetary contraction precisely because it is running hot, so there is no paradox in the commitment to reverse under those conditions. The commitment would not be “this money will never be withdrawn,” it would be “this money will be withdrawn when, and only when, the economy can absorb it.” That is the Taylor Rule applied to a different instrument. The Fed’s existing dual mandate already defines the observable indicators for monetary policy. Congress would need to authorize the instruments and codify the monetary financing framework; the operational mechanics already exist.

What the Public Accounts Would Show

The current system does not merely conceal the existence of monetary financing. It systematically distorts the public accounts in ways that make the fiscal situation appear worse than it is, and that distortion is used reliably to justify austerity measures whose necessity is partly an artifact of the accounting.

Every discussion of the national debt distinguishes between “gross debt” and “debt held by the public.” Gross debt is approximately $39 trillion. Debt held by the public is approximately $30 trillion. The difference — roughly $7 trillion — is intragovernmental debt, primarily the Social Security Trust Fund IOUs examined in the first essay of this series: accounting entries representing one government account’s claim on another, not genuine obligations to external creditors.

But “debt held by the public” includes the Federal Reserve. The Fed’s Treasury holdings — approximately $4.3 trillion — are counted as public debt alongside holdings by pension funds, foreign governments, and individual investors. A pension fund that holds Treasury bonds has a genuine claim on future tax revenues. The Federal Reserve holds Treasury bonds and collects interest on them, but returns most of that interest to the Treasury. The Fed is not a creditor in any economically meaningful sense. It is a government institution holding an obligation that generates a payment flowing back to the obligor.

The Fed is included in “debt held by the public” because the Federal Reserve banks are technically private corporations — chartered by Congress and owned by member banks, but not federal government agencies in the strict legal sense. Treasury’s definition of “held by the public” means held outside federal government accounts, and the Fed falls outside that boundary by legal classification. The terminology reflects a legal fact. It does not reflect an economic fact.

For much of the Fed’s history, before 2008, the Fed’s stock of Treasury securities was modest enough that including or excluding it in “debt held by the public” made little practical difference. Before QE, the Fed holdings were only about $700 billion, but after four rounds of QE, they expanded to $9 trillion at peak. The gap between “publicly held” and “privately held” — debt owed to genuine external creditors — is now measurable in trillions. The chart accompanying this essay shows four lines: gross debt, debt held by the public including the Fed, privately held debt excluding the Fed, and the Fed’s holdings alone.¹⁰ Before 2008, the publicly-held and privately-held lines are nearly indistinguishable. After 2008, they diverge, and the divergence tells the story of QE more accurately than any official description of those programs does.

Treasury’s own advisory body understands this. The Treasury Borrowing Advisory Committee noted in its February 2026 letter to the Secretary that Treasury securities in the Fed’s SOMA portfolio are “offsetting liabilities of the Treasury and assets of the Fed,” so that “the liability side of the consolidated balance sheet consists of only those Treasury securities which are privately held.”⁸ The government’s own senior fiscal advisors use the consolidated balance sheet framework internally, treating Fed-held debt as self-canceling, while the public debate focuses on gross figures that include it as genuine external obligation. The Federal Reserve Bank of Dallas publishes a FRED data series — MVPHGFD027MNFRBDAL — tracking the market value of privately held gross federal debt, excluding Fed holdings.⁹ It is available to anyone. It is cited by no one in the public debate.

The gap between the standard public accounts and an honest accounting produces two specific distortions.

The first is in the debt level. The headline figure — gross debt, approximately $39 trillion — includes intragovernmental holdings ($7 trillion, primarily trust funds representing the government owing money to itself) and Fed holdings ($4.3 trillion, debt owed to an institution that remits most of its earnings to Treasury). The genuinely external debt, owed to pension funds, foreign governments, individual investors, and financial institutions with real claims on future tax revenues, is closer to $26 trillion. This is still a large number. But it is meaningfully different from $39 trillion, and the difference matters when politicians and commentators argue that the debt level is at or near a crisis threshold.

The second distortion is in the interest burden. Federal interest payments are currently reported at roughly $900 billion annually — among the largest line items in the federal budget and the figure most frequently cited in arguments for fiscal consolidation. This figure includes interest paid to the Fed, which in normal times is returned to Treasury through remittances. The net interest cost — what the government actually pays to genuine external creditors — is substantially lower. During 2022 to 2025, when the Fed raised rates to combat inflation, its interest costs on bank reserves exceeded its earnings on the fixed-coupon bonds acquired during QE, creating an operating loss that suspended remittances entirely; Treasury was left paying full coupon rates on Fed-held debt with no offsetting return. Throughout this period the Fed continued meeting its obligations by crediting reserve accounts — a monetary sovereign proxy cannot be insolvent in the currency it manages — while recording the shortfall as a deferred asset to be recovered from future earnings. The deferred asset is an accounting entry, not a crisis. But it reveals the structural vulnerability of the covert system: the apparent zero-cost of Fed-financed debt holds only when monetary policy is accommodative.¹¹

Zero-interest instruments would eliminate this vulnerability entirely. A consol pays no coupon; there is no remittance mechanism to break down. The financing cost is genuinely zero in all rate environments, not contingently zero when the Fed happens to be making money. Worth noting in passing: the remittances themselves function economically less like an investment return than like a levy on private income streams — the government capturing a share of interest flows it inserted itself into by creating money, not by contributing real capital. The full implications of that observation belong to a later discussion of taxation and what a return on government co-investment actually means.¹²

When transparent instruments replace the current covert system, the monetary financing stock would be classified as government self-financing and removed from the external debt figures. Citizens evaluating fiscal policy would see an external debt several trillion dollars lower than the headline figure and an interest burden hundreds of billions of dollars lower than the gross figure. They do not have this now.

The Question That Cannot Currently Be Asked

Every year, Congress makes an implicit decision about how much of government spending will be financed by taxation, how much by borrowing from the public, and how much by monetary financing. The first two components are debated, reported, scored by the CBO, and subjected to democratic scrutiny. The third is left to the Federal Reserve under the label of monetary policy and never acknowledged as a fiscal decision at all.

“How much should be monetized and how much borrowed in any given year?” cannot be asked in any public forum, because one side of it — the monetization option — is officially nonexistent.

Serious economists and institutions have arrived at four distinct positions on the desirable quantity of monetary financing, none of which has penetrated public debate. The AIER and the fiscal conservative establishment hold that the answer is always zero — any positive amount of monetary financing leads inevitably to Weimar or Zimbabwe. They ignore conditionality, treat the worst available historical examples as universal, and decline to engage with the empirical record of quantitative easing which conducted monetary financing at $5 trillion scale without producing hyperinflation. The IMF, in its 2022 research paper, holds that the answer depends on conditions — strong central bank independence, low initial inflation, and manageable fiscal deficits make some amount of monetary financing safe; their absence makes it dangerous — but articulates this as a technical judgment for expert institutions rather than a question for democratic deliberation.³ Sheard and mainstream financial analysis hold that the answer depends on whether the operation is permanent or reversible and on the quality of the institutional framework — right as far as it goes, still a technical judgment.⁴ Turner, speaking at the IMF’s own conference in 2015, holds that the technical case is settled and the real questions are political — this is the most honest framing available — yet even Turner does not acknowledge that the debate is being conducted in the abstract while the practice occurs covertly at massive scale under a different name.⁵

None of these four positions asks who should make this decision, or whether the public should know it is being made. Making transparent monetary financing legal — not unlimited, not ungoverned, but explicit, subject to observable rules, and reported in the public accounts — would not settle the question of how much monetization is appropriate. It would make the question askable. That is democratic deliberation, and it is what is currently impossible.

Who Are We Fooling?

The invisibility of monetary financing is not primarily about hiding the truth from the public, though that is a consequence. It is about managing what Congress believes it can and cannot do.

The Federal Reserve’s institutional independence depends, in part, on Congress not asking why fiscal decisions are being made by an unelected committee under monetary policy cover. QE is fiscal policy. It decides how much of Treasury’s deficit spending is financed by money creation rather than by genuine public borrowing. That decision has distributional consequences — it affects asset prices, credit conditions, the real value of savings, and the cost of the government’s debt. It is made by twelve voting members of the Federal Open Market Committee, none elected, none describing what they are doing as fiscal policy. Sheard himself acknowledged that QE combined with large fiscal deficits “is probably best viewed as a monetary operation of the consolidated government” — a fiscal decision.⁴ The institutional framing prevents that characterization from becoming politically actionable.

The economics profession has a related interest. Economists who understand fiat currency and monetary operations routinely defer to commodity-currency language in public, in congressional testimony, and in op-eds, because that is the vocabulary their audiences recognize and because departing from it carries professional and political costs. The cumulative effect is a sustained public education in a framework that has been inaccurate since at least 1971. Each individual accommodation seems reasonable. The aggregate is systematic, and the public debates fiscal policy inside a frame that excludes one of the options on the menu.

The fiscal conservative policy community has a straightforwardly political interest in maintaining the fiction. “We can’t afford it” sounds like a statement of fact rather than a statement of values, it preempts deliberation about priorities, and it forecloses the question of who bears the cost of the things we allegedly cannot afford. If the existence of a third fiscal option were public knowledge, “we can’t afford it” would have to be replaced with “we have chosen not to finance it this way” — a statement that requires justification and invites scrutiny. Maintaining the fiction converts a political position into an apparent statement of natural law.

The primary dealer community’s interest is direct and measurable. The Fed’s open-market requirement generates billions of dollars annually in dealer spreads that would not exist if Treasury could deal directly with the Fed.¹⁴ Whether the requirement was originally inserted at dealer request, as Eccles believed, is historical interpretation. That it currently generates substantial private profit from a public operation is undeniable.

None of these actors needs to coordinate. Each operates within professional norms and institutional incentives that happen to point in the same direction. The result is a public debate conducted on the premise that a third fiscal option does not exist, even though that option is exercised at multi-trillion dollar scale by institutions that have every reason to call it something else.

Article I, Section 8 assigns to Congress the power to borrow money on the credit of the United States. The current system effectively transfers a portion of a closely related authority — the authority to decide how much of the government’s deficit spending is financed by money creation — to the Federal Reserve, by maintaining a fiction that prevents Congress from knowing the authority exists. Whether the Fed makes better decisions than Congress is a separate question. The question here is who is constitutionally empowered to make the decision, and whether democratic institutions can exercise authority they don’t know they have.

Who Benefits From Congressional Ignorance?

The actors identified above share a specific interest that unifies them despite their different institutional positions: they all benefit from Congress not knowing that a third fiscal instrument exists and is already being used.

The argument for maintaining congressional ignorance — made explicitly or implicitly — is that Congress cannot be trusted with the truth. Democratic institutions given access to zero-cost financing will use it irresponsibly. Political incentives systematically favor spending over fiscal discipline. The inflation risk is real and democratic institutions are poorly equipped to resist it.

This argument has genuine force. The historical record of democratic governments with direct access to monetary financing includes real inflationary disasters. The political economy logic is sound. The concern is not paranoid.

But the argument concedes something its proponents do not acknowledge. It is not an argument that monetary financing is always wrong. It is an argument that democratic institutions cannot be trusted to govern it. This raises a question: if democratic institutions cannot be trusted with transparent monetary financing, why can they be trusted with covert monetary financing? The current system does not prevent monetary financing. It prevents congressional oversight of monetary financing. The Fed makes the decision. Congress is excluded from it. The justification is democratic incapacity; the result is rule by technocracy under cover of a fiction about legal prohibition.

The people who benefit from congressional ignorance are the same people invoking congressional irresponsibility as the justification for that ignorance. “They can’t handle the truth” is the argument made by those who profit from their not knowing it.

Can We Trust Congress With the Truth?

The political discipline objection deserves full engagement. The IMF’s empirical findings confirm that monetary financing produces modest inflation effects under conditions of strong central bank independence and low initial inflation — and much stronger effects when those conditions are absent.³ The historical record includes genuine inflationary disasters. These are serious arguments, grounded in real experience.

The conditional-rule framework developed above answers them without dismissing them. The credibility-commitment problem dissolves when the right question replaces the wrong one. Can the government make a credible commitment to permanent non-reversal? That question has no satisfying answer. Can the government make a credible commitment that reversal will occur only when economic conditions make it benign? That question does, because central banks make exactly this kind of conditional commitment routinely — it is called forward guidance, and markets find it credible.

The accountability argument is equally important. A Congress that knows it is making a monetary financing decision, and knows the public knows, faces different incentives than a Fed committee making the same decision under cover of monetary policy language. Accountability changes behavior. A politician who votes for monetary financing that subsequently generates inflation faces that record. A Fed committee that expands the balance sheet through QE faces no comparable accountability, because the monetary policy framing insulates the fiscal decision from democratic review.

Covert monetary financing — the system we have — operates without a published rule, a public ledger, democratic accountability, or political cost for excess. Transparent monetary financing, governed by a publicly stated conditional rule tied to observable economic indicators, has all of these. The covert system does not protect against irresponsibility. It protects against scrutiny.

Stripped of its technical language, the political discipline argument makes a specific claim: that the public and their elected representatives cannot be trusted with an accurate understanding of how their government finances itself. The entire apparatus — the statutory prohibitions, the institutional conventions, the careful language of open market operations and quantitative easing — exists to maintain that position.

The case for the current system, stated honestly, is that Congress and the public cannot be trusted with an accurate account of their government’s fiscal options. The technical apparatus — the statutory prohibitions, the institutional conventions, the careful language of monetary policy — exists to maintain that position. So does the precedent. And so do the financial interests of those who collect a toll every time the government does, covertly, what the law says it cannot do at all.

“You want the truth? You can’t handle the truth!” — Col. Nathan Jessup, A Few Good Men (1992)

Notes

1. The statutory architecture relevant to money creation: Federal Reserve Act §14, codified at 12 U.S.C. §355 (direct purchase prohibition, enacted 1935; wartime exemption lapsed June 1, 1981); Second Liberty Bond Act of 1917, consolidated into the Public Debt Acts of 1939 and 1941, now at 31 U.S.C. §3101 (debt ceiling); U.S. Constitution, Art. I, §9, cl. 7 (Appropriations Clause). The Emergency Banking Act of 1933 explicitly authorized direct Fed purchases of Treasury securities; the Banking Act of 1935 reversed this without stated rationale. The Legal Tender Act of 1862, 12 Stat. 345, authorized the original issuance of United States Notes (Greenbacks), enabling the Civil War government to create money through physical currency delivered directly to payees; final issuance of that series concluded in 1971. Congress permanently repealed the Treasury’s note-reissuance requirements in 1994, Pub. L. 103–325, title VI, §602(f)(4)(B). Treasury’s authority to maintain its own depositary functions is recognized in 31 U.S.C. §3322, but in practice no modern counterparty accepts Treasury instruments unless they clear through the Federal Reserve system. The Antideficiency Act, 31 U.S.C. §1341, constrains spending authority rather than money creation and is discussed in the debt ceiling essay.

2. The 1935 prohibition and its contested rationale: Kenneth D. Garbade, “Direct Purchases of U.S. Treasury Securities by Federal Reserve Banks,” Federal Reserve Bank of New York Liberty Street Economics, September 29, 2014. https://libertystreeteconomics.newyorkfed.org/2014/09/direct-purchases-of-us-treasury-securities-by-federal-reserve-banks/. The three competing explanations — Eccles on dealer lobbying, Burgess on balance sheet limits, others on deficit deterrence — are documented in the Garbade source. The wartime exemption was authorized by the Second War Powers Act of 1942, which framed it explicitly as restoring a cash management safety net; Congress chose a $5 billion cap rather than full repeal even under wartime emergency, a pattern consistent with the dealer-protection explanation for the original prohibition. In 1977 and 1979, Treasury used the remaining direct purchase authority as a bridge during debt ceiling impasses — providing speed and market-independence rather than additional fiscal capacity, since direct purchases counted against the ceiling the same as open market sales. Representative Ron Paul’s characterization of those uses as an “end run around Congress” was constitutionally confused: Treasury was honoring spending Congress had already appropriated, accessing headroom that existed under the ceiling, doing so faster than market methods would allow. The political toxicity of the ceiling-bridge use, combined with the practical obsolescence of the authority after cash management bills were introduced in 1975, made renewal impossible in 1981.

3. IMF (2022): Itai Agur, Damien Capelle, Giovanni Dell’Ariccia, and Damiano Sandri, “Monetary Finance: Do Not Touch, or Handle with Care?” IMF Departmental Paper DP/2022/001, January 2022. https://www.elibrary.imf.org/view/journals/087/2022/001/article-A001-en.xml.

4. Sheard (2014): Paul Sheard, “A QE Q&A: Everything You Ever Wanted to Know About Quantitative Easing,” S&P Economic Research / Harvard Kennedy School, August 7, 2014. https://www.hks.harvard.edu/sites/default/files/centers/mrcbg/files/PSheardQEQAAugust2014.pdf.

5. Turner (2015): Adair Turner, “The Case for Monetary Finance — An Essentially Political Issue,” IMF Jacques Polak Annual Research Conference, November 5–6, 2015. https://www.imf.org/external/np/res/seminars/2015/arc/pdf/adair.pdf.

6. The primary dealer system: the Federal Reserve Bank of New York maintains the current list of designated primary dealers and their obligations. As of 2026, approximately 25 dealers are designated. Unlike other auction participants — including foreign central banks, pension funds, and asset managers who may bid directly — primary dealers are obligated to participate in every Treasury auction at competitive prices. More importantly for the monetary financing circuit, they are the exclusive counterparties for Federal Reserve open market operations; the Fed buys only from primary dealers, not from the broader universe of auction participants. The open-market requirement that makes dealer intermediation mandatory in the Fed purchasing channel is 12 U.S.C. §355.

7. Functional Finance: Abba P. Lerner, “Functional Finance and the Federal Debt,” Social Research, Vol. 10, No. 1 (February 1943), pp. 38–51. Lerner’s capacity-constraint formulation: fiscal policy should be adjusted “only insofar as this is needed for the purpose of keeping the rate of spending neither too low (so as to cause unemployment) nor too high (so as to cause inflation).” The Weimar and Zimbabwe hyperinflations are consistent with this framework: both occurred in economies that had lost productive capacity, making additional money creation inflationary regardless of the financing instrument used. The United States during QE1–QE3 had the opposite conditions: substantial slack, high unemployment, and deflationary pressure, producing the below-target inflation the framework predicts. The 2021–2023 inflation episode was primarily driven by pandemic-era fiscal transfers directly to households — bypassing bank intermediation and hitting goods markets directly — combined with supply chain disruption and energy price shocks, rather than by the QE reserves that had accumulated since 2009.

8. TBAC consolidated balance sheet framework: Treasury Borrowing Advisory Committee, Report to the Secretary of the Treasury, February 3, 2026. https://home.treasury.gov/news/press-releases/sb0385. The quoted language on SOMA holdings as “offsetting liabilities” appears in the discussion of privately held marketable borrowing.

9. Dallas Fed FRED series: Federal Reserve Bank of Dallas, “Market Value of Privately Held Gross Federal Debt,” FRED series MVPHGFD027MNFRBDAL. https://fred.stlouisfed.org/series/MVPHGFD027MNFRBDAL. Monthly data from January 1942.

10. FRED Blog decomposition: Federal Reserve Bank of St. Louis, “Who Holds US National Debt?” FRED Blog, March 6, 2025. https://fredblog.stlouisfed.org/2025/03/who-holds-us-national-debt/. The chart referenced in this essay shows four series as percent of GDP: gross federal debt, debt held by the public including the Federal Reserve, privately held debt excluding the Federal Reserve, and Fed holdings alone. The lower shaded band — the gap between publicly held and privately held — represents the Fed’s Treasury holdings: debt the government owes to an institution that remits most of its earnings back to Treasury. Before 2008 this band was negligible. After four rounds of QE it widened to several percentage points of GDP, narrowed during the quantitative tightening of 2017–2019, exploded during QE4, and has been narrowing again since 2022.

11. The Fed’s operating losses, suspended remittances, and deferred asset: when the Fed raises rates, it pays higher interest on the reserve balances that commercial banks hold at the Fed than it earns on the fixed-coupon bonds acquired during earlier QE rounds. Operating losses from late 2022 onward were recorded as a “deferred asset” — a bookkeeping entry representing future earnings needed before remittances resume, not a solvency event. The Fed continued meeting all obligations throughout by crediting reserve accounts. The deferred asset reached over $130 billion by mid-2024. Under consolidated accounting, the interest the Fed pays on bank reserves during such periods would appear in the federal budget as a transfer payment to the financial sector — a subsidy that in 2022–2025 exceeded $100 billion annually while appearing nowhere in the public accounts.

12. The remittances-as-levy observation: in Functional Finance terms, Fed remittances to Treasury function economically less like an investment return and more like a levy on private income streams — the government capturing a share of interest flows it inserted itself into by creating money, not by contributing real capital. The Cary Brown theorem, which shows that under full expensing the corporate tax becomes the government collecting its equity return on contributed capital, illuminates the contrast: Cary Brown involves a real contribution earning a real return; monetary financing involves a costless creation earning what is structurally a levy. The full implications of this parallel belong to the later essays on taxation and what a return on government co-investment actually means.

13. Gold revaluation as another hidden fiscal option. The United States holds approximately 261.5 million fine troy ounces of gold, nearly all of it monetized: gold certificates issued to the Federal Reserve equal the statutory value of the gold stock, currently $11.037 billion against a statutory value of $11.041 billion (the $4 million gap is a 100,000-troy-oz buffer set aside in 2002 as a bookkeeping safeguard). Per 31 U.S.C. §§5116–5117, the statutory price has been frozen at $42.2222 per troy ounce since 1973. At the market price of approximately $3,300 per ounce in 2025, the gold stock’s market value is roughly $863 billion — a latent fiscal capacity of approximately $852 billion accessible in principle through a statutory revaluation. This is not without precedent: the Gold Reserve Act of 1934 raised the statutory price from $20.67 to $35 per ounce, generating a windfall TGA credit used to capitalize the Exchange Stabilization Fund. In real (2025) dollar terms, the 1934 statutory price of $35 was equivalent to approximately $833 per ounce today; the 1980 market peak of $615 nominal was approximately $2,381 in 2025 dollars — substantial, but still below today’s market price of approximately $3,300. With holdings quantities nearly unchanged since 1980, the current total real market value of US gold holdings of approximately $863 billion is a genuine real as well as nominal all-time high. The current statutory price of $42.22, frozen for fifty-two years, represents less than 1.3% of the market price — a gap that grows larger in real terms with every year of inflation. The economic objections to revaluation today are substantially weaker than in 1934, because the dollar is no longer convertible to gold; a revaluation would be a pure accounting adjustment with no direct effect on exchange rates or the monetary system. The objections are primarily political: that it would make visible a large fiscal resource that “we can’t afford it” rhetoric requires to be invisible, and that it would blur the institutional line between monetary and fiscal operations that central bank independence doctrine requires to appear sharp. Congressional Research Service, “The Federal U.S. Gold Stock,” IF13109, September 23, 2025, https://www.congress.gov/crs-product/IF13109; Federal Reserve, “Does the Federal Reserve Own or Hold Gold?” https://www.federalreserve.gov/faqs/does-the-federal-reserve-own-or-hold-gold.htm.

14. Dealer revenue from Treasury market-making during QE periods is estimated at billions of dollars annually based on transaction volumes and typical spread ranges, but precise figures are not publicly reported as a separate line item in dealer financial statements. The Federal Reserve Bank of New York publishes dealer position data (FR 2004 reports) but not profit and loss by activity. The estimate follows from monthly Fed purchase volumes of $80–120 billion multiplied by typical round-trip spreads across auction and secondary market legs. This is an informed estimate grounded in published volume data rather than a verified figure.

This essay is the second in a series examining the gap between how American government actually works and how we are taught to think about it. The first essay, “The Trust Fund Is a Comfortable Lie — and We’re Paying for It in Confused Citizens”, examined how the Social Security Trust Fund encodes a gold-standard fiction into a fiat-currency world. This essay examines how the debt ceiling does the same thing — and how it became a weapon in the process.

Before the Ceiling: Congress and Its Two Hundred Bonds

The Constitution is clear about who controls the nation’s credit. Article I, Section 8 assigns to Congress — not the President, not the Treasury, not the Federal Reserve — the power to borrow money on the credit of the United States. The founders placed this power in the legislative branch deliberately, understanding that the commitment of a nation’s credit was too consequential to be left to executive discretion.

For the first 128 years of the republic, Congress took that assignment literally. Before 1917, every bond the Treasury issued required specific congressional authorization. Not a general delegation — a specific act, for a specific purpose, specifying the maturity date, the interest rate, the total quantity authorized, and the permitted uses of the proceeds. Between 1775 and 1920, Congress designed more than two hundred distinct bonds in exactly this fashion.¹ The Civil War required over a dozen separate borrowing acts between 1861 and 1865. The Spanish-American War required specific separate authorizations for short-term notes and long-term bonds. The 1902 Panama Canal bond specified thirty-year maturity at two percent interest — a level of congressional micromanagement that applied not just to major wars but to routine Treasury cash management, generating dozens of congressional sessions annually just to keep the government’s lights on.2

This was constitutionally faithful but operationally absurd. Congress was holding sessions to authorize the fiscal equivalent of rolling over a checking account overdraft. Something had to give.

Before examining what Congress chose to do, it is worth noting what it could have done instead — and could still do today. Every appropriations bill could carry a simple clause: “Treasury is authorized to borrow as necessary to fund this appropriation.” One vote, one decision, one moment of accountability. Members would own both the spending and its financing simultaneously. The political theater of voting for popular programs while performing outrage at the resulting debt would be impossible, because the two acts would be inseparable. The reason Congress has never done this — and the reason it protects the current arrangement — is precisely that the separation of the two votes is politically valuable. It allows members to be fiscally irresponsible in two directions at once, presenting both votes to constituents as consistent positions. That theatrical separation is not a byproduct of the debt ceiling; it is one of its primary functions.

1917: Congress Delegates What the Constitution Gave It

The standard account of the debt ceiling’s origin describes it as a limit on Treasury borrowing. This is the opposite of reality. The Second Liberty Bond Act of 1917 — the legislation that introduced the first debt limit in American history — was primarily an empowerment of Treasury: a grant of authority that the Constitution had assigned to Congress, now being delegated to the Executive. The ceiling was a guardrail on what Congress was giving away, not a restraint on what Treasury could do.

The distinction matters enormously. Congress was not asserting fiscal discipline over the Executive. It was relieving itself of a burden — the cumbersome, time-consuming work of approving hundreds of individual financing decisions — by making an unprecedented, but carefully limited, delegation of its own constitutional authority to the Executive. Congress benefited from this arrangement: freed from the micromanagement of Treasury financing, it could attend to other business. The debt ceiling was born not as an act of congressional oversight but as an act of congressional self-divestiture — an acknowledgment that Congress lacked the expertise, the time, and the institutional capacity to manage the detail of federal finance, and was therefore delegating that capacity to Treasury while retaining a nominal cap on the delegation.

The legislative record makes this unmistakable. Treasury Secretary William McAdoo testified before the Senate Finance Committee on September 11–12, 1917, asking Congress not to restrict him but to free him.3 His testimony is a sustained plea for discretion, flexibility, and latitude. He asked for authority over deposits, expenses, advertising, and debt instrument design. His recurring concern was that congressional micromanagement would “tie the hands” of Treasury at a critical moment. When challenged on specific provisions, he repeatedly said: give me the discretion; I will not abuse it; we must succeed.

The atmosphere surrounding the legislation was patriotic enthusiasm for government borrowing, not suspicion of it. The first Liberty Bond issue attracted over four million subscribers — ten times the pre-campaign estimate.4 Newspapers, banks, and department stores provided free advertising. Citizens bought bonds as acts of patriotism. The ceiling was introduced in a moment when the country was proudly lending to its government and actively looking for ways to make that lending easier — not worrying about whether its government was borrowing too much.

The 1917 act also represents a broader constitutional pattern that is rarely acknowledged. The modern administrative state rests on a series of congressional delegations to the Executive — to regulatory agencies, to executive departments, to the Federal Reserve. The debt ceiling is one instance of a general congressional habit: delegate a constitutional responsibility to the Executive when it proves cumbersome, then periodically threaten to reclaim it in the most destructive way available. The pattern repeats across the administrative state. Of all such instances, the debt ceiling is the most dangerous.

1939: The Modern Ceiling Is Born

The “debt limit” we actually argue about today — the comprehensive aggregate limit on all federal debt — was not created in 1917. The 1917 act set limits by instrument type: separate caps on bonds, notes, and certificates of indebtedness. The modern aggregate ceiling was created in 1939, when Congress passed the Public Debt Act establishing for the first time a single cap on the total amount of debt the Treasury could issue.5

The 1939 act was explicitly modeled on the demonstrated success of the 1917 delegation. Two decades of experience had shown that giving Treasury flexibility in debt management had worked well. Congress was not asserting new control; it was extending existing deference. As political scientist Sarah Binder has documented, Senate floor debates show the 1939 ceiling was designed to grant Treasury more flexibility in managing the mix of debt instruments — to let Treasury, rather than Congress, decide what combination of bonds, notes, and bills best served federal financing needs at any given moment. “Lawmakers recognized the limits of their financial know-how.”6

The direction of travel from 1917 to 1939 is unambiguous: more Executive discretion, less congressional micromanagement, greater Treasury flexibility. The ceiling set in 1939 was $45 billion — equivalent to roughly $1 trillion in today’s dollars.7 In nominal terms it now stands at more than $36 trillion, eight hundred times its original level; in real, inflation-adjusted terms, it has grown approximately thirty-five times. A limit raised approximately one hundred times, never meaningfully lowered, and now thirty-five times its original real value is not functioning as a constraint. It is functioning as a ritual. The ceiling was born in deference and extended in deference. It was a tool of congressional humility, not congressional discipline.

1953: The Weapon Is Discovered

On July 29, 1953, Senator Harry F. Byrd of Virginia took the Senate floor to oppose a request from President Eisenhower to raise the debt ceiling by $15 billion. The ceiling then stood at $275 billion. Congress was one day away from recess. The House had already agreed to the increase. Byrd objected.

His statement, entered into the Congressional Record that day and published in full in the New York Times the following morning, is the founding document of debt ceiling weaponization — and it is a precise demonstration of the category error that has defined the debate ever since.8

Byrd compared the federal debt of $272 billion to the total assessed value of all real estate in the 48 states — $143 billion — and applied municipal debt logic to a monetary sovereign. States, he noted, typically limit their municipalities’ debt to 18 percent of assessed real estate value. The federal debt was nearly twice the value of all real estate in the country. “A debt of $275,000,000,000 is as much or more than this country should be called on to stand.” What is striking, reading the Congressional Record, is that other senators engaged with this argument seriously rather than dismissing it — evidence of how completely commodity-money thinking still governed the chamber. What would be immediately recognizable as a category error today was treated as sound fiscal reasoning in 1953.

What is equally striking is that Byrd repeatedly encouraged Eisenhower to use executive orders to ration agency spending — urging the President to exercise unilaterally precisely the oversight over appropriations that the Constitution assigns to Congress. A senator arguing that the Executive should manage federal expenditures because Congress would not was, without quite realizing it, already enacting the constitutional inversion the debt ceiling would later institutionalize.

The argument is coherent — for a state, a county, a city, or a household. For any entity that cannot create its own currency, debt relative to asset value is a meaningful solvency measure. For the entity that issues the currency in which the debt is denominated, it is meaningless. Byrd was applying the logic of Virginia’s counties to a government that operates under entirely different fiscal physics. The category error was not obscure; it was fundamental. And it has been repeated, in essentially the same form, in every debt ceiling debate since.

In 1953, that error was at least historically understandable, as the next section will explain. But the government’s operational response confirmed how deeply commodity-money thinking still governed institutional behavior. With borrowing authority nearly exhausted, Treasury reduced expenditures across federal agencies and on November 9 monetized $500 million of free gold — gold held in Treasury vaults against which certificates had not yet been issued — by issuing those certificates and depositing them with the Federal Reserve, using the resulting account credits to retire $500 million of outstanding Treasury notes.9 The instinctive response to a fiscal bind was to reach for the gold.

The resolution: Congress returned from recess and approved a $6 billion increase — less than half of what Eisenhower had requested, but enough to restore operations. The first debt ceiling crisis was resolved. The template had been established.

The conservative press drew the explicit lesson within days. The Chicago Daily Tribune stated it plainly: “The lesson is clear — the way to get government expenditures down is to deny the administration authority to increase the debt.”⁹ The weapon had been discovered, named, and published. Everyone who read newspapers knew it was available.

The consequences arrived quickly. The 1957 debt ceiling fight forced the Air Force to drastically curtail spending. Economist Marshall Robinson of the Brookings Institution, writing in 1959, identified debt ceiling brinkmanship as a major cause of the 1957–58 recession.10 The mechanism is straightforward: forced spending curtailment is a contractionary fiscal shock; in a demand-constrained economy, demand withdrawal produces recession. Walter Heller, writing in 1958, was unsparing: “far from promoting fiscal prudence and expenditure restraint, as claimed by its protagonists, the federal debt limit has in fact eroded the integrity of our federal budget, interfered with efficient expenditure scheduling and effective debt management, endangered our defense program, and aggravated the 1957–58 recession.”11 That assessment is 68 years old. It reads as current events.

1971: The World Changed. The Weapon Remained.

The weapon Byrd discovered in 1953 at least made conceptual sense within the monetary world of its time. In 1917 and 1939, the ceiling was designed for a government that genuinely could run short of the commodity backing its currency. Congressional oversight of a fixed-supply resource was coherent. Byrd’s error was applying that logic after it had already been partially undermined — but the gold standard, though modified, was still a living institution in 1953.

That world ended on the evening of August 15, 1971, when President Nixon announced that the Treasury’s gold window — its standing offer to exchange dollars for gold at a fixed price — was permanently closed.12 The dollar became, irreversibly, a fiat currency issued by a monetary sovereign. The government’s fiscal constraint changed fundamentally: no longer the solvency constraint of a commodity-money system, but the inflation constraint of a fiat system. The right question for fiscal policy became not “do we have enough money?” but “will this spending cause inflation?”

The debt ceiling did not change with the world it was designed to describe. Neither did two additional legal choices that complete the architecture: the prohibition on direct Federal Reserve purchases of Treasury securities (12 U.S.C. § 355), and the requirement that Treasury spend only from its Federal Reserve account, replenished by taxing or borrowing.13 Both are statutory choices, not physical laws. The prohibition on direct purchases was enacted in 1935 partly to prevent the government from simply creating money to fund spending — a legitimate concern in any monetary system, including a fiat one. But the concern does not require the entire apparatus of artificial gold-standard constraints that has accumulated around it. Together, these three legal choices — the debt ceiling, the monetization prohibition, and the Treasury General Account requirement — produce a federal government that functions as if it were still on the gold standard: constrained by solvency rather than by inflation, forced to “find the money” before spending rather than spending and managing the inflationary consequences. None of this is economically necessary. All of it is legally chosen.

After 1971, then, the debt ceiling ceased to reflect any economic reality. It became a ghost — the institutional form of a constraint that had ceased to exist, haunting the fiscal debate without any underlying substance. Byrd’s error had been applying the wrong logic. Every subsequent debt ceiling crisis has been conducted using a concept that has not corresponded to economic reality for more than fifty years.

1979: The Gephardt Rule — An Admission of Dysfunction

Within eight years, the debt ceiling vote had become sufficiently toxic that Congress tried to procedurally neutralize it. The Gephardt Rule, named for Representative Richard Gephardt of Missouri, made debt ceiling increases automatic upon passage of the budget resolution: when Congress voted for a budget that required borrowing, it was deemed to have simultaneously authorized the borrowing. One vote, one decision, no separate opportunity for theater.14

The Gephardt Rule is itself an admission. Congress acknowledged, through the mechanism it created, that the separate debt ceiling vote served no useful fiscal purpose — that it was pure political theater, and sufficiently dangerous theater that it was worth suppressing procedurally. For twelve years this worked, more or less. Then came Newt Gingrich.

1995: Gingrich and the Anti-Democratic Turn

In 1995, Speaker Gingrich suspended the Gephardt Rule. Whether the calculation was entirely conscious or not, the effect was clear: restoring the debt ceiling as an independent vote recreated the hostage. The ceiling could now be used to extort spending concessions that the regular legislative process — introducing bills, defending them publicly, building majorities — might not deliver. Procedural leverage substituted for persuasion.

This substitution has a name: procedural extortion. Fiscal conservatism is a legitimate political position; there are serious arguments for lower spending and smaller government that should win or lose through democratic debate. What the weaponized debt ceiling does is remove that debate from the equation — not by persuading majorities but by threatening consequences severe enough that the majority concedes rather than tests the threat. The instrument becomes anti-majoritarian not because any particular actor intends it that way, but because that is how procedural hostage-taking functions structurally. The outcome is imposed on the democratic process rather than emerging from it.

It is worth noting, without belaboring the point, that Gingrich’s 1995 intervention was part of a broader assault on congressional institutional capacity whose consequences persist to this day: the defunding of committee staffs, the elimination of the Office of Technology Assessment, the radical centralization of power in party leadership, and the systematic degradation of Congress’s ability to develop independent expertise on complex policy questions. The debt ceiling was one weapon in that campaign. It was not the only one, and it was not chosen by accident.

In 2011, Speaker Boehner deployed Gingrich’s playbook against President Obama, and the weaponized ceiling reached its logical conclusion. Under the explicit threat of the first sovereign default in American history, the standoff produced the Budget Control Act — $917 billion in spending cuts over ten years.15 Standard & Poor’s downgraded U.S. debt for the first time in the nation’s history. The GAO later documented hundreds of millions in unnecessary borrowing costs from the 2011 and 2013 standoffs alone.16 These are the direct, measurable, monetary consequences of treating a legal fiction as an economic reality — money extracted from taxpayers and transferred to bondholders as the price of performing a crisis that served no economic function.

The constitutional inversion was now complete. The ceiling was designed as a constraint on Executive power — a limit on the delegation Congress had made to Treasury. After 1995 and 2011, it simultaneously constrains Congress and expands Executive power. Congress cannot honor its own appropriations without a separate vote that a minority can block; and when the ceiling binds, the Treasury Secretary must decide which bills get paid and which get deferred. That is the power of the purse — the power Article I assigns to Congress — being exercised unilaterally by the Executive, under duress, without authorization, because Congress’s own instrument has made it necessary. The Impoundment Control Act of 1974 was passed specifically to prevent presidents from unilaterally deciding not to spend appropriated funds.17 Debt ceiling crises force exactly that.

The instrument has inverted its own constitutional logic. Originally: Congress limits its delegation to Treasury. After weaponization: Congress appropriates spending; ceiling prevents honoring it; Treasury decides who gets paid. The ceiling went from constraining the Executive to empowering it at Congress’s expense — the precise opposite of its stated purpose, and an outcome the founders would have found alarming.

The Constitutional Absurdity

Strip away the history and the politics, and the debt ceiling presents a simple logical problem.

Congress votes to appropriate spending. The same Congress, separately, votes on whether to authorize the borrowing needed to cover what it has already appropriated. The two votes can contradict each other. When they do — when Congress votes to spend and then refuses to borrow — it has issued two irreconcilable instructions to the Executive. Treasury is ordered to spend appropriated funds and yet denied the authority to borrow the money to cover them. No coherent theory of constitutional government can justify this.

The House Budget Committee’s own fact sheet states the absurdity plainly: the debt ceiling “is a limit up to which Treasury can pay the bills and honor commitments that Congress has already made.”18 It does not limit future spending. It limits payment of past commitments. Congress authorizes; Congress then threatens to dishonor what it authorized.

The fix is simple and has been available all along: embed borrowing authority in each appropriations bill. “Treasury is authorized to borrow as necessary to fund this appropriation.” The constitutional responsibility returns to Congress where it belongs. It is Congress, not the Executive, who should decide what spending should be done and bear public accountability for that decision. Members cannot vote for a program and against its financing in the same breath. The theatrical separation that makes the current arrangement politically valuable would disappear — which is precisely why Congress has not done it.

The Name Is the Argument

George Lakoff has spent decades documenting how political language works not by describing reality but by pre-loading conclusions.19 “Tax relief” encodes taxation as an affliction before any debate about tax policy begins. “Pro-life” and “pro-choice” each occupy moral high ground before any argument about policy is made — one encoding the opponent as anti-life, the other as anti-freedom. The name forecloses the argument before it begins.

“Debt ceiling” works identically. It encodes a model of government finance — the household model, in which debt is a burden to be reduced through economy and sacrifice — before any debate about fiscal policy begins. Within that frame, the only available responses to a “ceiling crisis” are spending cuts or default. Tax increases are invisible: you do not solve a debt problem by spending more, and taxation is coded as spending — government taking money from people — rather than as revenue, government receiving a return on its co-production of economic value. Half the fiscal adjustment toolkit disappears before the debate begins.

The cognitive problem runs deeper than rhetoric. All mental models simplify reality — simplification is cognitively necessary, not inherently dishonest. The question is whether the simplification is symmetric or biased. “Debt ceiling” is asymmetrically biased: it makes the liability side of the government’s balance sheet vivid and the asset side invisible. A complete government balance sheet would show that $36 trillion in outstanding obligations coexists with assets — infrastructure, research institutions, the productive capacity of the educated workforce, the value of the rule of law and contract enforcement — that are orders of magnitude larger. The Congressional Budget Office does not publish a government balance sheet. The debt debate is conducted entirely on the liability side of a ledger whose asset side is never shown.

A 2024 study in PLOS ONE found that even numerically sophisticated citizens remain susceptible to debt ceiling framing effects.20 Cognitive sophistication does not inoculate against a defective mental model — it may amplify it, as people reason more efficiently within whatever conceptual framework they have been given. The public’s mental model of government finance is not false; it is insufficiently complex. People reasoning correctly within a defective model reach systematically wrong conclusions — not because they are unintelligent but because the available concepts are defective.

Consider what changes if the “debt limit” were called the “Public Investment Limit” — the cumulative total Congress has authorized for investment in public productive capacity. “We are approaching the Public Investment Limit” prompts: are we investing enough? What returns are we getting? Are there areas of underinvestment? These are productive questions. “We are approaching the debt ceiling” prompts: how do we cut spending? Who gets less? These questions presuppose the household model and produce austerity as the only legitimate answer.

The intergenerational argument reverses under the investment frame. “We are burdening our grandchildren with debt” becomes: are we investing enough in the economy our grandchildren will inherit? The post-World War II evidence is instructive. The highest debt-to-GDP ratio in American history — reached in 1945 — coincided with and preceded the fastest productivity growth, most rapid middle-class expansion, and most significant public investment in American history. The GI Bill, the Interstate Highway System, the research infrastructure that eventually produced the internet: all were funded in the shadow of “unsustainable debt.” Future generations did not suffer under that burden. They flourished because of that investment.

The austerity bias follows directly from the naming. Within the household model, spending cuts feel like fiscal responsibility and tax increases feel like admission of failure. Within the investment model, both spending reduction and revenue enhancement are symmetrically available responses to managing a portfolio. The debt frame systematically closes off one half of the fiscal toolkit — progressive taxation of concentrated wealth — in favor of the other: cuts to public services used by those without market power. This asymmetry is not accidental. It serves identifiable interests. The institutional incentives reliably reward continued use of the household model by those whose policy goals are advanced by austerity framing — and reliably punish those who challenge it with the accurate but politically costly language of monetary sovereignty.

What Would Be Better

The debt ceiling should be abolished. The constitutional alternative — borrowing authority embedded in each appropriations bill — is cleaner, more honest, and more faithful to the founders’ intent that Congress own its fiscal decisions. It is Congress, not the Executive, who should decide what spending should be done and bear public accountability for that decision.

But abolition alone is not sufficient. The real constraint on federal spending is inflation, and a responsible fiscal architecture needs to operationalize that constraint institutionally. The Functional Finance framework that Abba Lerner articulated in 1943 provides the right foundation: the correct criterion for fiscal policy is its economic effect — whether it produces full employment without inflation — not its conformity to any accounting rule about balanced budgets or debt levels.21

What follows is a sketch of the direction rather than a complete institutional design — the full architecture deserves its own treatment. The core idea is to replace a nominal dollar ceiling with an inflation constraint that has real economic content. Congress would mandate that the Congressional Budget Office publish, alongside its standard budget projections, estimates of the economy’s productive capacity and the current inflationary or deflationary gap — the distance between actual and potential output. This is technically difficult; measuring the output gap involves genuine uncertainty and honest disagreement among economists. But it is not categorically harder than what the Federal Reserve does for monetary policy, and we have accepted that difficulty there. A fiscal mandate would then replace the debt ceiling: Congress should keep net spending — spending minus taxation — within the non-inflationary range as estimated by the CBO. This is a real constraint, with real economic substance, unlike the nominal dollar ceiling, because it is tied to an economically meaningful quantity. Democratic accountability would be preserved through periodic congressional votes on whether the CBO’s assessment of fiscal space is correct, and whether proposed spending falls within it.

The difficulties are real and should be named honestly. The output gap is hard to measure. Inflation is not the only relevant social consequence of spending. The politics of delegating fiscal assessment to an independent body are severe. But the question of what is right is independent of what is currently possible. We accepted the institutional complexity of central bank independence because congressional management of interest rates was worse. The alternative to a Functional Finance fiscal constraint is the debt ceiling — and the debt ceiling is demonstrably worse. The logic is the same.

The Problem Is Still Congress

The debt ceiling has been raised or suspended approximately one hundred times. It has never meaningfully constrained spending. The debt has increased under every presidential administration since Herbert Hoover.

What the ceiling has produced is measurable and real: two credit downgrades, recurring market disruption, at least one recession attributable in significant part to its misuse, and billions of dollars in unnecessary borrowing costs paid to bond markets as the price of performing a ritual that serves no economic function. The GAO documented $1.3 billion in excess borrowing costs from the 2011 standoff alone. In 2013, between $38 million and $70 million more. These are not hypothetical costs. They are money extracted from taxpayers and transferred to bondholders because Congress chose to perform a crisis rather than govern.

The One Big Beautiful Bill Act of July 2025 raised the ceiling by $5 trillion — the largest single increase in the ceiling’s history — while simultaneously, through its reduction of income taxation of Social Security benefits, accelerating the Social Security Trust Fund’s projected insolvency. The party most associated with fiscal responsibility achieved both in a single bill. The ceiling was not a constraint. It was a prop.

This is not a new observation. On the Senate floor in 1953, during the very first debt ceiling crisis, Senator Homer Capehart of Indiana — a conservative Republican — made the constitutional accountability argument with characteristic bluntness. He noted that Congress had appropriated every dollar of the debt that now outraged it: “The President of the United States — whether it be Mr. Truman, or President Eisenhower — cannot spend a single penny unless it is appropriated by Congress.” As for those senators who had voted for every expenditure and were now crying alarm about the resulting debt, Capehart had a suggestion: “Go home tonight and take a good look in the mirror, and you will find who is responsible for this situation.”22

The instrument was designed in 1917 to make governing easier. It has made governing more difficult, more contentious, and more dangerous. It was designed as a constraint on Executive power. It now simultaneously constrains Congress and empowers the Executive. It was born as a tool of congressional humility about finance. It is now a weapon of congressional extortion against the public interest.

The institutional incentives that maintain this arrangement are powerful and self-reinforcing. Those who benefit from austerity framing — from a public that cannot distinguish between a monetary sovereign and a household, that accepts “we can’t afford it” as an answer to questions about public investment, that watches recurring hostage crises and concludes that federal finance is genuinely precarious — have every reason to preserve the instrument that generates that confusion. The debt ceiling manufactures the crisis that justifies austerity.

It is worth being precise, one final time, about what the debt ceiling actually is. It is not a fiscal guardrail. It is not a constitutional check. It is not a meaningful limit on spending or debt. It is a legal artifact, born of an unprecedented delegation of congressional authority, designed for a gold-standard world that has not existed for more than fifty years, repurposed as a weapon by a Speaker who found procedural leverage more reliable than democratic persuasion, and maintained by those who benefit from the public’s inability to understand the difference between a monetary sovereign and a household.

The problem has never been the ceiling. The ceiling is empty. The problem is Congress — and a political culture that has preferred a comfortable fiction, a defective mental model, and a corrupted vocabulary to the honest conversation about public investment, democratic accountability, and shared obligation that a self-governing people are owed.

That conversation is overdue. It will go better when the people having it understand what the words actually mean.

“In the name of budgetary integrity, financial prudence, adequately financed national security, and aggressive policies to combat inflation and counter recession — in other words, in the name of everything that is fiscally holy and wholesome — our anachronistic federal debt limit should be abolished.”

— Walter W. Heller, Chairman, Department of Economics, University of Minnesota, 195823

Notes

A government that cannot understand its own fiscal capacity cannot govern in the public interest. This essay examines how Congress built that misunderstanding into law.

When Franklin Roosevelt’s advisors designed Social Security in 1935, they deliberately made it look like an insurance system — even though it was never structured like any viable insurance company.1 A real insurer must maintain actuarial reserves: assets sufficient to cover projected liabilities as they come due. Social Security was designed from the start to operate as a pay-as-you-go transfer system, collecting from current workers and paying current retirees. Roosevelt understood this. The insurance framing was a political choice, not an actuarial description or an economic necessity — and Roosevelt said so, privately. When Luther Gulick, a public administration expert, challenged the payroll tax as economically indefensible, Roosevelt replied: “I guess you’re right on the economics, but those taxes were never a problem of economics. They are politics all the way through. We put those payroll contributions there so as to give the contributors a legal, moral, and political right to collect their pensions and their unemployment benefits. With those taxes in there, no damn politician can ever scrap my social security program.”2

Workers would pay in; the fund would accumulate and earn compounded interest; future retirees would draw down. The mechanism made intuitive sense — because in 1935, it was intuitive, and it was correct. The United States federal government was still operating under a gold standard. Money was, in a meaningful and literal sense, a thing the federal government could run out of. Saving it up before it was needed was not just prudent — it was the only honest way to design a long-term commitment. The people who built Social Security were not naive; they were working with the only economics available to them.

But that world was already beginning to dissolve. The domestic gold standard had been weakened in 1933, when Roosevelt signed Executive Order 6102, requiring Americans to surrender their gold to the Federal Reserve, effectively ending private gold convertibility.3 Bretton Woods in 1944 maintained an international gold peg for the dollar — foreign central banks could still exchange dollars for gold at $35 per ounce — but the tether grew increasingly strained through the 1950s and 1960s as other nations accumulated dollar claims that exceeded American gold reserves.4 The end came on the evening of August 15, 1971, when Richard Nixon, following a secret three-day meeting with senior economic advisors at Camp David, announced that the Treasury’s gold window — its standing offer to exchange dollars for gold at a fixed price — was closed. Foreign governments could no longer redeem their dollar holdings for gold. The gold window was never reopened.5 The dollar became, permanently and completely, a fiat currency issued by a monetary sovereign. The institutional structure Roosevelt had correctly built for a gold-standard world remained. The economic reality it had been designed to reflect did not.

What Monetary Sovereignty Actually Means

A monetary sovereign is a government that issues its own currency, borrows in that currency, and faces no external constraint on its ability to create more of it. The United States federal government is a monetary sovereign. It cannot run out of dollars the way a household runs out of dollars — or the way a gold-standard government could run out of gold — because it is the source of dollars. Its real fiscal constraint is not solvency but inflation: it can always spend, but spending beyond the economy’s productive capacity generates inflation. A monetary sovereign never needs to ask ‘can we afford this spending?’ It must ask instead whether the economy has the real productive capacity to absorb that spending without generating unacceptable inflation.

This is not a radical claim. It is the operating reality of every major central bank and treasury in the developed world.

Before proceeding, one distinction is worth clearing up, because federal fiscal rhetoric constantly blurs it: subordinate governments — states, counties, municipalities — are not monetary sovereigns. They do not issue currency. They cannot create money. Like households and businesses, they can only spend what they can tax, receive in transfers, or borrow. And because they face real future obligations — pension payments, infrastructure repair, debt service — they often must do what households do: save in advance. A state’s rainy day fund is genuinely useful. A county’s reserve account is genuinely necessary. A city that accumulates assets against future pension liabilities is doing something economically meaningful, for exactly the same reason a family saves for retirement: because neither can just create money when the bill comes due. The fiscal logic that applies to every level of government below the federal level simply does not apply one level up.

The Trust Fund as Scorekeeping

The Social Security Trust Fund does not contain money in any economically meaningful sense. It contains special-issue Treasury securities — a particular class of non-marketable government obligations that exist solely as intra-governmental accounting instruments. Unlike the Treasury bonds, notes, and bills available to the public, these securities cannot be sold on the open market and their interest rates are set by a statutory formula rather than by market conditions.6 They are, in the most precise sense, merely IOUs from one part of the federal government to another.

When payroll taxes exceed benefit payments, the surplus is “invested” in these securities, which means the Treasury spends the money and records a liability back to the Trust Fund. When benefit payments exceed payroll taxes, the Trust Fund redeems these special unmarketable securities, which means the Treasury covers the gap from general revenue or new borrowing. At no point in this process is there a vault, a “lock-box”, an account at a bank, or a pool of saved resources waiting to pay future retirees. There is only a ledger entry.

The government’s ability to pay Social Security benefits in 2040 or 2060 without causing inflation will depend on two things: the productive capacity of the economy and Congress’s willingness to authorize the spending. A monetary sovereign always has the nominal capacity to pay — it is the source of the currency. What changes over time is the real economic impact of those payments. The Trust Fund balance will have no independent effect on either.

This means the famous Trust Fund “insolvency” dates — the years when the fund is projected to run dry — are accounting events, not economic ones. Current projections put the Old-Age and Survivors Insurance Trust Fund insolvency at roughly 2033, at which point continuing payroll tax revenue would cover only approximately 77 - 81 percent of scheduled benefits. The shortfall would represent, for a typical retiring couple, something on the order of $18,000 in lost annual benefits.7

The mechanism that produces this outcome is worth examining closely, because it turns out to be traceable to a single provision of federal law. Under 42 U.S.C. § 401(h), all retirement and survivor benefit payments “shall be made only from the Federal Old-Age and Survivors Insurance Trust Fund.”8 When the Trust Fund is empty, that clause — as currently written — means benefits cannot be paid in full, because there is no other authorized source. The “crisis” is not an economic inevitability; it is a legal artifact. A simple amendment replacing “shall be made only from” with “shall be made from [the Fund], or if that Fund is insufficient, from the general fund of the Treasury” would dissolve it entirely. The underlying economics would be unchanged. Only the accounting instruction would differ.

Congress could resolve this tomorrow — by raising the payroll tax, by adjusting benefits, or by amending that clause. The economic capacity to pay full benefits without causing inflation would be unaffected by any of those choices. Only the legal permission structure would change.

Payroll taxes, in this light, are best understood as what they actually are: federal revenue. They offset the inflationary pressure of federal spending, as all federal taxation does. They do not “fund” Social Security in the sense of accumulating resources that later “pay for” benefits. This year’s payroll tax receipts are this year’s federal revenue. Once collected, they are indistinguishable from any other tax dollar — American currency comes in only one color. The payroll tax dollar and the income tax dollar are identical once they enter the Treasury. Next year’s benefit payments will be next year’s federal spending, authorized or not by whatever Congress then sits.

The Property Claim That Isn’t

Consider Ida May Fuller of Ludlow, Vermont — the first person to receive a Social Security retirement benefit.9 She paid into Social Security for three years, from 1937 to 1939, contributing a total of $24.75 in payroll taxes. Her first monthly check, received in January 1940, was $22.54 — nearly equal to her entire lifetime contribution, paid out in a single month. She lived to be 100, and collected $22,888.92 in total benefits before she died in 1975.

Miss Fuller’s case is extreme, but the general pattern is not unusual. Most Social Security recipients — particularly those with average or below-average wages who live to typical ages, which describes most recipients — receive substantially more in lifetime benefits than the actuarial present value of their payroll tax contributions, even assuming compound interest. No viable insurance policy can make such a promise. The program is redistributive, intentionally so. That is one of its genuine virtues.

Yet one political function of the Trust Fund framing is to give beneficiaries a quasi-property claim: I paid in, so I am owed a return. This claim is comforting, but doubly false.

At the macro level, the money paid in did not “go somewhere” to be retrieved years later when needed. It was tax revenue like any other. The government’s obligation to pay future benefits is a political commitment backed by the productive capacity of the future economy — not a retrieval of saved funds from a vault.

At the individual level, the actuarial math does not support the claim even on its own terms. Most recipients receive far more than they contributed. The “I’m just getting back what I paid in” story is simply inaccurate for most people.

And there is a deeper problem with the framing: any taxpayer has contributed to the general revenue that funds government operations. The payroll tax creates no stronger moral claim to Social Security benefits than income taxes create to highway benefits, or than excise taxes create to national defense. The earmarking is an accounting convention — a political device — not a property relationship. Accepting the “I paid in” framing uncritically means accepting a fiction that privileges one bookkeeping convention over another.

A Litany of Federal Savings Fictions

The Social Security Trust Fund is the most politically charged example of a federal “savings” mechanism, but the same analysis applies to all others.

The Medicare Hospital Insurance Trust Fund (Medicare Part A) operates much like the Social Security Trust Fund. Its projected insolvency dates generate the same alarm, serve the same accounting-trigger function, and have the same non-relationship to the government’s actual capacity to fund health care for the elderly.

The Highway Trust Fund and the Airport and Airway Trust Fund present a related but distinct fiction: the idea that users “pay for” these services through earmarked excise taxes. The history of the federal gas tax illustrates how layered this fiction is. When President Herbert Hoover signed the federal fuel tax into law in 1932, it was intended as a temporary measure to cover national defense spending — not highways. Persistent budget deficits kept it in place, and it was eventually redirected to road maintenance through the Highway Trust Fund.10 The tax was never designed for the purpose we are now told that it serves.

Nor does it actually serve that purpose. Because the tax rate has not changed since 1993, the revenue it raises does not come close to meeting road maintenance needs, and Congress regularly supplements the Trust Fund from general revenue. The earmark creates the political impression of self-financing without the economic reality of it.

The fiction becomes visible in moments of political stress. When President Trump proposed suspending the federal gas tax in May 2026 to offset oil price increases driven by his confrontation with Iran, the debate in Congress centered on whether drivers would actually save money — not on whether highway construction would stop. Everyone understood that the spending would continue regardless of what happened to the earmarked revenue. “Nobody wants to create a hole in the Highway Trust Fund,” one economist noted — but the hole, if it appeared, would simply be filled from general revenue, as it routinely is already. The spending and the earmark are separate things: one a policy commitment, the other an accounting convention. Everyone in Washington understands this, even if the public is not encouraged to.

Proposals for a federal sovereign wealth fund — periodically floated as a way to “invest” government surpluses — rest on a category error. A monetary sovereign that runs a surplus is not accumulating wealth in any sense that parallels household or state savings; it is withdrawing net financial assets from the private sector. What such a fund might accomplish is federal investment in productive assets — which may or may not be sound policy — but it is not “saving” in any economically coherent sense at the federal level.

The pattern across all these cases is consistent: the federal government adopts the institutional form of saving because saving is legible, politically useful, and deeply intuitive to citizens whose own fiscal lives are genuinely constrained. But the “savings” form does not reflect the underlying economic reality at the federal level.

Why the Fiction Persists — and Why That’s a Problem

The Trust Fund fiction persists because it does real political work. It provides an accounting structure that forces periodic congressional attention to long-term commitments. It frames a benefit that might otherwise be politically vulnerable as an earned entitlement. It creates friction against benefit cuts — and it provides a seemingly reasonable, though false, justification for why Congress cannot increase benefits, even when the real resources and productive capacity exist to support them. None of these functions are trivial.

So the argument here is not that the Trust Fund should be abolished, or that Social Security should be reframed as a general revenue program. The institution has protected a genuinely valuable program for ninety years, and that matters. Reasonable people can disagree about whether the political benefits of maintaining the fiction outweigh its costs.

The argument is about those costs. And the primary cost is civic rather than fiscal.

The Trust Fund fiction teaches citizens that the federal government operates like a household or a pension fund: that it must collect before it can spend, that it can “run out of money,” that its long-term commitments are constrained by accumulated savings rather than by real resources and political will. These lessons are false. And because they are systematically taught — through every news cycle about insolvency dates, every congressional debate framed around “fiscal responsibility,” every political argument that treats federal solvency as analogous to personal solvency — they produce a citizenry whose intuitions about government finance are reliably wrong.

This matters for democracy. Citizens who believe the federal government operates under a household budget constraint will make systematically different political choices than citizens who understand it operates under an inflation constraint. They will support austerity during recessions, when expansion is warranted. They will accept false explanations for why Congress cannot act — why their parents or grandparents cannot receive the increased benefits that might make the difference between a decent old age and one spent rationing medications. They will misread every major fiscal debate. They will be, in a precise sense, less informed participants in democratic self-governance — not because they are unintelligent, but because the institutional language of their government has been systematically misleading them for ninety years.

The Problem Is Congress

It is worth being precise about what the Trust Fund insolvency “crisis” actually is — and what it is not.

It is not a shortage of money. There is no money in the Trust Fund — there never has been. There are unmarketable Treasury securities: accounting entries recording that one part of the federal government owes another part of the federal government a sum that the first part will cover by taxing, borrowing, or spending as it does for everything else. When those accounting entries reach zero, no vault empties. No reserves are depleted. The federal government’s capacity to pay Social Security benefits without causing inflation is exactly what it was the day before “insolvency.”

What changes at insolvency is simpler and more uncomfortable: Congress loses its excuse.

For decades, the Trust Fund fiction has allowed Congress to avoid a straightforward decision. Social Security is a transfer program. It moves money from current taxpayers to current retirees. Its long-term fiscal balance depends on the ratio of workers to retirees, on wage growth, on political choices about benefit levels and tax rates — not on a ledger balance in a non-marketable securities account. The program’s designers knew this. Roosevelt knew this. He said so, privately, to Luther Gulick.

The insolvency date is not an economic event. It is a legislative deadline — a moment when the accounting fiction that has allowed Congress to defer real decisions about Social Security will no longer hold. At that point, Congress will do what it has always had the power to do: raise the payroll tax, adjust benefits, broaden the funding base, or amend the single clause in 42 U.S.C. § 401(h) that currently prohibits payment from general revenue when the fund runs dry. Any of these is a genuine policy choice, with genuine distributional consequences, that should be made through genuine democratic deliberation.

The problem has never been the fund. The fund is empty. The problem is Congress — and a political culture that has preferred a comfortable fiction to an honest conversation about what we owe each other and how we will pay for it.

That conversation is overdue. And it will go better if the people having it understand what is actually at stake.

Notes

The Surplus Capacity Tax: Principle over Compromise

Under fiat currency, taxation’s economic function is not to fund government but to constrain aggregate consumption and thereby control inflation — the central insight of Abba Lerner’s Functional Finance framework. A tax should therefore be evaluated by its effect on consumption, not by revenue raised.

Equity requires that the burden of reduced consumption be distributed in proportion to each taxpayer’s ability to pay. The Benefit Principle and the Ability-to-Pay Principle, long treated as rivals, are in fact the same idea: each person’s economic position above the Hobbesian baseline — a normative reference point: the state of nature in which government does not exist and surplus accumulation is impossible — is simultaneously the measure of benefit received from the institutional order and the measure of ability to pay. Those who have achieved more within that order have benefited more from it, and bear a correspondingly greater obligation to sustain it.

Ability to pay, so understood, is measured by surplus consumption capacity — the fraction of income not required for consumption, as proxied by (1 − PC), where PC is the propensity to consume, whether measured as a marginal (MPC) or average (APC) rate. Because (1 − PC) rises with income, equal proportional sacrifice in consumption requires higher rates on higher incomes. A tax indexed to (1 − PC) appears progressive when measured against income, but it is strictly proportional when measured against what a monetary sovereign’s taxation actually constrains: consumption capacity. A tax designed on this principle is a Surplus Capacity Tax.

Taxation must not be ruinous. Every person is entitled to a subsistence exemption — a living-wage deduction sufficient for a minimal dignified life. Those whose incomes fall below subsistence should receive a negative tax (analogous to the EITC) rather than paying tax. No marginal rate should reach 100%, as that would eliminate the incentive to earn additional income. These constraints define the floor and ceiling of the Surplus Capacity Tax rate schedule.

Above the subsistence floor, the base Surplus Capacity Tax equalizes the sacrifice of consumption capacity across all taxpayers — not the sacrifice of income, but specifically the fraction of consumption capacity that taxation removes. Equity further supports a progressive surtax, reflecting the fact that wealthier individuals have derived greater benefit from the social and legal infrastructure — property rights, contract enforcement, stable currency, public investment — that makes the accumulation of surplus capacity possible in the first place. Those who have benefited more from that infrastructure bear a correspondingly greater obligation to sustain it.

This surtax should follow a sigmoid curve (an S-curve), which is the natural mathematical form satisfying three constraints simultaneously: it rises slowly at low and middle incomes, preserving the incentive to work and accumulate; it accelerates above a politically determined threshold, where surplus capacity is sufficiently large that equity demands a steeper contribution; and it approaches asymptotically a legislatively fixed maximum rate, ensuring that no marginal rate reaches the point of confiscation. The sigmoid is not an arbitrary choice — it is a natural and tractable class of curves that respects all three constraints at once, though the precise parameterization remains a matter of political determination.

For corporations, the same logic applies with one substitution. The normal competitive rate of return plays the role that subsistence income plays for individuals: it is the minimum return capital requires to remain in productive use, just as subsistence is the minimum income a person requires to remain a functioning member of society. Accordingly, returns below the normal rate should be exempt from the Surplus Capacity Tax. Above that threshold, the excess of investor returns over the normal rate measures corporate surplus capacity, and a sigmoid surtax rises from zero to a legislatively fixed maximum, following the same form as the individual surtax. The principle — tax surplus capacity, spare subsistence, preserve incentives — is the same in both individual and corporate cases.

The general form of the Surplus Capacity Tax rate schedule is:

t(X) = t_min + (t_max − t_min) · S(X)

where X is surplus capacity as defined for the taxpayer class, S(·) is a sigmoid function, t_min is the minimum rate, and t_max is the legislatively fixed ceiling. For individuals, X = (1 − PC(I)), where I is income above subsistence and PC(I) is the propensity to consume at income level I, whether measured as a marginal (MPC) or average (APC) rate; when PC > 1, X becomes negative and the formula delivers a transfer automatically, without special cases. For corporations, X = (R − R_n), where R is the actual rate of return and R_n is the normal competitive rate of return.

The result is a unified tax framework, governed by a single principle and a single mathematical form, applicable to both individuals and corporations under the name Surplus Capacity Tax. Its purpose is not to prescribe a perfect tax system but to establish an objective baseline — grounded in principle rather than political compromise — against which any actual tax system can be measured, its deviations identified, and its fairness debated.

Imagine that a Supreme Court decision has just changed how courts interpret a standard testamentary capacity clause — the language in a will that establishes whether the testator understood what they were signing. The decision narrows the standard of evidence required to challenge a will on capacity grounds. Millions of existing wills may now be more vulnerable to contest than their authors believed.

How does that information reach the people who need it? Through law review articles, bar association notices, continuing legal education (CLE) seminars, and word of mouth across the profession. It is a slow, uneven, lossy process, but it exists. A lawyer who drafted that will has professional obligations, relationships, and channels through which the news will eventually arrive. The legal profession’s mechanisms for propagating a ruling are imperfect, but they are real.

Now consider the person who asked an LLM to draft their will last year. They used one of the general-purpose AI assistants — a perfectly reasonable thing to do, and the resulting document looked entirely professional. Call it vibe-writing¹ — the document-drafting equivalent of vibe-coding: natural language in, plausible artifact out, no verified constraints in between. That person has no professional obligation to stay current on estate law. No bar association will notify them. No CLE requirement will surface the ruling. The will still looks correct. It may not be.

There is a subtler problem underneath that one. Even if the LLM that drafted the will had encountered the SCOTUS ruling — even if it exists somewhere in the model’s training data — the model would still tend to generate the old clause language, because the training distribution is overwhelmingly weighted toward the old boilerplate. Hundreds of thousands of prior documents use the superseded language; a handful use the new. Frequency beats recency in a probabilistic system. The model does not retrieve the current standard; it reconstructs the most statistically probable output from everything it has ever seen.

This is an architecture gap, not a knowledge gap. Injecting the ruling at query time through retrieval-augmented generation helps but does not solve it: the probability mass of the prior distribution still dominates generation. The only fix is to bypass generation entirely for verifiable components: to retrieve the current clause rather than reconstruct a probabilistic approximation of it.

Sophisticated law firms already understand this, at least implicitly. A substantial commercial ecosystem of clause library tools has grown up to address exactly this problem within professional practice: ClauseBase, Henchman/Lexis Create+, Draftwise, Spellbook, and others.² These tools allow firms to build, version, and track verified clause libraries; to surface pre-approved language at drafting time; to benchmark clauses against market-standard terms. The private library layer the spec library describes is already commercially well-developed. The gap is not in the tools; it is in what the tools do not do: they are firm-centric and backward-looking. When the SCOTUS decision changes the meaning of a clause, none of them automatically notify every firm and every individual who has deployed that clause. That alert mechanism is missing.

The contrast is striking. A law firm using ClauseBase has curated, versioned, tracked clause libraries and some process (however informal) for learning about legal developments and updating its library. The individual who asked an LLM to draft their will has none of these protections, no professional update channel, and no awareness that anything has changed.

These are solvable problems. But before proposing a solution, it is worth stating what kind of solution is being sought: what counts as a real answer rather than a speculative one. In 1976, Russell Ackoff, designing an idealized system for scientific communication, proposed two constraints that any viable redesign must satisfy: it may use only technology known to be feasible today, and it must be operationally viable — capable of functioning if it were actually built. In his words, such a design is “not a work of science fiction; it is a feasible, however improbable, design.”³ Both constraints apply here.

Now imagine that testamentary capacity clause language were standardized and versioned — a verified entry in a shared library, accessible to LLMs, identified by a stable version number, with a record of which documents deployed it. The decision arrives. The spec maintainer annotates the entry: clause v4.2, interpretation changed by Smith v. Jones, decided [date], challenge threshold narrowed in these specific circumstances. Every practitioner and individual registered as a user of clause v4.2 receives an alert. The review is triggered automatically, not eventually. And when the next LLM generates a testamentary capacity clause, it retrieves the current version — not a probabilistic reconstruction of the old one.

This is not a fantasy. It is the direct consequence of taking the spec library seriously. And it is not only about legal documents.

Figures 1a and 1b show why this matters in practice.

Figure 1a shows the process of LLM artifact generation with which we’ve all recently become familiar. The LLM is trained on a massive corpus of information — of good, bad, and indifferent quality and accuracy. The artifacts it produces are WAGs (“Wild Ass Guesses”): probabilistic reconstructions with no verified grounding, assembled from whatever patterns the training distribution happens to contain.

Figure 1a. Current LLM Generation. WAG (“Wild Ass Guess”): probabilistic reconstruction with no verified grounding, tagged for human review.

Figure 1b shows the proposed architecture. Here the LLM draws on two distinct sources: the broad training data common to current systems, and a spec library — a structured collection of verified, versioned components retrieved at generation time rather than baked into the model’s weights. The output artifact is then a combination of three kinds of material: the user’s specific inputs, retrieved verified components, and WAGs where no verified component exists. Critically, the LLM reports the source of each kind of material to the user — ideally in the body of the artifact itself, as comments in generated code, as margin annotations in a generated contract, or as explicit provenance metadata in any other kind of document.

Figure 1b. Spec library architecture. Three sources feed a constraint layer that routes each component to the appropriate generation path.

The pattern in Figures 1a and 1b holds across every domain where LLMs are asked to generate professional artifacts. Each artifact combines two kinds of material: verified components and the results of novel inference. A sorting algorithm implementation relies on known algorithmic methods; a legal contract relies on established clause language; a clinical protocol relies on accepted dosing guidelines. The novel contribution lies in how these components are applied to a specific situation.

Human practitioners generally know where this boundary lies. They recognize when they are applying established practice and when they are exercising judgment. LLMs do not. When asked to generate an artifact, LLMs reconstruct both parts from their training distribution. The known components — the algorithm, the legal clause, the dosing protocol — are reconstructed probabilistically from examples seen during training, rather than retrieved as verified components. (This has a well-known technical term: WAG, or “Wild Ass Guess.”⁴) That reconstruction is where many silent and sometimes deadly errors originate. The pattern is the same across domains: the boundary between the library’s contents and WAGs is invisible to the automated generator. Making it explicit eliminates a large and consequential class of silent errors — not every LLM failure, but a class large enough to matter enormously in professional practice.

This architecture appeared in Essay 1⁵ as a solution to AI-assisted software development failures. The same architecture applies whenever an LLM is asked to generate consequential artifacts from established knowledge in any domain built on stable, reusable components. That covers most professional fields.

One further point before we proceed: private libraries are part of this ecosystem from the start. A law firm’s proprietary clause library is not in tension with the spec library concept. It is the organizational layer of a hierarchy running from public foundational knowledge through standards bodies through organizational policy to individual practice. The framework accommodates private knowledge. It does not require contributing it to a commons.

For readers arriving without having first read Essay 1: a spec library is a versioned, human-readable collection of specifications that constrain what an LLM generates — the way a type system constrains what a compiler accepts. When an LLM generates code, a legal clause, or a medical protocol, it fills unspecified gaps with probabilistic inference. The spec library approach makes the boundary between verified components and the results of novel inference explicit, so that the verified parts are retrieved rather than reconstructed, and the novel parts are clearly marked as requiring human review.

II. Vannevar Bush’s Unsolved Problem — and Why It Matters Now

In July 1945, Vannevar Bush published an essay in The Atlantic that has shaped how we think about knowledge and machines ever since.⁶ His diagnosis was precise: human knowledge had outgrown its organizational infrastructure. The problem was the organizing principle, not the volume: alphabetical and categorical classification was misaligned with how knowledge actually works, which is by associative link. One idea leads to another not because they share a category but because they share an associative relationship. The path between ideas is itself knowledge.

Bush proposed the Memex: a machine for storing, navigating, and sharing a personal library through associative trails. A researcher could follow connections between documents, record the path as a named trail, and share that trail with colleagues. The trail itself — the record of which associative links were productive — would become an intellectual contribution. Bush imagined that building such trails would become the primary scholarly contribution: synthesis of existing ones into navigable, shareable paths rather than discovery of new facts.

Why does this matter now? Because the problem Bush diagnosed has compounded. Knowledge is no longer merely hard to navigate; it is being consumed directly by machines that generate new artifacts from it. The organizational failure Bush identified now has a second consequence he could not have anticipated: machines filling the gaps in unorganized knowledge with probabilistic inference (WAGs), invisibly, at scale.

The partial realizations of Bush’s vision each solved part of the problem. Ted Nelson’s hypertext formalized the associative link as the hyperlink. Tim Berners-Lee’s World Wide Web deployed it at global scale. The W3C’s Annotations specification added a layer for attaching meaning and commentary to those links — a closer approximation of Bush’s trails, though still underused in practice. The Semantic Web project attempted full machine-readable semantics through formal ontology languages (RDF and OWL), but their overhead proved too high for mainstream adoption.

What all of these preserved was the link. What none delivered at scale was Bush’s annotation: an associative link that carries meaning, epistemic status, provenance, and verification.⁷ A hyperlink tells you a connection exists. It does not tell you what the connection means, how reliable it is, who established it, or when it was last checked.

There is a distinction worth naming. Bush’s Memex was a knowledge navigation tool, helping human researchers find and connect existing material. The spec library is a generation constraint, helping machines produce new material reliably from what has been verified. These are related but not identical. The spec library extends Bush’s associative trail to an active consumer he could not have anticipated: a system that does not merely traverse the knowledge graph but generates new nodes in it, constrained by what has already been established.

The formal methods tradition in computer science has argued for decades that separating specification from implementation is the precondition for reliable construction — from Hoare logic and Design by Contract to TLA+ and OpenAPI schemas. These are powerful but demanding tools; most practitioners never use them. The spec library pursues the same architectural principle with a far lower barrier to entry: natural language specifications and structured metadata rather than formal logic or proof systems. The target of construction has changed; the architectural insight has not.

What the Semantic Web failed to deliver at scale, the spec library achieves through pragmatic design: natural language specifications with structured metadata, versioned and attributed, machine-consumable without requiring formal logic. Bush’s associative trails, made semantic, versioned, verified, and connected to the generation process. The dependency relationship between specs carries meaning, not just adjacency. The verification carries epistemic status. The provenance carries authorship and history.

III. The Pattern Across Domains

Practitioners across every consequential professional domain already compose artifacts from verified fragments rather than generating from scratch. This is not new. What is new is that LLMs are now being asked to perform the same generation tasks — without access to the kind of verified-component infrastructure that many professions have worked hard to develop, and without the update propagation mechanisms that keep that infrastructure current. Most professionals in these fields already work this way. A spec library simply extends the scope — from the firm to the ecosystem — and adds the update propagation layer that local implementations cannot provide. Crucially, alerting becomes easier as scope increases: a firm-level clause library can notify its own lawyers; an ecosystem-level spec library can notify everyone who has deployed that clause, everywhere.

Five domains follow, each anchored in documented failures rather than hypotheticals. The pattern of failure is the same in each: verified components reconstructed probabilistically, with no mechanism to flag when the reconstruction is outdated or wrong.

Legal — from documented failures to the solution

The documented failures are already extensive. The well-known confabulation problem — LLMs inventing or hallucinating case citations — is now extensively documented: over 900 documented cases worldwide by early 2026 — approximately 90% from 2025, at a rate that grew from a handful per week before spring 2025 to several per day by year end.⁸ But confabulation is not the subtler or more dangerous failure. More concerning are the cases of substantive misstatement: LLMs misrepresenting the holdings of real cases, generating legal standards that are plausible but wrong.⁹ An invented case is at least detectable — a citation that leads nowhere. The misrepresentation of a real case is harder to catch, and harder to explain when it reaches a judge.

The testamentary capacity scenario from Section I is an instance of a third failure mode, distinct from both: correct law from the wrong time. The LLM generates language that was legally sound when most of its training data was written, but has since been superseded. No citation is invented. No holding is misrepresented. The clause simply reflects an earlier state of the law, reconstructed with full confidence from a training distribution weighted heavily toward older documents.

Most people believe that contracts are drafted from scratch by lawyers exercising unique judgment. In sophisticated practice, this is largely not the case.¹⁰ Contracts are assembled from clause libraries — standardized, tested fragments that have accumulated meaning through use, interpretation, and precedent. A lawyer deploying a verified clause is not just recycling boilerplate — they are deploying interpretive history that freshly drafted language cannot have.¹¹ The boilerplate is good precisely because it has been tested.

The problem is that these libraries exist at the firm level, not the ecosystem level. An LLM asked to draft a will does not consult any firm’s clause library. It reconstructs from training data — which includes decades of document language in varying states of currency, jurisdiction-appropriateness, and drafting quality — without any record of what it chose or why.

Standardization would change this — but even a partially standardized system would be an improvement in one fundamental way: it would provide names. Right now when an LLM generates a will, there is no way to say “this clause came from X” because there is no X — only a probabilistic reconstruction with no identity and no provenance. One of the oldest insights in philosophy of language applies here: you cannot discuss, critique, update, or be alerted about what you cannot name. A spec library gives components identities. Those identities enable everything else: the journal article analyzing v2.37, the alert when v4.2 is superseded, the record of which documents used which version. Naming is the first gift of standardization; verification is what you build on top of it.

Imagine a law journal article titled “Testamentary Capacity Clause v2.37: Consequences and Inadequacies.” Fifty pages of dense analysis. Proposed modifications for cognitive impairment cases, foreign nationals, blended families. This article is possible only because there is a v2.37 for everyone to analyze together. Standardization does not flatten the intellectual landscape. It maps it, which is the precondition for exploring it productively. A will drafted using verified language for testamentary capacity and execution requirements need not be a standard will. The verified components handle the routine questions correctly, freeing the drafter’s full attention for the genuinely circumstance-specific ones.

Medical — documented failure even in a professional system

The medical domain is the most striking example because the failure is documented in a professional, standardized-component system, not in consumer AI use. A 2025 peer-reviewed study examined Electronic Health Record (EHR) order sets for alcohol withdrawal syndrome across multiple hospital organizations.¹² Order sets are exactly the verified-component model applied to clinical care: standardized collections of treatments, medications, and protocols assembled by clinical experts and embedded in EHR systems for point-of-care use. They are professionally composed from approved components, not vibe-written.

The study found that a widely-shared templated order set — used across multiple hospital organizations — was still based on 2004 American Society of Addiction Medicine (ASAM) guidelines, despite substantially revised 2020 guidelines having been available for five years. The time lag was not caused by carelessness or incompetence. It was caused by the absence of an update propagation mechanism. No one was automatically notified that the template they were using was misaligned with current guidelines. The component was verified at the time of creation. It had not been re-verified since.

When a clinician asks an LLM to confirm a dosing protocol or summarize current guidelines, the LLM reconstructs a WAG from its training distribution rather than retrieving a verified current guideline.¹³ An outdated dosing protocol reconstructed with high confidence is more dangerous than an uncertain one, because it does not signal the need for verification. The EHR order set study shows that the update propagation problem exists even when practitioners are using standardized components — which makes it more urgent, not less. The alert mechanism is not just a feature for consumer AI use. It is a feature every professional system needs.

Education — verified components without verified currency

Teachers already compose lesson plans from verified components: standards-aligned learning objectives, pedagogical frameworks, pre-built activity blocks. Teachers Pay Teachers (TPT), the largest marketplace for K-12 educational resources, hosts over three million materials with more than a billion downloads — a mature ecosystem of shared, reusable teaching components.¹⁴

The structural gap is familiar. A 2022 study of over 500,000 TPT resources found that Common Core alignment, while claimed by many resources, was not a focal point of most content — meaning the verified-component layer is nominal rather than rigorous. When a state revises its curriculum standards, there is no systematic mechanism to notify teachers whose lesson plans reference the old version. Verified components exist. The update propagation layer does not.

Scientific Papers — where citation infrastructure already exists

Scientific papers already have a mature citation infrastructure. When a paper says “following the method of Smith et al. 2019,” it is gesturing at a verified fragment. The spec library extends that infrastructure to make method references machine-consumable, versioned, and checkable against the current version. The novel contribution of a paper is precisely the residual after all standard methods are tagged.¹⁵

Engineering Calculations — where regulatory liability follows the certifier

Structural engineering reports, thermal load studies, and environmental assessments all reference approved standards. Building codes already specify which calculation methods are permitted — a constraint spec system operating informally, without the infrastructure to connect it to generation workflows.¹⁶ An LLM generating a thermal load calculation that silently uses a method superseded by an updated energy code produces a document that looks authoritative but is professionally defective. The liability follows the professional who certified it, not the LLM that generated it.¹⁷

The pattern across all five domains — legal, medical, education, scientific, engineering — is identical. Practitioners already compose verified fragments rather than generating from scratch. The spec library is not proposing something new. It is proposing the infrastructure that connects this universal practice to LLM generation workflows, plus the update propagation layer that none of the existing tools provide.

These failures have costs: in legal sanctions, in vacated court orders, in compromised patient care. Standard language is part of the infrastructure that prevents them.

IV. Standard Language Is Not the Enemy of Good Language

A predictable objection: if all wills use the same verified testamentary capacity clause for common needs, if all NDAs use the same verified non-compete language for common circumstances, do we lose the diversity from which better practice emerges? The evolutionary analogy presents itself — monocultures are fragile, genetic diversity is the raw material of adaptation.

The objection is worth taking seriously. The response has three parts.

First: standardization redirects scrutiny rather than reducing it. When everyone uses the same verified clause for common needs, collective professional attention concentrates on whether that clause is correct and sufficient. The verified text becomes a focal point for improvement. When a court interprets it unexpectedly, the entire community learns simultaneously. The law journal article analyzing testamentary capacity clause v2.37 is possible only because there is a v2.37 to analyze. That is more evolutionary pressure on the standard language, not less.

Second: the spec library standardizes components, not artifacts. A will using verified language for testamentary capacity and execution requirements need not be a standard will — the routine parts are correct and the specific circumstances of the testator shape everything else. The drafter’s full attention is available for the genuinely circumstance-specific questions because the approved components have already answered the routine ones. Standardization amplifies the novel contribution by handling the routine, not by suppressing it.

Third, and most important: the alternative to standard language is not creatively diverse language. If created by an undisciplined LLM, it is random language. This is where the evolutionary analogy cuts the opposite direction from what the objection implies.

Evolutionary diversity is variation under selection pressure. Current LLM generation without spec constraints is genetic drift, not directed evolution.

The spec library introduces selection pressure: only verified, tested components enter the library. The result is directed evolution: the novel contributions are the variation, the library is the selected gene pool. Novelty detection — identifying which untagged patterns keep appearing independently across artifacts — is the mechanism that incorporates successful variations into the library over time.

Evolution is expensive selection. Most mutations are insignificant or damaging. A defective contract clause can destroy a business. A wrong dosing specification can kill a patient. The approved baseline is not a constraint on human judgment — it is the precondition that makes human judgment effective.

The monoculture fragility argument deserves a precise answer. The adaptive immune system offers a better analogy. Its power comes not from uniformity but from a mechanism for learning: the thymus teaches T-cells to distinguish self from non-self before they are deployed, and the system updates its trained responses when new pathogens appear. The spec library works the same way. The verified components are the trained baseline. The update and revocation mechanism is the adaptive response — when a new legal interpretation or a security finding appears, the spec is updated, and every registered user is notified.¹⁸ An agricultural monoculture has no such adaptive mechanism. A spec library ecosystem does.

V. Higher Quality at Lower Cost

Using spec libraries can lead to higher quality output at lower cost — directly addressing the urgent need to reduce LLM energy consumption and compute expenditure. That may seem like an odd claim. Let me explain.

When a developer asks an LLM to implement string search, the model does not retrieve a verified sorting algorithm implementation. It samples from a probability distribution shaped by every string search implementation in its training data — correct ones, subtly wrong ones, outdated ones, ones appropriate for different contexts — and produces a WAG that merely resembles a known artifact. The output looks like a string search implementation. Whether it is a correct one is a separate question.

The same thing happens when a lawyer asks an LLM to draft a non-compete clause, or a researcher asks it to describe a spectroscopic method, or a clinician asks it to confirm a dosing protocol. The model reconstructs from its training distribution, producing something that resembles the target well enough to pass a quick review. The subtle errors — the outdated assumption, the jurisdiction-specific exception omitted, the edge case not handled — are invisible precisely because the output looks so plausible.

This reconstruction is expensive in two ways simultaneously. First, compute: the model evaluates probability distributions over a vast space of possible outputs to produce something that merely resembles a known artifact. Second, quality: the result inherits whatever errors were present in the training distribution, with no mechanism for selecting the correct variant over the merely plausible one.

Probabilistic reconstruction of what is already known is the hidden waste in every unconstrained LLM generation. It is expensive and unreliable simultaneously. Use of a spec library eliminates it.

With a spec library, the known components are retrieved rather than reconstructed. The sorting algorithm implementation, the non-compete clause, the spectroscopic method, the dosing protocol — each arrives as a known-correct artifact with explicit provenance. The error modes introduced by unconstrained generation — silent omissions, outdated variants, jurisdiction-inappropriate choices — are eliminated for the retrieved portions. The residual error rate applies only to the novel contributions, which is where human review should be concentrated anyway.

This has a direct consequence for the cost of human oversight. Current practice requires reviewing the entire output of an unconstrained LLM, because any part of it might be wrong — the verified components reconstructed with errors, the novel inference applied incorrectly, and everything in between. Use of a spec library makes the boundary explicit. The verified portions arrive with provenance; a reviewer can confirm that the correct spec was retrieved without re-deriving it from first principles. The WAG portions are explicitly flagged as such, so the reviewer knows precisely where circumstance-specific judgment is required and where it is not. Human attention is then available in full for the genuinely novel parts — the circumstance-specific judgment that no library can supply and no reviewer should skip. This is a concentration of oversight where it creates value, not a reduction in it.

The compute cost falls for a related reason, though not primarily through token-level effects. The primary cost of LLM generation is repeated probabilistic reconstruction and the rework it induces — retries, corrections, extended review cycles — not token selection. In practice, LLM workflows follow a generate-review-find-issue-regenerate loop that is extremely expensive. Spec libraries change that structure: retrieve known components, generate only novel parts, review only novel parts. Cost reduction comes from eliminating reconstruction of what is already known, compressing prompts, and shifting computation from repeated online inference to amortized offline validation.

This economic structure has been demonstrated at scale in search.¹⁹ A predictive caching architecture showed that precomputing results against anticipated demand produces both lower latency and higher quality simultaneously. A cache of roughly 20 million precomputed results represented a substantial fraction of daily query volume and saved several hundred thousand processor cores. The spec library applies the same inversion to LLM generation: the library’s contents precomputed by domain experts, retrieved at generation time, novel computation concentrated on what actually requires it.

The result: higher quality at lower cost. The combination runs against the standard tradeoff in computing: higher quality costs more. Use of a spec library inverts it by replacing reconstruction with retrieval — a structural improvement, not a capability improvement.

Much of what looks like a model capability problem is actually an architecture problem. The model is not failing because it lacks capability. It is failing because it is being asked to do expensive probabilistic reconstruction of things that should be retrieved. Better models will not fully solve that problem. Better architecture will.

VI. The Simple Structure Underneath

The image below was produced by applying a set of eight rules first studied by Stephen Wolfram.²⁰ The starting condition: a single black cell, everything else white. The rules specify, for each cell, whether its next state should be black or white based on its own state and the state of its two immediate neighbors. That is all. Nothing more.

The complexity in the output is real. The pattern never repeats, and its statistical properties are indistinguishable from random noise. But the complexity lives in the output, not in the rule. The rule is trivial. Identifying the rule dissolves the apparent mystery without eliminating the complexity.

Rule 30 (Wolfram, 1983): eight simple rules, one black cell, 300 generations. Wolfram used this output as a pseudorandom number generator in Mathematica.

The failure modes surveyed in Section III look more complex than they are. Rule 30 suggests why: surface complexity can conceal a simple underlying mechanism. The reliability problem in LLM generation has the same structure. Across domains — software, legal drafting, scientific methods, medical protocols, engineering calculations, education — the failure modes look diverse and domain-specific. They reduce to a single mechanism: the boundary between the library’s contents and WAGs is invisible to the automated generator, and making it visible is the remedy. That reduction covers a large and important class of silent errors — a class that is both large and consequential, though not every LLM failure.

The KISS principle demands honesty about what is and is not simplified. What the spec library genuinely simplifies: the mechanism of the reliability problem. One structural observation covers the class of errors described above.

What remains genuinely complex: the governance problem — who verifies or certifies, who resolves disputes, how standards evolve — is not simple. The incentive problem is not simple either: unlike open source code, which maintainers need for their own work, spec libraries are closer to public goods and may require stronger institutional structures to sustain. The authoring cost — writing good specs is expensive and most organizations struggle to maintain documentation — is not simple. The semantic drift problem — natural language specs can change meaning without changing bytes, as when a court reinterprets “good faith” — is not simple. The version proliferation problem — spec ecosystems tend to accumulate versions that are never retired — is not simple. The fragmentation problem — legal standards vary by jurisdiction, medical guidelines by country, educational standards by state — means that federated governance structures, not a single universal authority, are likely the realistic path.

The more careful claim: use of a spec library simplifies the right thing — it operates at the correct level of abstraction at which the problem has a tractable structure. The remaining complexity is visible and addressable rather than hidden and compounding.

VII. The Infrastructure We Should Build

Bush ended “As We May Think” hoping that we could find a way to organize the record of human knowledge, making it accessible and usable rather than buried and lost. Eighty years later the problem is not accessibility. Every professional in every domain has more relevant literature than they can possibly read. The problem is that the record is not organized in a form that the machines now doing much of the reading and writing can use reliably.

In the pre-LLM world, knowledge systems helped humans find information. In the LLM world, knowledge systems must help machines compose artifacts safely. That shift in function is the reason the spec library is needed now in a way it was not before. The Memex was a navigation tool. The spec library is a generation constraint. Both address the same underlying problem: the organization of human knowledge, but at different moments in the relationship between knowledge and the machines that process it.

The spec library could be that organizational infrastructure. Not just for software. Not just for law. Not just for medicine. For every domain where knowledge is composed into consequential artifacts and where getting it wrong has consequences that matter.

An alert system is a natural extension of the same infrastructure. As specs are updated — a security finding, a Supreme Court interpretation, an updated clinical guideline — every registered user of the affected version could receive notification.²¹ An LLM with memory of the specs it has previously used could go further: scanning prior work to flag documents that drew on a now-superseded version, or identifying WAGs in existing documents that could be replaced by verified components added to the library since the document was drafted. This is only possible if the language is standardized and the usage is recorded. The SCOTUS scenario that opened this essay is not a fantasy. It is the direct consequence of taking the spec library seriously.

The bootstrap starts with what already exists: law firms maintain private clause libraries, medical systems have order sets, schools use standards-aligned lesson components. The spec library framework does not require these to be replaced — it requires them to be connected to a shared versioning and update infrastructure. The private library is the organizational layer. The public spec is the foundational layer above which private variation operates.

A further bootstrapping mechanism is available that existing tools like Henchman approximate but do not fully exploit. An LLM processing documents at inference time — reviewing contracts, analyzing briefs, summarizing clinical notes — sees patterns across millions of documents from different firms, jurisdictions, and practice areas simultaneously. It can identify which fragments appear repeatedly and consistently across contexts: the candidates for verified spec entries. The feedback loop is self-reinforcing: the LLM identifies candidates from what it processes, human experts verify them, verified entries enter the spec library, and the next round of novel contributions becomes the next round of candidates. Human experts do not start from a blank page. They start from LLM-identified patterns in real professional practice. The benefit is not confined to individual users. Each verification adds to a shared library of ever-increasing utility — a public good whose value grows with every contribution and every use.

The cryptographic layer that would make conformance claims machine-verifiable: W3C Verifiable Credentials, Decentralized Identifiers, signed provenance records in code and documents — exists and is ready to be applied.²² That application will be developed in a subsequent essay. The point here is simpler: the spec library’s value does not depend on cryptographic verification. It begins with versioned, human-readable, attributed specifications. The first verified entry for a testamentary capacity clause, a standard spectroscopic analysis method, or a GDPR-compliant data handling protocol could be written today by a domain expert with an afternoon and a text editor.

Bush’s trails are waiting to be built. The machines are waiting to follow them. The knowledge exists. The infrastructure does not. Yet.

Footnotes

1. “Vibe coding” was coined by Andrej Karpathy in February 2025 to describe directing AI systems to write code through natural language prompts without detailed specification of the desired behavior. “Vibe-writing” is not universally welcomed; see Sirsh. “Vibe Writing Shouldn’t Be a Thing.” Drafter, Substack, April 2025 (https://drafts4life.substack.com/p/vibe-writing-shouldnt-be-a-thing), which argues that the value of writing lies in transparency of process and maintained human agency — a concern fully compatible with the spec library’s provenance and WAG-tagging mechanisms.↩︎

2. Clause library software for legal professionals has matured into a substantial commercial category. Leading tools include ClauseBase (now acquired by LawVu; clausebase.com), which allows lawyers to build, version, and track clause libraries across their entire document portfolio; Henchman (now Lexis Create+), which “creates a database from a firm’s past contracts” and surfaces pre-approved clauses at drafting time — the private library layer the spec library describes; Draftwise, which uses historical firm deal data to guide drafting; and Spellbook, which benchmarks clauses against market-standard terms. Enterprise clients include CMS, Deloitte Legal, AES, and Axel Springer. All of these tools solve the within-firm consistency and retrieval problem well. None address cross-firm post-generation update propagation — the alert mechanism that is the spec library’s contribution above the existing infrastructure.↩︎

3. Ackoff, Russell L., Martin C. J. Elton, Thomas A. Cowan, James C. Emery, Peter Davis, Marybeth L. Meditz, and Wladimer M. Sachs. The SCATT Report: Designing a National Scientific and Technological Communication System. University of Pennsylvania Press, 1976. Chapter 1, pp. 6–7. Ackoff’s two constraints for idealized design are stated in italics in the original. The characterization of idealized design as “not a work of science fiction” appears on p. 7. SCATT stands for Scientific Communication and Technology Transfer. The project was funded by the National Science Foundation’s Office of Science Information Service.↩︎

4. WAG: “Wild Ass Guess” — an estimate of lesser worth than a SWAG (Scientific Wild-Ass Guess). The term originated in US military usage and is applied to any rough estimate made without rigorous calculation or verified grounding. Wikipedia: https://en.wikipedia.org/wiki/Scientific_wild-ass_guess↩︎

5. Wyman, Bob. “The Next Turn of the Spiral: Fixing Vibe Coding Without Reinventing Software Engineering.” Substack, March 2026. https://mystack.wyman.us/p/the-next-turn-of-the-spiral-fixing. That essay acknowledged prior art explicitly: Design by Contract (Meyer), TLA+ (Lamport), Alloy, SyGuS, OpenAPI, and W3C conformance testing. The spec library concept is not a replacement for these traditions but an architectural layer connecting them to the LLM generation workflow.↩︎

6. Bush, Vannevar. “As We May Think.” The Atlantic, July 1945. https://www.theatlantic.com/magazine/archive/1945/07/as-we-may-think/303881/↩︎

7. W3C Web Annotation Data Model. W3C Recommendation, 23 February 2017. https://www.w3.org/TR/annotation-model/. The W3C Annotations specification provides a standard for attaching structured commentary, provenance, and meaning to web resources — the closest existing infrastructure to Bush’s associative trails with epistemic status. Its limited mainstream adoption is itself evidence of the gap the spec library addresses.↩︎

8. Mata v. Avianca, Inc., 678 F. Supp. 3d 443 (S.D.N.Y. 2023), is the landmark case: lawyers submitted a brief containing six completely fabricated cases generated by ChatGPT. Since then the phenomenon has grown dramatically. Researcher Damien Charlotin (HEC Paris / Sciences Po) maintains a public database of legal decisions worldwide involving AI-generated hallucinated content in court filings. As of March 2026, the database documents over 900 such cases across 31 countries, with approximately 90% issued in 2025. Charlotin reported the rate escalated from approximately two cases per week before spring 2025 to two or three per day by September 2025, and five or six per day by December 2025. Database: damiencharlotin.com/hallucinations. A Stanford RegLab study found hallucination rates of 33% for Westlaw AI-Assisted Research and 17% for LexisNexis Lexis+ AI on legal queries. See Magesh et al., “Hallucination-Free? Assessing the Reliability of Leading AI Legal Research Tools,” Journal of Empirical Legal Studies (2025); arXiv:2405.20362.↩︎

9. Two documented failure modes beyond outright fabrication. (1) FABRICATED CITATIONS: Johnson v. Dunn, No. 2:21-CV-01701-AMM (N.D. Ala. July 23, 2025): ChatGPT-generated citations in briefs submitted by attorneys at a prominent national law firm; three attorneys sanctioned, publicly reprimanded, disqualified from the case, and referred to the Alabama State Bar. Shahid v. Esaam, Georgia Court of Appeals, June 30, 2025 (not Florida as sometimes misreported): a pro se litigant’s brief contained 15 citations of which 11 were hallucinated; one supported a claim for attorneys’ fees; the trial court issued an order partially relying on the hallucinated holding; on appeal the court vacated the order and remanded. (2) MISREPRESENTATION OF REAL CASES: Fivehouse v. U.S. Department of Defense, No. 2:25-cv-00041 (E.D.N.C. 2025–2026): the court’s own order identified fabricated quotes and misstatements of the holdings of multiple real circuit court opinions — cases that existed but whose holdings were described in ways the court found inaccurate; the assistant US attorney responsible was fired. The New York Appellate Division noted the same failure mode in Deutsche Bank Nat’l Trust Co. v. LeTennier, 2026 NY Slip Op 00040 (3d Dep’t Jan. 8, 2026), observing that even where AI provides accurate case citations it may misrepresent the holdings of those cases, often in a direction favorable to the party supplying the query. The currency/jurisdiction mismatch problem — correct law for the wrong time or place — is a structural variant of failure mode 2 and is illustrated by the testamentary capacity scenario in Section I; it is less documented in sanctions decisions because it is harder for courts to detect.↩︎

10. That legal drafting already operates largely through assembly of standardized fragments is underappreciated outside the profession. See Adams, Kenneth A. A Manual of Style for Contract Drafting. 4th ed. American Bar Association, 2017.↩︎

11. The ISDA Master Agreement and its definitional booklets represent the most mature domain-specific verified fragment library in commercial practice. A direct ancestor of the “verified clause” concept is Ian Grigg’s Ricardian Contract (1990s), in which a document is simultaneously a legal agreement and a machine-readable parameter set. See Grigg, Ian. “The Ricardian Contract.” First IEEE International Workshop on Electronic Contracting, 2004.↩︎

12. Electronic health record order sets represent verified-component composition in clinical practice. A 2025 study of hospital EHR order sets for alcohol withdrawal syndrome found that a widely-shared templated order set was still based on 2004 American Society of Addiction Medicine (ASAM) guidelines rather than the substantially revised 2020 guidelines — illustrating the time lag associated with implementing new guidance even in a professional system with standardized components. The study noted that “the wide reach of a single templated order set suggests an opportunity to broadly shift adoption of guidelines” — precisely the alert mechanism argument. See: Crotty et al. “Guideline concordance of electronic health record order sets for hospital-based treatment of alcohol withdrawal syndrome.” Journal of General Internal Medicine, 2025. PMC12276874.↩︎

13. FDA DailyMed provides structured drug labeling data; SNOMED CT provides a versioned clinical terminology standard; ICH guidelines govern pharmaceutical regulatory submissions. None are currently connected to LLM generation workflows in any systematic way.↩︎

14. A study analyzing over 500,000 resources on Teachers Pay Teachers found that Common Core standards, while present, were “not a focal point of most content” — meaning the marketplace is large but the verified-component layer is thin and inconsistently applied. See: Rosenberg, J.M. et al. “Analyzing 500,000 TeachersPayTeachers.com Lesson Descriptions Shows Focus on K-5 and Lack of Common Core Alignment.” Teaching and Teacher Education, 2022.↩︎

15. Scientific reproducibility problems arise from multiple sources: data availability, undocumented experimental conditions, statistical misuse, and code differences. Method tagging addresses the specification gap specifically and does not claim to solve the broader reproducibility problem. See Baker, Monya. “1,500 scientists lift the lid on reproducibility.” Nature 533, 452–454 (2016).↩︎

16. Engineering standards bodies whose specifications constitute proto-spec library entries include AISC (steel construction), Eurocode (European structural standards), ASCE, ASHRAE (mechanical/energy), and ISO. None are currently connected to LLM generation workflows in any systematic way.↩︎

17. Additional domains: insurance policy language (ISO clause forms), patent applications, actuarial reports (Actuarial Standards of Practice), pharmaceutical regulatory submissions (ICH Common Technical Document), international trade documentation (UCP 600, Incoterms), and academic curriculum accreditation materials.↩︎

18. The alert system requires no new infrastructure beyond what the spec library already provides. The CVE (Common Vulnerabilities and Exposures) system maintained by MITRE Corporation demonstrates the model: vulnerability discoveries come from distributed contributors, MITRE maintains the registry, the process is transparent, and no single authority is the source of trust. Who bears responsibility if a verified spec is wrong is a governance question addressed in a forthcoming essay.↩︎

19. Wyman, Robert M., Trevor Strohman, Paul Haahr, Laramie Leavitt, and John Sarapata. “Predictive Searching and Associated Cache Management.” US Patent Application US20100318538A1. Filed June 12, 2009. Assigned to Google LLC. The system precomputes search results against anticipated queries during idle time, achieving lower latency and higher result quality simultaneously by shifting computation away from the moment of user demand — the same economic structure the spec library proposes for LLM generation.↩︎

20. Rule 30 is one of 256 elementary cellular automaton rules identified and studied by Stephen Wolfram. See Wolfram, Stephen. A New Kind of Science. Wolfram Media, 2002. The image was generated independently using Python’s standard library; no Wolfram code or proprietary material was reproduced. Rule 110, similarly trivial, has been proven Turing complete. See Cook, Matthew. “Universality in Elementary Cellular Automata.” Complex Systems 15, no. 1 (2004): 1–40.↩︎

21. See note 17 above.↩︎

22. The W3C Verifiable Credentials specification and Decentralized Identifiers (DID) provide cryptographic infrastructure for making conformance claims machine-verifiable rather than merely human-readable. The application to spec library verification — including novel questions about LLMs as credential subjects and credential chaining for compositional specs — will be developed in a subsequent essay. Note that the original W3C working group was named the “Verifiable Claims Working Group,” a name that more accurately captures the epistemic structure: a spec library entry is a verified claim, not a credential in the identity-management sense.↩︎

Here is an apparent paradox: everything about LLM-based programming is genuinely new, and none of it is new at all. Both halves of that statement are true simultaneously, and understanding why is the point of this essay.

The enthusiasts are right. Natural language is genuinely a new kind of programming language. The accessibility is genuinely broader. The experience of building things has genuinely changed in ways that would have seemed remarkable even a decade ago. When someone describes what they want in plain English and working code appears, something real has happened.

The skeptics are also right. The underlying principles of specification, complexity, abstraction, and certification are unchanged and inescapable. Complexity doesn’t disappear because you express it in English. Systems built on underspecified foundations fail, regardless of what language was used to specify them. The old rules apply with exactly the same force they always have.

Both are right because they are describing different levels of the same phenomenon. The enthusiasts are observing the surface correctly. The skeptics are observing the structure correctly. The apparent contradiction dissolves once you recognize that we have been exactly here before — not metaphorically, but precisely. Every previous turn of the abstraction spiral in programming history presented the same paradox: a genuinely new language or mode of expression — machine code to assembler, procedural to object-oriented, formal syntax to natural language — that nonetheless had to obey all the old rules, because those rules admit no exceptions regardless of the language in which the work is expressed. And every time, the resolution was the same. The ecosystem eventually reconstructed, at the new level of abstraction, the same disciplines and institutions that had proven necessary at every prior level. Not because anyone planned it that way, but because the underlying principles left no other choice. This essay argues that we are at that moment again — that the apparent conflict between “everything changes” and “nothing changes” is not a paradox to be resolved but a pattern to be recognized, and that recognizing it tells us precisely what work lies ahead.

The Principle That Doesn’t Care What Language You Use

In 1956, the cyberneticist Ross Ashby formulated what became known as the Law of Requisite Variety: a system that regulates another system must have complexity at least equal to the system it regulates.¹ Applied to software specification, the implication is straightforward and unforgiving. A specification that is less complex than the program it describes cannot uniquely determine that program. The missing complexity has to come from somewhere.

In traditional programming, missing complexity surfaces as compiler errors, failed tests, or runtime crashes. The system fails loudly and points at the gap. In LLM-based programming, missing complexity is filled silently by the model’s statistical priors — by whatever the training data suggested was most likely in this context. The program doesn’t fail to compile. It compiles beautifully into something that may or may not be what you wanted, and you may not be able to tell the difference until much later.

This is the actual problem with vibe coding. Not that natural language is imprecise — English can express statements as precise as anything written in C++ or Python. The problem is that natural language, unlike formal programming languages, permits imprecision. It doesn’t enforce complexity. The failure mode is invisible rather than loud, and invisible failures in software are the dangerous kind.

When a vibe coder writes “encrypt the user’s password,” the LLM doesn’t leave a gap where the encryption specification should be. It makes choices — about algorithm, key management, padding scheme, error handling — based on statistical inference from training data. If the training data contains a high proportion of incorrect, insecure, or naïve implementations — and the internet is generously supplied with exactly those, in tutorials, Stack Overflow answers, and code samples written to illustrate something other than security — then the LLM will do its best to ensure that its output has a roughly similar probability of being wrong. The statistical process is not biased toward correctness; it is biased toward resemblance to what it has seen. This is not a flaw in the system. It is the system working exactly as designed. What is commonly called an LLM “hallucination” — a confident wrong answer — is often the architecture functioning correctly: producing output that faithfully resembles the distribution of its training data, errors included. The hallucination framing implies a departure from normal behavior. The more precise framing is that an LLM trained on the internet will replicate the internet, including its confident mistakes, its cargo-culted patterns, and its widespread misunderstandings — at scale, without fatigue, and with complete syntactic fluency. What is missing is any mechanism that introduces correctness as an explicit constraint on a process that has no such constraint by default. The result looks complete. It may even work, in the sense that it runs without errors. But those choices were never made by the developer. They were made by a probability distribution, without the developer’s knowledge or review.

The rest of this essay is about what that mechanism looks like and why we have built versions of it before.

We Have Been Here Before

This is not the first time a new programming language created this problem. It is, in fact, the same problem we have encountered at every previous turn of the abstraction spiral.

Grace Hopper understood it in the 1950s. When she worked to make COBOL as English-like as possible, she was trying to collapse the gap between specification and program entirely — to create what she called an executable specification.² The dream was that a business manager could write something close to plain English and have that be both the description of what was wanted and the artifact that executed it. COBOL was a partial success. The gap narrowed but never closed.

The first solution to the problem of specification complexity was the subroutine. A programmer who needed to search a string did not have to implement string search, and did not have to specify it either. A single terse call stood in for both the implementation and its specification simultaneously — the first complexity voucher. The subroutine was small at first: a few instructions for computing a square root, converting a number to a string, sorting a short list. But the principle it established was unlimited in scope. The complexity encapsulated inside a subroutine call could grow indefinitely while the call itself remained a single terse instruction.

Over decades the subroutine grew. Small utilities became library functions. Library functions became operating system services. Operating system services became entire subsystems. At each step the same principle held: the person calling the function did not need to understand it, only to trust it. The voucher remained valid regardless of how much complexity it encapsulated. By the time this trajectory reached its logical conclusion, the voucher could represent not a few instructions but millions of lines of carefully engineered code — an entire product, callable as though it were a simple function.

There was once a time when a database system, a text editor, a calendar, and a communications tool were entirely separate artifacts. Each stood alone. A programmer who wanted to combine their capabilities faced an impossible task: the products had no interfaces designed for composition, no way to be called rather than simply used. Integration was not merely difficult; it was architecturally excluded. This changed with the recognition that a product could expose its function while keeping its form separate — that the million-line complexity of an entire product could be made available through a defined interface that any other system could call.

In the 1980s at Digital Equipment Corporation, the same problem appeared in a different form. Products were being built as monolithic systems that bundled their interfaces tightly with their underlying functions. A business manager who needed to pull figures from a database, check a calendar, do some arithmetic, and write a memo had to run four separate programs, each with its own commands, its own conventions, its own mental model. There was no way to call the function of one product from within another. Integration was architecturally excluded.

The solution was callable interfaces — DTR$DTR for Datatrieve, TPU$TPU for the Text Processing Utility, and similar interfaces across Digital’s product line. The underlying function was separated from the form through which it was accessed. Anyone could now compose a new interface from existing functional components, calling certified, well-understood capabilities without reimplementing them. The complexity of the underlying function was encapsulated, accessible by reference, no longer requiring full understanding by everyone who used it.

The payoff was demonstrated concretely in Digital’s ALL-IN-1 integrated office system. ALL-IN-1 was able to provide a remarkable breadth of capability — document creation, database access, electronic mail, calendar management, and more — precisely because it was built as an integration of Digital’s other products as callable components into a consistent whole with a common user interface. The ALL-IN-1 engineers did not reimplement what Datatrieve or TPU already did. They called certified, well-understood components and concentrated their effort on the integration itself — the novel interface, the consistent user experience, the coherent composition of capabilities that none of the underlying components provided alone. The result was greater than the sum of its parts, and it required far less code than a ground-up implementation would have demanded. This is what the callable interface architecture made possible: integrated solutions whose complexity was managed through composition rather than reimplementation.

At Microsoft in the early 1990s, the same architectural vision found its fullest expression in the OLE and COM family of technologies. OLE — Object Linking and Embedding — provided the framework for compound documents and inter-application communication; COM Automation extended that foundation to make it accessible to solution builders. The goal was explicit: allow solution developers using Visual Basic or VBA to call components written in C++ without understanding what C++ was doing. A solution builder could interact with a database, an editor, a spreadsheet engine — not by understanding how these things worked, but by calling certified interfaces that guaranteed specific behaviors.

The Microsoft Solutions Development Framework that emerged from this work made a distinction that proved durable: solution developers, who understood user needs and worked in high-level languages, and component builders, who understood machine behavior and worked in lower-level languages. Each group had tools appropriate to their domain. You couldn’t write a device driver in Visual Basic, and writing a user interface in C++ was laborious where VB made it trivial. The distinction wasn’t about intelligence or status. It was about which complexity each role was responsible for managing.

The architectural decision that made this separation durable was the immutability of COM interfaces. COM relied on composition rather than inheritance, which meant that component builders were free to refactor, optimize, or completely reimagine their implementations — but the interface contract could not change without explicit versioning. This was not merely a technical constraint. It was a guarantee to solution builders that the complexity of the component world, however much it evolved, would never silently reach across the boundary and become their problem. The interface was a promise as much as a specification: changes below the line stayed below the line.

The Voucher That Cashes Out

These historical examples share a common structure that goes to the heart of the Ashby problem.

When a VBA programmer wrote database.Query(), they were writing a terse statement that nonetheless carried enormous specification complexity — not because the programmer understood that complexity, but because someone else had already fully specified and implemented it, and the interface was a certified reference to that work. Call it a complexity voucher: a short, accessible token that points to a fully elaborated specification.

The alternative — and the failure mode of vibe coding — is what might be called a fake voucher. When a vibe coder writes “query the database,” they are superficially doing something similar: using a terse phrase to gesture at a complex domain. But database.Query() is a pointer to a determinate specification. “Query the database” is a pointer to a distribution over possible specifications, weighted by the LLM’s training data. The compression is illusory. There is no agreed decompression key.

Experienced developers use LLMs more effectively than novices for exactly this reason. They know which terse phrases function as genuine complexity vouchers — because they know the underlying specifications those phrases reference — and which phrases are merely gestures toward a vague neighborhood of solutions. They know the difference between “use RSA encryption” and “use OpenSSL’s RSA_PKCS1_OAEP_PADDING with 2048-bit keys generated from a hardware entropy source.” Each additional constraint narrows the distribution of possible implementations the LLM samples from. Full specification collapses the distribution to something approaching determinism — but at that point, you’ve essentially written the code yourself and are using the LLM as a syntax formatter.

The leverage of LLM-based programming exists in the gap between specification and implementation. That gap is also where the probabilistic uncertainty lives. The challenge is to capture the leverage while managing the uncertainty — and the answer, as it was at every previous turn of the spiral, is certified abstraction.

A Specification Library

The solution is straightforward to state, though not trivial to build: a library of specifications for common computational functions, design patterns, and system-level invariants, maintained as versioned, human-readable, publicly accessible documents — and a development discipline that requires LLMs to consult that library before generating any code.

This is less novel than it sounds. It is precisely what the software development community already does with code — through open source repositories, package managers, standards bodies, and platform-bundled libraries. npm, PyPI, CTAN, GitHub: each is a versioned, distributed, reputation-governed ecosystem of reusable components. The difference is that the artifacts being shared, versioned, and governed would be specifications rather than implementations. The infrastructure, the trust mechanisms, the contribution patterns, and the governance models are all already proven at scale. What is new is applying them one level up.

The existence of these prior traditions is not a limitation the essay needs to manage. It is confirmation of the essay’s central argument. Design by Contract, developed by Bertrand Meyer in the Eiffel language, showed that software elements should carry specifications governing their interaction. TLA+ and related formal tools showed that system behavior can be specified precisely and machine-checked. Alloy showed that constraint exploration could surface design flaws before implementation. OpenAPI and schema-first API design showed that a human- and machine-readable specification can serve as the source of truth from which validation, documentation, and code artifacts are generated. W3C and OASIS conformance methodology showed that reference implementations and conformance test suites are achievable at scale across diverse implementors. These traditions exist. They work. They are underused precisely because they were developed for a previous turn of the spiral, at a level of formality and tooling investment that the mainstream of software development was unwilling to sustain. The LLM context changes the calculus. A specification that once required a programmer to learn TLA+ before it could do any work now requires only that it be readable — by a human and by an LLM. The prior art does not need to be reinvented. It needs to be recognized, connected, and applied to the generation loop that is already in use. We have been here before. We know what to do. The question is whether we will do it.

There is a more fundamental distinction worth drawing, one that also clarifies why this approach differs from its predecessors more than it might first appear. Spec-first development — write the specification, then implement against it — is a familiar methodology with a mixed track record. It leaves two artifacts that can drift apart: the spec describes intent, the code is reality, and synchronizing them is a perpetual maintenance burden. Literate Programming tried to collapse that gap by interleaving document and code in a single artifact, but the practitioner still authored both, and two things that could diverge still existed.

The vibe coding context is categorically different, and the spec library proposal is designed for that context. The practitioner does not write code. The natural language prompt is not a precursor to code — it is the artifact the practitioner authors and controls. The LLM output is not an implementation of a separately maintained specification; it is the direct execution of intent, constrained by the spec library. This is not spec-first. It is spec-as-code. The specification is not documentation of what the code should do; it is a constraint on what the execution produces. The analogy is not a comment preceding a function — it is a type annotation enforced by a compiler. The synchronization problem of spec-first development disappears because there is only one artifact: the specification and its execution are the same act.

The discipline would work roughly as follows. When given a programming task, the LLM first decomposes the problem into its functional components. It then queries the specification library to find whether certified specifications exist for any of those components. Where specs exist, the LLM’s code generation is constrained by those specs rather than by unconstrained statistical inference. Where no spec exists, the LLM flags the gap explicitly rather than filling it silently.

This changes the character of LLM output fundamentally. Instead of a single artifact — here is your code — the output becomes a structured report: components handled by certified specs, with version numbers and references; components where specs exist but organizational policy decisions are required; components where no spec was found and probabilistic generation was unavoidable; platform and context dependencies where the LLM’s knowledge is asymmetric; and ambiguities in the original specification that decomposition revealed.

The unresolved components list is particularly valuable. It converts invisible uncertainty into a visible engineering artifact — a prioritized research agenda that tells a development team exactly where to focus human attention. The LLM is no longer hiding its uncertainty inside apparently complete code. It is surfacing uncertainty where a human can act on it.

What Goes In the Library

The specification library would contain three broad categories of entries.

Behavioral specifications describe functions whose outputs are deterministic given their inputs. Cryptographic algorithms, network protocols, sort algorithms, encoding schemes. These are the computational equivalents of mathematical definitions — there is a ground truth, and a sufficiently detailed spec can converge on it. The gold standard here already exists: IETF RFCs are precisely this kind of artifact. RFC 8017 is the RSA specification.³ The framework doesn’t need to reinvent these — it needs infrastructure that connects them to the code generation process.

A well-written spec also transcends its original domain. The abstract problem that KMP solves — find a pattern within a longer sequence efficiently — is not specific to text. Bioinformatics researchers recognised this early: the Smith–Waterman algorithm (1981)⁴ and the tools built on it, including BLAST and FASTA, apply the same foundational insight to DNA and protein sequences, adding a domain-specific scoring layer to handle biological realities such as tolerated substitutions and insertion events. An LLM presented with a DNA sequence matching problem and a sufficiently abstract string search spec could reason by analogy, recognising the structural similarity and adapting the specification to the new domain, rather than either failing to make the connection or inventing an approach from scratch that silently violates known requirements. The generality of a spec is a design choice: a spec written at the right level of abstraction becomes applicable far beyond the problem that motivated it.

Constraint specifications describe components that must satisfy a set of conditions while retaining legitimate degrees of freedom. User interface components, accessibility requirements, layout systems, style guides. No single reference implementation exists because many valid implementations exist simultaneously. These specs define what is outside the valid space — no red for colorblind accessibility, minimum font size, right border alignment — while a curated suite of example implementations populates the interior of the valid space. The spec is the statute; the example suite is the case law.

System-level invariants are cross-cutting constraints that apply regardless of which components are used or how they are assembled. Security requirements, privacy constraints, regulatory compliance obligations. A private key must never be stored in plaintext. Personal data must never be logged without consent. These are constraints that currently live in the heads of experienced developers, in post-mortem reports after failures, or in compliance documents that developers may or may not have read. Encoding them as active constraints on code generation converts known requirements from passive reference material into enforced properties of every generated system.

The organizational policy layer sits above all three categories. A company specifies which spec versions are approved for its projects, which system-level invariants apply to all generated code, and what confidence thresholds require human review. Solution builders are generally relieved of these decisions until a breaking change requires organizational action — exactly as COM Automation shielded VBA programmers from C++ implementation changes until an interface version required updating.

Reference Implementations and Conformance Tests

Every specification that can be implemented should be accompanied by a reference implementation and a conformance test suite. This is not a novel idea — the best standards bodies have always done this, and the practice reveals why it matters.

A reference implementation is an executable proof that the specification is internally consistent and complete enough to produce a working system. Any specification that cannot generate a reference implementation has failed its own standard. It contains degrees of freedom the spec author didn’t resolve — places where an LLM generating code would be sampling from a distribution rather than converging on a determinate result. The failure is diagnostic: it tells the spec author exactly what work remains.

A conformance test suite converts the qualitative question “does this implementation conform to the spec” into a quantitative and automatable one.⁵ An LLM-generated implementation can be compared against the reference implementation’s behavior across the conformance test suite, producing a conformance score rather than a binary judgment. Organizational policy can specify minimum conformance thresholds for different risk levels — 98% might be acceptable for a logging component, 100% required for authentication.

Together the spec, the reference implementation, and the conformance test suite constitute a complexity voucher of the highest order — simultaneously human-readable, machine-executable, and LLM-interpretable. The spec is what the solution builder references. The reference implementation is what the component builder produces. The conformance tests are what the CI/CD pipeline checks. Each serves a different consumer without contradiction.

There is an important quality gate implicit in this structure, applicable to behavioral specifications: if a behavioral spec cannot be used to generate a reference implementation and a conformance test suite, the specification is broken. The failure is not in the tooling — it is in the spec itself, which contains unresolved ambiguity or internal contradiction. This gate requires no external authority to apply. A spec that fails to bootstrap its own verification has failed by the very standard the framework is built on. Constraint specifications and system-level invariants require a different quality gate: their constraints must be testable, non-contradictory, and precise enough that a generated system can be checked against them. A constraint spec that any implementation trivially satisfies, or that no implementation could satisfy, has failed its own standard — differently but just as completely.

Design Patterns and the Shape of Interaction

Functional component specs handle the what-each-piece-does problem. But a complete system also requires specifications for how pieces relate to each other — and this is a categorically different kind of specification complexity.

The design patterns literature, beginning with the Gang of Four’s 1994 catalog, addressed exactly this.⁶ Observer, Factory, Command, Mediator, MVC — these are not functions but templates for component interaction. A vibe coder might successfully invoke certified encryption and certified database queries while assembling them in a pattern that has subtle timing or state assumptions violated, producing failures that are extremely difficult to diagnose precisely because each component in isolation appears correct.

The specification library should include pattern specifications alongside functional specs. These would specify constraints on interaction rather than deterministic behaviors — more like constraint specs than behavioral specs. A pattern spec for Observer would define what properties must hold between subject and observer regardless of how either is implemented, what failure modes arise from common violations, and what the observable behavioral differences are between correct and incorrect implementations.

Joel Spolsky, writing in 2002, formulated what he called the Law of Leaky Abstractions: all non-trivial abstractions leak.⁷ The complexity that an abstraction was designed to hide will surface at the worst possible moment, and when it does, the person using the abstraction must understand what is beneath it to handle the failure. This law doesn’t invalidate abstraction — it sets an honest ceiling on what abstraction can achieve and identifies a requirement for spec authorship. Every spec should document its known leakage points: where is this abstraction thin, what underlying complexity is most likely to surface, what should a solution builder know to diagnose failures at this level. The spec library doubles as a leakage diagnostic resource — not preventing leaks but ensuring that when they occur the knowledge needed to understand them is findable, legible, and attributed to people who can be consulted. Spolsky’s law also sets an honest bound on the solution builder / component builder distinction. Abstraction can reduce how often low-level knowledge is required — but it cannot reduce that requirement to zero. The component builder’s role becomes rarer and more specialised at each turn of the spiral; it never disappears. What the spec library provides is not the elimination of that need but the assurance that when the abstraction leaks, the knowledge required to handle it is findable rather than locked in someone’s head.

An Ecosystem, Not a System

The specification library is not a single system maintained by a single authority. It is an ecosystem — diverse, distributed, variably authoritative, and governed by reputation rather than by fiat.

This is how the software library ecosystem works, and for good reason. Some specifications will be formally certified through years of work in IETF, W3C, IEEE, or ISO. Others will be maintained by open source communities with strong reputations and transparent processes. Others will be organizational or product-specific, maintained privately. Others will be developed by individuals for specific purposes. The trust hierarchy is emergent — determined by track record, adoption, transparency of process, and response to discovered errors — rather than assigned by a central authority.

The interesting governance question is not who certifies but how you decide which certifications to trust. And we already know how to solve that problem because we solve it everywhere else. An IETF RFC carries more weight than a university research group’s spec, which carries more weight than an individual’s GitHub repository — not because a central authority enforces this hierarchy but because of accumulated reputation and transparent process. The same mechanisms that make you trust a software package with ten million downloads over one with twelve operate on specifications. The trust signal is emergent and self-correcting.

Applied to the spec library, this means that a contribution to an existing spec — the insight that forced a revision — is as citable and as permanent as the original authorship. Your two lines might appear not in the initial spec but in the update that corrected it. That attribution belongs to you no less than the original author’s belongs to them.

The attribution system also creates a reputation economy for specification knowledge that does not currently exist anywhere in software development. A developer whose track record shows accurate, widely-adopted updates to specs across multiple domains — cryptographic protocols, accessibility constraints, network error handling — has demonstrated something rare and currently invisible: the ability to reason precisely about behavioral requirements across different problem spaces. There is no credential for this today, no reputation system that captures it, no way for an organisation to find such a person or for such a person to demonstrate their value.

Attribution also expands who can earn deserved reputation. The current software ecosystem is deeply winner-take-all: the people who wrote the Linux kernel, created Python, or built React occupy a permanent reputation aristocracy that later contributors cannot challenge regardless of the quality of their subsequent work. The foundational components are taken; the opportunity to build them passed before most current developers were born. The skills required to conceive a foundational component are not the same skills required to find the subtle flaw in a widely-deployed specification that has been in production for a decade. Both are valuable. The current ecosystem rewards only the first. The spec library’s attribution model rewards both — and creates a path to deserved reputation for the analyst, the edge-case finder, and the careful reader who arrives after the foundations are already laid.

The more open a specification is, the more eyes see it, and the more quickly an error in one implementation triggers a spec update that benefits everyone. This is not a complete answer to the failure mode problem — a widely adopted incorrect spec can propagate errors at scale before the error is detected — but it is the same mitigation that open source software relies on, and it has proven reasonably effective there.

New Affordances

Every previous turn of the abstraction spiral reconstructed the same disciplines at a new level. But this turn has properties that previous turns lacked, and they produce genuinely new capabilities.

Specifications are human-readable in a way that code is not. A VBA programmer who encountered a bug in a COM component could not read the C++ source and understand it. A solution builder who encounters unexpected behavior in an LLM-generated system can read the spec that governed the generation and understand what was intended, what constraints applied, and where the leakage occurred. The abstraction is transparent to the people who use it in a way that compiled code never was.

Knowledge can be active rather than passive. Documentation has always been inert — it informed humans who then wrote code. Specifications in this framework are active inputs to the generation process. The distance between knowing a constraint and enforcing it collapses. A security requirement encoded in a system-level invariant spec is enforced in every generated system that touches its domain, automatically, without requiring any developer to remember it.

Tacit knowledge can be externalized at scale. The most dangerous knowledge in software development is the kind that lives only in experienced developers’ heads — the constraints they apply automatically without thinking about them. The framework creates institutional pressure to make that knowledge explicit and encoded, because an unapplied constraint is a visible gap in the spec rather than an invisible absence in someone’s mental model.

Intellectual contribution becomes attributable and durable. Code obscures authorship almost immediately — it gets refactored, dependencies change, languages evolve, systems are rewritten. A specification, being human-readable and implementation-independent, has a much longer half-life. The insight outlasts any particular artifact that embodies it. The people who contribute foundational insights to widely used specs will have their contributions embedded in the infrastructure of software development for generations — legible, citable, and attributed in a form that even non-engineers can find and read.

Specifications can represent legal and regulatory obligations as first-class constraints. Laws and regulations are already a form of specification — they define required behaviors, prohibited actions, and conditions of compliance. GDPR is a privacy constraint spec. HIPAA is a system-level invariant for healthcare data. PCI-DSS is a security behavioral specification for payment systems. Currently these exist as documents that developers may or may not have read, encoding requirements that LLMs have no systematic way to apply. Representing them as spec library entries — with the same versioning, conformance testing, and organizational policy machinery as any other spec — converts regulatory compliance from a manual audit process into an enforced property of generated systems. The longer-range possibility, which the human-readability of specs makes conceivable for the first time, is that legislatures might eventually adopt specification formats for original drafting: law written to be simultaneously human-readable and machine-applicable, without the translation layer that currently separates legal text from enforcement. That remains speculative, but the affordance that makes it possible — a rigorous, human-readable format that machines can also consume — is precisely what the spec library provides.

The feedback loop between practice and specification closes. Currently when a production system fails, the lesson might reach a blog post or a conference talk. Under this framework, the lesson has a natural home — an update to the relevant spec, reviewed and certified, automatically applied to all future generation using that spec. The distance between learned lesson and enforced constraint collapses dramatically.

The Spiral

In 1988, Barry Boehm published his spiral model of software development.⁸ But there is another spiral in the history of programming itself — one that turns more slowly and whose implications are still unfolding.

Each turn of this spiral introduced a new language that allowed problems to be expressed more naturally and at higher abstraction than what came before. Machine code to assembler. Assembler to high-level languages. High-level languages to object orientation. And now to natural language. At each turn, the same pattern repeated: initial euphoria, the same problems resurfacing at the new level of abstraction, and eventually the reconstruction of the same disciplines — libraries, tools, methodologies, role distinctions, institutional structures — at the new level.

And at each turn, the previous level didn’t disappear. It became the component layer for the new level, managed by specialists when needed. Assembly language didn’t vanish when C appeared. C didn’t vanish when Python appeared. Python will not vanish because natural language prompts can generate it. It will become what natural language compiles to, managed by component builders whose specialized skills become more critical precisely as they become less common.

The natural language turn is the one that was always the destination. Every previous turn was pointing here — toward the point where the specification and the program are the same artifact, where human intent and machine execution are separated by the minimum possible distance. We have not quite arrived. The gap is narrower than it has ever been, and it is still there.

What fills that gap — what has always filled it — is certified abstraction. Complexity vouchers that allow the person working at the new level to reference the fully elaborated work of the people working at the level below, without having to understand it, without having to reimplement it, and without having to trust that the LLM made good choices in the dark.

It is worth pausing to note that both Ashby and Boehm, whose work underlies this essay’s central argument, have faded from active reference in recent years in ways that are difficult to justify given the continued force of their ideas. Ashby’s Law of Requisite Variety is routinely rediscovered independently in adjacent fields — under different names, without citation, without the accumulated debt being acknowledged. Boehm’s spiral model was absorbed into software engineering curricula as a historical artifact rather than a living analytical tool, then largely displaced by Agile methodologies that inverted his intentions. Boehm was explicit that iterative development was meaningful only when driven by rigorous risk analysis at each cycle — the hard thinking had to happen, iteration was how you managed its consequences, not how you avoided it. In common practice, some Agile implementation devolved into precisely what Boehm warned against: iteration used as a substitute for understanding rather than a discipline for managing it. The result is what it must be under Ashby’s law — a random walk through implementation space that gives the appearance of forward motion while systematically avoiding the specification complexity the system actually requires. That the essay arguing for the return of specification discipline must also argue for the return of the thinkers who made that discipline rigorous is itself an instance of the problem it describes. The knowledge exists. It simply stopped being cited.

The First Step

It is worth being honest about where this framework will not help much. Exploratory work — prototyping a product idea, discovering requirements through iteration, building glue code between systems, refactoring existing code, writing tests for known behavior — does not require a spec library and would gain little from one. The framework is most valuable where the problem domain is already stable enough to admit specification: security primitives, network protocols, data serialization, accessibility constraints, regulatory compliance. In domains where the user does not yet know what they want, or where the value comes precisely from rapid informal iteration, certified abstraction is not the bottleneck. Informal iteration is not always evasion of thought; often it is how requirements are discovered. The spec library addresses one class of LLM coding failure — silent underdetermination in well-understood domains — not all of them.

None of this requires waiting for institutions to act, standards bodies to convene, or platform vendors to build new tooling.

There is in fact an existing infrastructure that demonstrates the problem and the solution simultaneously. The Model Context Protocol — MCP — allows LLMs to discover and invoke tools at runtime. An LLM’s ability to match a tool to a need is entirely dependent on the quality of the tool’s description. A poorly described tool is invisible to the LLM even when it is exactly what the task requires. A well-described tool is reliably found and correctly applied. This is the fake voucher problem in miniature and in real time: a terse description that does not cash out to a determinate specification of what the tool does, under what conditions, and with what constraints, produces the same silent probabilistic failure as an underspecified natural language prompt. MCP tool descriptions are proto-specs — they exist, they are structured, they have a recognised quality problem, and the spec library discipline applies to them directly. Improving the description of a tool you already use, applying the spec format to make its behavior, constraints, and failure modes explicit, and observing that the LLM uses it more reliably as a result: that is perhaps the most immediately actionable first step available, requiring no new infrastructure and producing measurable results within a single session.

The minimal viable version of this framework exists today. A markdown file containing a careful specification for a common function. A GitHub repository making it publicly accessible. A prompt that instructs an LLM to decompose a problem, check whether components match attached specifications, use those specs to constrain generation, and report what it couldn’t match.

That is step one. It is embarrassingly simple, which is either a sign that something important is being missed or a sign that the idea is correct. The history of consequential open source projects suggests that embarrassingly simple first steps are exactly how durable ecosystems begin. The first demonstration is simple; the ecosystem is not. Building a living library of trusted, versioned, composable specifications that practitioners actually use is a collective action problem requiring sustained authoring effort, governance, tooling, and incentive structures that do not yet exist. That work is harder than writing a markdown file and pushing it to GitHub. The point is that it starts there.

The first public spec could be written today. The first library could be created this week. The first demonstration — showing that a spec-constrained LLM prompt produces more deterministic, auditable, and trustworthy output than an unconstrained one — could be running before the end of the month.

And then, as it has always been, more would follow.

Notes and Open Questions

Notes on threads from the conversation that generated this essay, for possible future development:

[1] String search as a candidate first spec library entry: Knuth, Morris, and Pratt’s linear-time string matching algorithm provides a clean example of a commonly needed function with a precisely specifiable behavior, a reference implementation, and a well-defined conformance criterion.⁹ Word boundary detection (the “what is a word” problem) is a natural companion spec that would depend on it.

[2] The formal verification community and its relationship to this framework. The framework is complementary rather than competitive — more specs and less code to formally verify is likely welcome — but the relationship deserves explicit treatment.

[3] The spec format question: should there be a standard structure for spec library entries? Metadata (title, author, date, version, review status), behavioral description, constraint layer, conformance tests, reference implementation, leakage documentation, known limitations. Something like IETF RFC format adapted for this purpose.

[4] Spec identity and naming requires a three-tier architecture, analogous to Zuko’s triangle extended with a fourth edge for persistence. A cryptographic hash identifies one specific version of one specific document — this exact text, these exact bytes — verifiable without trusting any authority. A GUID or IPFS address identifies a spec or family of related specs sharing the same interface and observable effects, persisting across editorial improvements until a breaking interface change requires a new identifier. A tag URI (RFC 4151) provides persistent, human-readable, authority-derived identity for the entire lineage across all versions and interface revisions — the name a human cites, an organisation governs by, and a grandchild finds decades later. The three tiers map to a directory structure: the tag URI is the root directory, GUID subdirectories represent interface versions, hashed files within each subdirectory are specific editorial versions. One tag URI contains a history of GUIDs; each GUID contains a history of hashes. A complete reference carries all three: tag URI for human citation and governance, GUID for compatibility checking, hash for verification.

Colophon

This essay was developed through an extended dialogue between the author and Claude Sonnet 4.6 (Anthropic), conducted in March 2026. The author provided the conceptual framework, the historical account, the architectural insights drawn from direct experience, and the intellectual direction throughout. Claude contributed elaboration, structural organisation, connection-making across sources, and occasional extension of ideas. The text was reviewed, revised, and approved by the author.

Model: Claude Sonnet 4.6 (claude-sonnet-4-6). Interface: claude.ai. Session date: March 2026. The dialogue that generated this essay is available as part of the conversation history in the author’s Claude account.

A draft was subsequently reviewed by Google Gemini and OpenAI ChatGPT, whose critical responses sharpened several arguments. The author remains responsible for all claims.

There is a pleasant irony in the colophon itself. This essay argues that intellectual contributions should be attributed explicitly, that the provenance of ideas matters, and that knowledge which is made visible and citable outlasts knowledge that is not. The colophon is an instance of that argument — a record of how this particular artifact came to exist, made available to whoever cares to read it.

Footnotes

1 W. Ross Ashby, An Introduction to Cybernetics, Chapman & Hall, London, 1956. The Law of Requisite Variety is developed in Part Three, which Ashby identifies as the “central theme” of cybernetics. The law is stated formally in Chapter 11. A digitized edition is freely available via the Internet Archive: archive.org/details/introductiontocy00ashb

2 Grace Hopper’s work on COBOL and her concept of the executable specification are documented in: Jean E. Sammet, “The Early History of COBOL,” ACM SIGPLAN Notices, Vol. 13, No. 8, August 1978, pp. 121–161. Hopper’s own account of her motivations is given in numerous interviews; see also: Kathleen Broome Williams, Grace Hopper: Admiral of the Cyber Sea, Naval Institute Press, 2004.

3 The RSA cryptographic standard is specified in: “PKCS #1: RSA Cryptography Specifications Version 2.2,” RFC 8017, Internet Engineering Task Force, November 2016. Available at: tools.ietf.org/html/rfc8017. RFC 8017 obsoletes RFC 3447 (2003) and RFC 2437 (1998), illustrating precisely the versioned, immutable specification model described in this essay.

4 Erich Gamma, Richard Helm, Ralph Johnson, and John Vlissides, Design Patterns: Elements of Reusable Object-Oriented Software, Addison-Wesley, 1994. The four authors became known collectively as the “Gang of Four,” and the book’s catalog of 23 patterns remains the foundational reference for object-oriented design patterns.

5 Joel Spolsky, “The Law of Leaky Abstractions,” Joel on Software, November 11, 2002. Available at: joelonsoftware.com/2002/11/11/the-law-of-leaky-abstractions/. Spolsky states the law as: “All non-trivial abstractions, to some degree, are leaky.” The essay remains one of the most widely cited in software engineering practice.

6 Barry W. Boehm, “A Spiral Model of Software Development and Enhancement,” IEEE Computer, Vol. 21, No. 5, May 1988, pp. 61–72. DOI: 10.1109/2.59. The spiral described in this essay is distinct from Boehm’s: his spiral addresses iterative risk management within a single development project; ours describes the historical succession of abstraction levels across the entire history of programming. The resonance of the metaphor is intentional. It is worth noting that Boehm himself warned explicitly against what he called “hazardous spiral look-alikes” — processes that adopted the iterative form of the spiral while discarding the risk analysis that gave it meaning. Common Agile practice, in which iterative delivery substitutes for upfront specification rather than managing the consequences of careful analysis, falls squarely within Boehm’s category of dangerous impostors. The random walk through implementation space that results is not a failure of iteration as a technique; it is a failure to apply the specification discipline that makes iteration converge rather than merely explore.

7 Donald E. Knuth, James H. Morris, Jr., and Vaughan R. Pratt, “Fast Pattern Matching in Strings,” SIAM Journal on Computing, Vol. 6, No. 2, June 1977, pp. 323–350. The KMP algorithm achieves O(n+m) string matching versus the O(nm) naive approach — a difference that becomes significant at scale and that a vibe-coded implementation would silently get wrong. The algorithm is also discussed in Knuth’s The Art of Computer Programming, Volume 2. Knuth’s planned Volume 5 (Syntactical Algorithms) will cover lexical scanning and word boundary problems more generally; it remains in preparation as of 2026.

8 The ‘tag’ URI scheme is specified in: Tim Kindberg and Sean Sayer, “The ‘tag’ URI Scheme,” RFC 4151, Internet Engineering Task Force, October 2005. Available at: tools.ietf.org/html/rfc4151. A tag URI mints a globally unique, persistent identifier from a domain name or email address controlled by the minting authority at a specific date, plus a locally assigned string. Unlike URLs, tag URIs make no claim about the location of the identified resource, providing persistence independent of hosting infrastructure.

9 Temple F. Smith and Michael S. Waterman, “Identification of Common Molecular Subsequences,” Journal of Molecular Biology, Vol. 147, No. 1, 1981, pp. 195–197. The Smith–Waterman algorithm performs local sequence alignment across nucleic acid and protein sequences using dynamic programming, achieving guaranteed optimal local alignment. It was subsequently extended into the BLAST family of tools (Altschul et al., “Basic Local Alignment Search Tool,” Journal of Molecular Biology, Vol. 215, 1990, pp. 403–410), which trades exhaustive optimality for the speed required to search large sequence databases. The lineage from KMP to Smith–Waterman to BLAST is a demonstration in the biological sciences of exactly the domain-transcendence property the essay attributes to well-abstracted specifications.

Practical men, who believe themselves to be quite exempt from any intellectual influence, are usually the slaves of some defunct economist. Madmen in authority, who hear voices in the air, are distilling their frenzy from some academic scribbler of a few years back.

— John Maynard Keynes, The General Theory of Employment, Interest and Money (1936)

There is a question that lies beneath every argument in this series of essays — beneath the claim that taxation is a public service, beneath the convergence of the benefit principle and the ability-to-pay principle, beneath the evidence that some countries deliver on capitalism’s promise while others do not. The question is this: what was the government legitimately doing in the first place?

Without an answer, the downstream arguments float free of their foundation. If the libertarian is right that government’s proper role ends at the watchman functions — enforcing property rights, protecting physical security, adjudicating disputes — then the “benefit” against which tax and civic obligation is measured is far smaller than this series assumes, and the restorative justice argument has no anchor. The libertarian objection, stated at its strongest, is not that government is incompetent or corrupt, though it may be both. It is that government is overbuilt — that it does too much, that everything beyond the watchman minimum is coercive overreach dressed up as public service.

But the libertarian is not the only objector. The revolutionary — in a tradition that long predates Marx, running through him to the various socialisms of the twentieth century — arrives at an equally sweeping rejection from the opposite direction. The institutional order is not overbuilt, on this view; it is irredeemably captured. The welfare state, the regulatory apparatus, the rule of law itself — these do not represent genuine reforms but the permanent co-optation of state power by capital. Tinkering within the system cannot work. The only honest response is to sweep the existing order away and replace it with something built on different foundations.

These sound like opposite positions. One wants government cut back to a principled minimum; the other wants the entire institutional structure replaced. But they share a deeper common structure, and that shared structure is what this essay addresses. Both treat the state capacity available at a particular historical moment as a permanent verdict on what institutional development can accomplish. The libertarian looks at the seventeenth century and before — when the watchman state was genuinely the ceiling of what people knew how to build as collective institutions — and concludes that the watchman state is the principled limit for all time. The revolutionary looks at the nineteenth century — when the administrative tools that might have made capitalism reformable barely existed — and concludes that capitalism is permanently unreformable within any institutional framework. Both positions were approximately defensible given the evidence available when they were formed. Both have been overtaken by a century of institutional learning for which neither can adequately account. And both, as a consequence, demand that learning stop at precisely the point where it becomes inconvenient for their preferred conclusions.

This essay argues that both objections fail — for the same reason. The mandate of government is not fixed by the institutional frontier of any particular century. It is indexed to ever-changing state capacity: to what we have collectively learned to do, and to what, having learned, we therefore may be obligated to do. That is Lincoln’s criterion for the proper scope of government, and it is where this essay begins.

I. Lincoln’s Criterion for the Proper Scope of Government

Abraham Lincoln left behind an unfinished fragment, written around 1854, that contains one of the most useful sentences in the literature on government’s proper role. Lincoln wrote:

The legitimate object of government, is to do for a community of people, whatever they need to have done, but can not do, at all, or can not, so well do, for themselves — in their separate, and individual capacities. In all that the people can individually do as well for themselves, government ought not to interfere.

— Abraham Lincoln, “Fragment on Government” (c. July 1, 1854)1

Each phrase of that criterion carries weight, and the unpacking rewards care.

“Need to have done” — not merely useful or convenient, but required for the functioning of a decent society. Lincoln is not saying that government should do whatever people would prefer not to do for themselves. He is setting a threshold: something that the community genuinely needs.

“Cannot do” — structural impossibility or inadequacy. This is not about preference or inconvenience. It is about whether private action, however well-intentioned and well-resourced, can actually accomplish what is needed.

“So well do” — and here is the crucial qualification that makes the criterion dynamic rather than static. Government is warranted not only when private action is impossible but when it is inadequate. The test is comparative: can the community do this better together than separately? If yes, and if the need is genuine, the mandate applies.

There is a fourth element, implicit but essential: a cost-benefit constraint. The mandate applies when collective benefit is reasonably expected to exceed collective cost — where cost is understood broadly to include not only financial expenditure but lost freedoms, constraints on private action, administrative burden, and unintended consequences of any kind. But this constraint must be applied symmetrically. Opportunity costs are real in both directions. A government that could prevent structural harm and does not is not neutral; it is choosing the costs of inaction as surely as an interventionist government chooses the costs of action. A genuine cost-benefit accounting counts the costs of not having public health surveillance, deposit insurance, or financial system stability alongside the costs of providing them. The asymmetric version — which counts the costs of action but not the costs of inaction — is not economics. It is ideology dressed as analysis.

The Fiduciary Obligation to Act

Lincoln’s criterion, so understood, is both descriptive and normative. It asks what government can do, and from the answer begins to derive what government should do. But scope alone does not settle the question of obligation.

If government’s legitimate object is to serve the community’s needs — as Lincoln’s criterion presupposes — then a government that has developed the capacity to meet a genuine need, at reasonable cost, and declines to do so is failing its own stated purpose.

Establishing that something falls within government’s legitimate scope — that it satisfies Lincoln’s criterion — is necessary but not sufficient to create an obligation to act. Scope defines the domain within which government may legitimately operate; an obligation to act arises only when three additional conditions coincide: a genuine collective need, sufficient state capacity to address it effectively, and a reasonable expectation that collective benefit will exceed collective cost.

Where all three conditions are clearly met, government is obligated to act. Where they are partially met — the need is real but capacity is still developing, or the cost-benefit calculation remains uncertain — there is a reasonable expectation of action as conditions mature, but not yet a full obligation. Where none are met, neither obligation nor expectation arises.

This three-condition test prevents the expansion of legitimate scope from becoming an unlimited warrant for government action. A government that has developed the capacity to do something it has no genuine need to do, or whose intervention would cost more than it delivers, is not thereby obligated to act. Lincoln’s criterion opens the door; the three conditions determine whether obligation requires walking through it.

This positions Lincoln between two inadequate alternatives. The libertarian freezes the criterion at a historical moment, treating the contingent limits on state capacity in the seventeenth century as if they were a permanent principle derived from the nature of government itself. The unrestricted progressive untethers the mandate from capacity and need entirely, asserting that government should do whatever justice requires regardless of whether the three conditions are met. Lincoln threads between them: the domain of legitimate government action expands as state capacity grows, but the obligation to act within that domain is disciplined by genuine need and honest cost-benefit accounting.

II. The Watchman State as the Limit of State Capacity

The minimal state that libertarians defend was codified at a particular moment in history. At that moment, it was also roughly the maximum of what people had yet learned to do collectively through government. The coincidence is worth examining.

Before going further, a word about the term “watchman state” itself — because its origin is instructive. The phrase was coined not by a libertarian but by Ferdinand Lassalle, the nineteenth century German labor organizer and socialist, in a speech delivered in 1862. Lassalle used it as a term of contempt. The bourgeois conception of the state, he argued, was nothing more than a Nachtwächter — a night-watchman — whose sole function was to prevent theft and burglary. This, he said, was “a policeman’s idea,” and he drew out its logical implication with characteristic sharpness: if the state exists only to protect property from thieves, there would be no reason for a state at all if thieves did not exist. Libertarians and minarchists subsequently reclaimed the term as a badge of honor, apparently without noticing — or without minding — that they were adopting a label coined specifically to mock their position. Keynes observed that practical men who believe themselves exempt from intellectual influence are usually the slaves of some defunct economist. The watchman state is a case in point: a concept distilled from an 1862 socialist pamphlet, whose original critical force may have been forgotten by those who now wave it as a flag.

The irony deepens when we notice the date. Lassalle coined the term in 1862. Marx published the first volume of Capital in 1867. The watchman and the revolutionary are not opposites from different eras. They are contemporaries, arguing within the same nineteenth century intellectual moment, about the same set of industrial-era problems, against the same backdrop of underdeveloped institutions. This essay’s title names both prescriptions for government in order to decline both — and the fact that both belong to the same historical moment is part of the reason why.

Consider what people in early modern societies could accomplish collectively through government. They could enforce peace within their territory — imperfectly, but substantially. They could define and protect property rights, adjudicate disputes through courts, field armies for external defense. They could, in short, do the things that Hobbes and Locke described as the essential function of government: maintaining the sovereign order that made civil society possible. These are genuinely important functions, and nothing in this essay diminishes them. The Hobbesian baseline — the minimum institutional order without which economic life above subsistence cannot exist — is real, and communities that fail to provide it produce misery in abundance.

But what people in early modern societies could not yet do collectively is equally important to the argument. They could not regulate food safety — bacteriology did not exist, and the germ theory of disease was not established until the late nineteenth century. They could not provide unemployment insurance — the administrative machinery to identify eligible workers, collect contributions, process claims, and distribute payments at population scale did not exist and had not been imagined. They could not manage a central bank or conduct macroeconomic stabilization — the theoretical framework that would make such management possible was not developed until the twentieth century. They could not pool health risk at population scale — actuarial science was in its infancy and the administrative capacity to implement social insurance programs had not been built.

The minimal state, in other words, was not a principled choice. It was the ceiling of what people had yet learned to build as collective institutions. Hobbes and Locke were describing the institutional frontier of their time, not deriving a permanent limit from the nature of political authority. Hobbes’ Leviathan is a remarkable work of political philosophy; it is not a permanent constitution for government’s scope.

One further note on Hobbes: his “state of nature” — the war of all against all from which the sovereign rescues us — is best understood as a theoretical construct about what social life would be like without institutional order, not as a historical description of what pre-political societies were actually like. Anthropological and archaeological evidence suggests that human communities before the state were considerably more cooperative and socially organized than Hobbes’s picture implies. The analytical baseline — the point that institutional order creates possibilities that disorder forecloses — survives this critique. The historical picture does not. And noting this matters, because the libertarian argument for the minimal state often depends on presenting the Hobbesian baseline as the natural default to which we return when government is reduced, rather than as a theoretical abstraction that has never precisely existed.

Robert Nozick’s minimal state, the most rigorous philosophical reconstruction of the libertarian position, has similarly never existed at the scale of a modern economy. It is a theoretical construction — admirable in its internal consistency, essential to engage seriously, but not a historical baseline to which we might return. The question is not whether to go back to something that once was. The question is how to move forward from where we are.

The libertarian error is specific: it treats a historical limit on state capacity as a normative boundary. The fact that government couldn’t do X in 1651 or 1862 does not mean it shouldn’t do X in 2026. That inference requires an additional premise — that the state capacity ceiling of the seventeenth century reflects a permanent principle rather than a contingent limitation — and that premise is simply false. State capacity grew, and continues to grow. The normative boundary, if the libertarian argument were correct, should have grown with it. The libertarian who concedes that government may legitimately do all that the watchman state could do has already conceded the principle; the argument about scope is then an argument about current state capacity and current need, which is exactly the argument Lincoln’s criterion invites.

III. The Right’s Own Thinkers Against the Watchman Position

There is a temptation, when making the case for an expanded government mandate, to draw primarily on thinkers associated with the political left — Keynes, Rawls, the welfare state theorists. That temptation should be resisted, for a simple reason: the case is stronger than that, and the intellectual honesty of engaging the other side’s best thinkers is itself part of the argument. The expansion of government’s legitimate mandate is not a left-wing imposition. The right’s own best thinkers, read carefully and completely, support it.

Hayek and evolutionary institutionalism

Friedrich Hayek is the most frequently invoked intellectual authority for the minimal state position, and the invocation is almost always based on The Road to Serfdom — the 1944 work that established his popular reputation as the great opponent of big government. The Road to Serfdom is a serious book, more nuanced than its reputation, and it deserves to be read rather than merely cited. Hayek’s argument there is specifically against central planning — the comprehensive direction of economic activity by a central authority. It is not a general argument against government intervention, and Hayek was explicit about this: he endorsed a social safety net, public health measures, and regulation of working conditions in the text itself, in passages his libertarian admirers tend to skip.

But the more important point is what Hayek argued in his mature work — The Constitution of Liberty (1960) and Law, Legislation and Liberty (1973–79) — which most conservatives seem not to have read because it is considerably less convenient than The Road to Serfdom. Hayek’s central argument in those works is that social institutions embody accumulated knowledge that no individual mind could reconstruct — knowledge dispersed across millions of participants, encoded in rules and practices that have survived because they work, not because anyone designed them to. Institutions evolve through a selection process: what works is retained, what fails is discarded.

Read consistently, this argument supports the welfare state. The welfare state is precisely what survived a century of evolutionary selection across virtually every developed democracy. It was not imposed by central planners; it emerged through democratic competition, institutional experimentation, and the retention of what worked. The libertarian proposal to replace the evolved state with a theoretically derived minimal state is precisely the rationalist constructivism Hayek warned against: replacing institutions shaped by historical experience with a theoretical model of how society ought to work. Hayek’s own argument, applied honestly, counsels respect for what has survived, not for what a theoretical model says should have survived.

There is also a deeper point about Hayek’s epistemic argument against central planning. Hayek objected that central planning requires knowledge that the state doesn’t have — the dispersed, local, tacit knowledge that only markets can aggregate. This is a powerful argument, developed most fully in his 1945 essay “The Use of Knowledge in Society.” But it is a conditional argument about current state capacity, not a permanent structural barrier. The same logic that says “the state shouldn’t do X because it lacks the knowledge to do it well” implies “the state should do X once it develops the knowledge to do it well.” This is Lincoln’s criterion restated in Hayekian terms. Hayek identified a real problem — the knowledge problem — without recognizing that the problem has a solution: institutional learning and development. States can, over time, find new ways to obtain relevant knowledge and develop the capacity to act in domains where they were once genuinely incompetent. To exclude all information-requiring government action because the state once lacked the information is an unnecessarily permanent solution to a contingent problem.

Burke and the conservative case for evolved institutions

In 1790, Edmund Burke made the conservative case for respecting institutions that have survived. His argument against the French Revolution was not that the ancien régime was good — it was that destroying evolved institutions in favor of abstract principles derived from pure reason was Jacobinism,² and Jacobinism ends badly. Whatever emerges from revolutionary rupture is likely to be worse than what was replaced, because what was replaced had at least been tested by time.

This argument cuts sharply against contemporary libertarian institutionalism. The welfare state, the regulatory apparatus, the central bank, social insurance — these have survived. They emerged through a century of democratic deliberation, institutional experimentation, and political competition. By Burke’s own standard, proposals to dismantle them in favor of a theoretically derived minimal state are the radical position. The Burkean conservative should be the defender of the institutional status quo, not its opponent. True conservatives should be devoted to the preservation of those institutions that have proven themselves successful. Today’s small-government institutional demolitionists, whatever they may call themselves, are, by Burke’s standard, the Jacobins.2

Smith and the invisible hand, correctly read

Adam Smith is the libertarian’s most frequently invoked authority, and the invocation is based almost entirely on the concept of the invisible hand — the idea that individuals pursuing their own interests are, as if by an invisible hand, led to promote the public interest. This has become the foundational metaphor for the claim that markets, left alone, produce optimal outcomes and that government intervention can only make things worse.

The metaphor appears exactly three times in Smith’s published work. In The Wealth of Nations, it is a narrow observation about one specific mechanism: the preference of domestic over foreign investment. It is not a general theorem about market optimality. It is not Smith’s central argument. It is a passing observation in a much larger work that spends Book V detailing the extensive and legitimate functions of government — roads, education, defense, the administration of justice, and more.

More importantly, Smith was deeply suspicious of the merchants and manufacturers whose interests the libertarian reading of him is most inclined to protect. His warning about collusion is direct and famous: “People of the same trade seldom meet together, even for merriment and diversion, but the conversation ends in a conspiracy against the publick, or in some contrivance to raise prices.” The invisible hand works when competitive conditions are maintained. Smith understood that those conditions do not maintain themselves — that concentrated private interests will, if unchecked, systematically undermine the competitive markets that make the invisible hand mechanism operate. The libertarians who cite Smith against regulation are making exactly the error Smith warned against: assuming the mechanism works regardless of the institutional conditions that are necessary for it to function.

Friedman and the negative income tax

Milton Friedman is among the most prominent conservative economists of the twentieth century and a vigorous opponent of many forms of government intervention. But in Capitalism and Freedom (1962), Friedman proposed something remarkable: a negative income tax — a universal basic income delivered through the tax system, providing every household with a guaranteed minimum income regardless of circumstances. This was at the time a genuinely novel institutional form, something government had never done before in quite that way. Friedman’s proposal concedes, at minimum, that the state has a legitimate insurance function that goes beyond pure market failure correction in the traditional sense, and that new institutional forms for discharging that function are worth developing. A consistent advocate of the watchman state cannot endorse the negative income tax. Friedman endorsed it anyway.

Nozick’s legitimacy condition

Robert Nozick’s Anarchy, State, and Utopia is the most rigorous philosophical defense of the minimal state. Nozick argued, correctly, that redistribution to achieve preferred distributional patterns is morally problematic — it treats individuals as means rather than ends, using some people’s holdings to serve others’ visions of the good society. This is a serious argument that deserves serious engagement, and this series engages it seriously in other essays.

But Nozick’s argument depends entirely on an assumption that is frequently overlooked: that existing holdings are legitimately acquired. His entire framework rests on this. If holdings were not legitimately acquired — if they rest, as so many concentrated fortunes do, on rent extraction, monopsony power, state-backed coercion, or the appropriation of what belongs to the community — then Nozick’s own principles condemn them. Correcting illegitimate holdings is an application of Nozick’s entitlement theory, not a violation of it. The libertarian who invokes Nozick against redistribution without examining the legitimacy of existing holdings is using Nozick selectively — accepting his conclusions while ignoring his premises.

IV. A Taxonomy of Government Functions

The preceding section argued against the libertarian case for limiting government to the watchman minimum. But arguing against a limit is not the same as describing what a properly understood government should be and do. That requires a positive account — a taxonomy of the functions that Lincoln’s criterion, applied honestly and systematically, actually warrants. That is what this section provides.

Before working through the categories, a clarification is needed about the economic concept most often used to justify government action.

A public good, in the technical sense that economists use the term, has two properties. It is non-excludable: once it is provided, you cannot prevent people from consuming it, whether or not they contributed to its provision. And it is non-rival: one person’s consumption does not diminish what is available to others. National defense is the canonical example: it protects everyone within the territory regardless of who paid for it, and protecting one citizen does not reduce the protection available to others. Clean air, lighthouses, the rule of law, and public knowledge are similar.

These properties together create a market failure: since non-payers cannot be excluded, no private actor can capture enough of the benefit to justify the cost of provision. The provision of public goods is therefore systematically neglected by markets, and government provision is warranted on straightforward economic grounds.

But here is what matters for the argument that follows: the coincidence that the most historically familiar and intuitive government functions — watchman and infrastructure — happen to involve public goods in the technical sense has contributed to a widespread but mistaken impression that the provision of public goods exhausts the economic case for government action. It does not. The provision of public goods is the first and most visible category of justified government action, but not the last. The others follow from different but equally compelling market failures — adverse selection, catastrophic tail risk, uninternalized externalities, natural monopoly, and coordination problems — each with its own logic, and each satisfying Lincoln’s criterion through a different mechanism. The state of knowledge in 1776, when Adam Smith wrote, or in 1854, when Samuelson’s formalization of public goods theory still lay a century in the future, did not yet provide the analytical tools to see this. We have those tools now.

The four categories are: foundational functions, risk and stability functions, investment functions, and constitutive functions. Each is examined in turn.

Foundational Functions

The watchman functions — property rights enforcement, contract law, physical security, the legal infrastructure within which economic life takes place — are public goods in the strict sense. Everyone benefits from a common legal order, no one can exclude themselves from it, and no private actor has the incentive or scale to provide it at the level a modern economy requires. These functions are the Hobbesian minimum. Without them, economic life above subsistence is impossible, and everything else in this taxonomy rests on them.

But even this most minimal category carries a lesson the libertarian position tends to miss. The watchman state is not cheap, not simple, and not self-sustaining. Property rights must be defined, constantly re-adjudicated as circumstances change, and defended against attempts to capture or corrupt them. Contract law must evolve to cover new forms of transaction. Legal infrastructure requires professional training, institutional maintenance, and ongoing investment. The “minimal state” is more demanding than its advocates typically acknowledge — and that recognition is the thin edge of the wedge against the watchman-state position as a stable resting point.

Infrastructure extends the foundational category. Physical networks — roads, railroads, electrical grids, the Internet backbone — involve natural monopoly: network economics favor consolidation to the point where competition either cannot survive or produces unnecessary duplication. Institutional standards — weights and measures, time zones, technical protocols — require coordination: someone must establish the standard, and once established, everyone benefits from using it regardless of who did the establishing. The monetary system and the payments system require scale and trust that no private actor can unilaterally provide. These are things individuals genuinely cannot do as well for themselves in their separate, individual capacities. Lincoln’s criterion is plainly satisfied.

Risk and Stability Functions

Social insurance — health coverage, unemployment benefits, disability protection, deposit insurance, pension security — is not a public good in the technical sense. Benefits are excludable (you can prevent specific people from receiving them) and rival (a dollar paid to one beneficiary is unavailable to another). The market failure that justifies government provision here is of a different kind, and it is important to be precise about what it is.

The first is adverse selection. Voluntary insurance markets face a structural problem: those most likely to need the insurance are most likely to buy it, while those least likely to need it have less reason to participate. As the risk pool skews toward high-risk participants, premiums rise; as premiums rise, more low-risk participants exit; as they exit, the pool skews further. Private health insurance without compulsory participation tends toward this death spiral — it is not a problem that can be solved by better products or more competition, because it is a structural consequence of voluntary participation in the presence of information asymmetry. Compulsory participation solves the problem by preventing adverse selection from operating. This is something we have learned — and it would have seemed draconian, or simply unnecessary, before the mechanism of adverse selection was understood.

The second is catastrophic tail risk. Some losses are simply too large relative to individual resources for private insurance to pool adequately across the full population. A serious cancer diagnosis, a prolonged disability, the loss of a job at the wrong moment in an economic cycle — these can exhaust a lifetime of savings. Adequate pooling requires a population large enough and diverse enough that the law of large numbers operates effectively, and that typically requires either compulsion or very substantial government involvement.

The third is macroeconomic externality. Unemployment insurance does something that private insurers cannot provide as a product: it maintains consumer spending during recessions, dampening the downward multiplier effect that would otherwise amplify economic contractions. This macroeconomic stabilization benefit accrues to everyone operating in the economy, but no private insurer can capture it. It is, in effect, a public good produced as a side effect of what looks like a private insurance arrangement.

Macroeconomic stabilization itself — the management of aggregate demand to control inflation, the automatic stabilizers, the maintenance of financial system stability — is a public good in the strict sense. Stable prices and full employment, once achieved, benefit everyone in the economy and cannot be withheld from those who did not contribute to achieving them. No private actor has the scale to manage aggregate demand; the attempt by central banks and fiscal authorities to do so is precisely the kind of collective action that Lincoln’s criterion captures. Individually rational financial behavior is, under some conditions, collectively destabilizing — the bank run is the canonical case, but the financial panics of the nineteenth century and the 2008 financial crisis are the large-scale versions. Government provision of financial stability corrects a market failure that private actors cannot solve for themselves.

Investment Functions

Markets systematically underinvest in knowledge production. Knowledge, once created and published, is non-rival — my knowing the structure of DNA costs you nothing — and largely non-excludable — once the paper is published, the idea is out. Private firms can capture some of the returns to knowledge through patents, trade secrets, and first-mover advantages, but they cannot capture the full social return. The gap between private and social returns is the measure of underinvestment, and it is large. The entire modern technology economy rests substantially on publicly funded basic research: the Internet emerged from DARPA, the World Wide Web from CERN, the pharmaceutical revolution from NIH-funded basic science, the semiconductor industry from government research programs and procurement. None of these would have been privately funded to the degree necessary, because no private investor could have captured enough of the return to justify the investment.

Public education is the human capital case. The returns to education accrue partly to the individual — in higher lifetime earnings — partly to employers — in a more productive workforce they did not pay to train — and partly to society in ways that markets cannot capture at all: civic participation, scientific progress, the cultural and political life of a community. Each private employer free-rides on the educated workforce that public investment in education created. Left to private provision alone, education would be systematically underprovided, concentrated among those who can afford it, and oriented toward immediately marketable skills rather than the broader formation that a democratic society requires.

Long-horizon investment presents a different problem. Private markets discount the future at rates that reflect investors’ time preferences and the opportunity cost of capital. Those rates are, from the perspective of society as a whole and of future generations who cannot bid in today’s markets, too high. Environmental protection, infrastructure maintenance, and scientific infrastructure are chronically underinvested by markets because their returns accrue over time horizons that private investors systematically underweight. Future generations have genuine interests in decisions made today — interests that the market cannot represent, because future people cannot buy and sell in current markets.

A contemporary example makes the state capacity threshold argument vivid. We have only recently learned that the thymus is the organ responsible for maintaining the immune system’s capacity to adapt to novel threats — training the T-cells that are the immune system’s learning arm. After early adulthood, the thymus undergoes a process of involution: it gradually loses function, and with it the immune system progressively loses its capacity to respond to pathogens it has not previously encountered.3 The consequences accumulate with age: reduced response to novel infections, reduced response to vaccines, increased susceptibility to cancers and other conditions that a fully functional immune system would suppress.

We have recently crossed a threshold in understanding this process well enough to ask whether the functional decline of the thymus can be interrupted or reversed. The social returns from answering that question affirmatively would be enormous — longer, healthier lives, substantially reduced late-life healthcare costs, maintained cognitive and immune function in aging populations. But no pharmaceutical company can capture enough of that value to justify the basic research investment. The returns are too diffuse, too long-horizon, and too difficult to appropriate. This is precisely the kind of investment that Lincoln’s criterion requires government to make: something people need done that they cannot do as well for themselves, at a cost that is reasonably expected to be exceeded by the collective benefit. The capability threshold has been crossed in real time, making newly obligatory what was recently impossible.

Constitutive Functions

The first three categories all rest, ultimately, on market failure arguments of one kind or another. The constitutive functions are different. They rest not on the failure of markets to provide something but on the right of a self-governing community to decide what kind of community it wants to be.

Public libraries, national parks, arts funding, historic preservation, public broadcasting — these are not justified primarily by market failure, though market failure arguments can sometimes be constructed for them. They are justified by the democratic community’s decision to express certain values collectively, to maintain certain shared spaces and cultural goods, to preserve certain things for their own sake rather than for their market value. Lincoln’s criterion in the narrow sense — the community cannot do this as well for itself — may not always be strictly satisfied here. Democratic self-determination provides the justification instead.

Redistribution, properly understood, belongs here — but as a much more constrained category than progressive politics typically assumes. Most of what is framed as a case for redistribution is better understood as restoration: returning to households and communities the surplus that rent extraction, monopsony power, and state-backed coercion transferred upward. Restoration is not redistribution. It corrects illegitimate transfers rather than overriding legitimate holdings, and it has a natural stopping condition — when the extraction has been corrected — that redistribution lacks. The moral logic is entirely different: restoration says that wealth was taken from you; redistribution says that someone else’s need is greater than your entitlement. Both may lead to a transfer of resources, but they are grounded in fundamentally different claims.

Restoration is not unlike what Nozick called rectification — the correction of unjust acquisition and transfer, which his entitlement theory not only permits but requires. The moral logic is similar: correcting what was wrongly taken is justice, not redistribution. Whether Nozick’s own principle, as he developed it, reaches the full range of extractions that restoration addresses is a question for another occasion.

What remains after restoration has done its work is the genuinely residual redistributive category: inequality that persists because it reflects differences in circumstances of birth that individuals did not choose — the family into which one is born, the community in which one grows up, the natural talents with which one arrives. A democratic community may legitimately decide, through its normal political processes, that children should not bear the full consequences of their parents’ failures, or that human dignity requires a material floor regardless of market outcomes. These are genuine moral claims. But they are claims about the residual inequality that remains after the more powerful claim — that most of what looks like inequality is extraction that should never have occurred — has been addressed. Treating redistribution as the primary tool for addressing inequality, before restoration has done its work, obscures the more fundamental problem and generates a weaker and more contested response to it.

Finally, fiscal citizenship. Edwin Seligman, the great American tax theorist of the early twentieth century, argued that membership in a political community generates obligations that are not reducible to either market failure correction or benefits received. Citizens and government are partners in a joint project — the maintenance and development of the institutional order that makes civil and economic life possible. Taxation, on this view, is not a price for services rendered or a cost imposed by state power. It is a form of civic participation, an expression of the partnership rather than a burden extracted from it. This dimension of taxation — the dimension that pure technical analysis cannot capture — is developed more fully in other essays in this series.

V. The Learning Dynamic

The taxonomy in the previous section is not a list of ideological preferences. It is what centuries of institutional learning have built.

Each category of government function became possible — and therefore, under Lincoln’s criterion, necessary — at a specific historical moment when state capacity crossed a threshold. Before the germ theory of disease, public health regulation was not possible; there was nothing coherent to regulate toward. Before actuarial science and modern public administration, social insurance was not possible; the machinery to implement it did not exist. Before the development of macroeconomic theory in the twentieth century, stabilization policy was not possible; there was no framework within which it could be conducted. Before the modern research university and the institutional infrastructure of basic science, large-scale public investment in knowledge production was not possible in the organized way we now take for granted.

The watchman state was not abandoned on ideological grounds. As state capacity grew, it was supplemented. The watchman functions remained and remain foundational. Each new category was added as a new capacity threshold was crossed, and each addition satisfied Lincoln’s criterion at the moment of its addition: something people needed done that they could not do as well for themselves, at a cost that collective benefit exceeded.

Two levels of learning must coincide for a new capacity threshold to become operational. The first is technical: germ theory, actuarial mathematics, macroeconomic theory, network engineering, epidemiology. The second is institutional: the civil service that can administer programs without corruption, the regulatory agencies that can apply technical knowledge to practical problems, the central bank that can conduct monetary policy, the public health infrastructure that can implement what the science recommends. Technical knowledge without the institutional capacity to deploy it is insufficient — a point the early Progressive Era repeatedly demonstrated. The ideas often existed before the machinery to implement them did, and the gap between idea and implementation was measured in decades.

The expansion of government’s mandate over the past two centuries has the structure of evolution. Variation comes through institutional experimentation — different countries and different periods trying different approaches to common problems. Selection occurs through democratic accountability and empirical performance — what works gets retained, what fails gets reformed or discarded. The welfare states that survived the twentieth century are not the ones that were theoretically optimal; they are the ones that worked well enough to retain popular support and to adapt when they didn’t. This is what Adam Smith described for markets, applied to the collective institutions that capitalism requires. It is not planning. It is adaptation.

Hayek’s insight that evolved institutions embody accumulated knowledge applies here with full force. The welfare state embodies a century of accumulated knowledge about how to address problems that markets cannot solve — problems of adverse selection, catastrophic risk, macroeconomic instability, knowledge underinvestment, and constitutive self-determination. It emerged through exactly the evolutionary selection process Hayek described. The libertarian proposal to dismantle it in favor of a theoretically derived minimal state is precisely the rationalist constructivism Hayek warned against — substituting a theoretical model for the accumulated wisdom of what has actually survived.

Two Groups Frozen at Different Levels of State Capacity

This history of institutional learning also helps explain why certain influential theories of the state continue to misread the trajectory of modern governance. Both the libertarian and the orthodox Marxist make the same structural mistake, just at different centuries.

The libertarian is frozen at the seventeenth century state capacity ceiling. The watchman state was the operational limit of Hobbes’s time; the libertarian treats it as the principled limit of all time. This is a mistake of category: a contingent constraint on state capacity is mistaken for permanent principle.

The orthodox Marxist is frozen at the nineteenth century ceiling. Marx’s conclusion that capitalism was unreformable — that the gap between its promise and its reality was too fundamental to be closed through institutional reform — was, in its time, an empirically grounded assessment of what the available state capacity could accomplish. Marx’s despair was well-founded. The administrative state barely existed. Civil service was in its infancy. There was no central banking worthy of the name, no securities regulation, no antitrust law, no progressive income taxation, no unemployment insurance, no labor law protecting collective bargaining. The franchise was restricted to property owners in most places. Given those constraints, the conclusion that the working class could not achieve meaningful reform through political processes was not unreasonable. It was approximately correct given the state capacity Marx could observe. In Marx’s time the administrative machinery required to reform capitalism simply did not exist. The century that followed would build precisely that machinery, often in response to the very problems Marx had identified.

The New Deal, the welfare state, central banking, antitrust enforcement, progressive taxation, labor law, the extension of democratic participation to the full adult population — these are precisely the institutional responses that Marx’s framework said were impossible within capitalism. They were built not everywhere, not completely, and not without capture and backsliding, but substantially across the developed world. Their existence is an empirical challenge to the impossibility claim, even if they remain incomplete and contested.

The orthodox Marxist response — that these institutions merely legitimate capitalism without fundamentally altering it, that they represent co-optation rather than genuine reform — is structurally identical to the libertarian’s dismissal of evidence that government intervention can work: in both cases, the theory is insulated from empirical falsification rather than updated in light of evidence. When evidence that should challenge a theory is explained away as the theory’s prediction, the theory has become a doctrine.

There is a further irony in the Marxist case. Marx’s own method — historical materialism, the insistence that analysis must be grounded in the material conditions of its time — argues directly against treating his nineteenth century conclusions as permanent structural truths. A genuinely Marxist analysis of twenty-first century capitalism, conducted with Marx’s method rather than frozen in his conclusions, would have to grapple seriously with the vastly more capable administrative state, democratic accountability, and the empirical record of welfare state capitalism. The Marxists most faithful to Marx’s method are probably not the ones most faithful to his conclusions.

The Immune System as Model

The immune system is an adaptive learning architecture. It does not arrive pre-programmed with responses to every pathogen it will encounter. It has a learning structure — a capacity to encounter novel threats, generate responses, retain what worked, and build those responses into a catalog of acquired immunity. It is, in Hayek’s terms, an evolved institution for managing biological risk.

At the center of this learning structure is the thymus. The thymus is where T-cells mature and are trained — where the immune system learns to distinguish self from non-self, friend from threat. Through this process, the immune system continuously updates its capacity to respond to novel pathogens.

But the thymus undergoes involution after early adulthood. It gradually loses function, and with it the immune system progressively loses its capacity to learn. An aging immune system can fight the threats it learned to recognize when young; it responds poorly to threats it has not previously encountered. The decline in adaptive immunity with age is not a design feature; it is a failure mode. An immune system that can only fight yesterday’s pathogens becomes progressively less capable of meeting the challenges that a changing biological environment generates.

The parallel to government is direct. The watchman state is the government that developed the state capacity to respond to the threats and collective action problems of the eighteenth century and before. It learned well. But the institutional thymus — the capacity for government to keep learning, to develop new responses to new problems as they emerge — is precisely what the libertarian position would suppress. A government constitutionally prohibited from building new state capacity is not a minimal state. It is a government with thymic involution: capable of addressing the threats it learned to recognize at the watchman stage, progressively unable to respond to the novel challenges that a more complex economy and society continuously generates.

The revolutionary position produces a different failure mode: rather than involution, rupture. Destroy the existing immune system and build a new one from scratch. The problem is that immune systems carry crucial memory — accumulated knowledge about threats encountered, responses refined, failures corrected. Destroying them destroys that knowledge. The new system must begin learning from zero, vulnerable to everything until it has rebuilt what was lost. Revolutionary rupture with existing institutions faces exactly this problem: the accumulated institutional knowledge embedded in what is destroyed does not automatically transfer to what replaces it.

The Demand to Stop Learning

Both the libertarian and the revolutionary, in their different ways, are asking that institutional learning stop.

The libertarian asks that we freeze government at the watchman stage. The revolutionary asks that we treat capitalism as permanently unreformable and stop trying to improve it. Neither is claiming that learning is impossible. Both are claiming that learning should stop at the point where it becomes inconvenient to their preferred conclusions.

This is not analysis. A government deliberately kept incapable of addressing problems it could address is not a minimal state. It is a chosen failure. An economic system declared permanently unreformable despite a century of evidence to the contrary is not rigorous theory. It is a doctrine. The learning dynamic makes both demands untenable on the same grounds: the conditions that made those positions approximately correct have substantially changed, and honest analysis must change with them.

VI. Fiduciary Accountability

State capacity and obligation expand together. This is not merely an observation about the growth of the state; it is a claim about accountability. A state that has developed the capacity to prevent the extraction of the community’s share of the wealth the institutional order enables, and then stops doing so — or allows that capacity to erode through regulatory capture — is not taking a neutral position. It is in breach of a duty to the community whose institutional order it was charged to maintain.

The mechanisms that drive structural inability to accumulate among lower-income households — rent extraction from concentrated housing markets, wage suppression from concentrated labor markets, asset concentration self-reinforcing through the returns to capital, extractive credit arrangements that capture rather than enable — are partly state capacity failures. The capacity to address these problems was developed. Antitrust law, labor law, progressive taxation, financial regulation — these represent the hard-won expansion of state capacity to prevent the reconcentration of wealth that unrestricted market dynamics tend to produce. In many cases that capacity was allowed to erode: regulatory capture, judicial reinterpretation, legislative attrition.4 The failure is not the absence of government. It is government in breach of its own mandate.

What distinguishes countries that achieved broad-based accumulation from those that received similar initial conditions but ended in oligarchy is not the starting point. It is institutional follow-through — the state’s continuous discharge of its duty to maintain the conditions under which the institutional order’s benefits are broadly shared. Rule of law rankings, read carefully, measure something more specific than procedural justice. They measure the degree to which the state actually discharges this duty rather than being captured by those most able to benefit from its neglect.

The full development of this argument — including the concept of the community as equity partner in the wealth the institutional order enables, and the evidence from cross-country comparison — belongs in companion essays. This essay’s contribution is to establish the legitimate mandate that makes the fiduciary role possible in the first place. The state cannot be held accountable for failing to protect what it was never legitimately charged to protect. Once the mandate is established — and this essay has argued that it is real, grounded in Lincoln’s criterion, indexed to state capacity, and validated by the evolutionary selection of what actually works — the accountability follows.

VII. Conclusion: Institutional Maturation, Not Scope Creep

The four categories of government functions described here are not a wish list. They are a record of learning.

The watchman state was not wrong. It was doing everything its state capacity permitted, and it was doing what the institutional frontier of its time allowed. Each subsequent category represents a state capacity threshold crossed, a problem identified that people could not address as well for themselves, and an institutional response that survived the evolutionary selection process — surviving not because it was theoretically optimal but because it worked well enough to retain legitimacy and to adapt when it fell short.

The cost-benefit constraint is real and binding in both directions. Government should not exercise state capacity merely because it has it. It should exercise state capacity when the collective benefit exceeds the collective cost, accounting for the costs of inaction as carefully as the costs of action. Opportunity costs run in both directions. The asymmetric version of the constraint — which counts only the costs of government action, never the costs of government inaction — is not economics. It is advocacy.

The process has no endpoint. Complexity, made possible by earlier learning, continuously generates new collective action problems. State capacity continuously approaches new thresholds, making possible what was previously impossible. Democratic communities continuously refine their understanding of what kind of community they want to be. There is no final foreseeable institutional design, no equilibrium state toward which the learning dynamic converges. There are only better and worse responses to the problems of the moment, and the state capacity — or its absence — to develop them.

The essays that follow in this series rest on the foundation laid here. The institutional ecology that legitimate government functions have built over two centuries is what constitutes the benefit against which tax obligation is measured. The state’s authority to act on behalf of the community’s share of the wealth the institutional order enables is grounded in the same legitimate mandate this essay has traced from Lincoln’s criterion through the evolutionary selection of what actually survives. Progressive taxation is not extraction. It is proportional participation in the joint project that made individual achievement possible — the project of building and maintaining the institutional order that capitalism requires.

The watchman says government’s role was fixed at the beginning. The revolutionary says the whole structure must be torn down and rebuilt. Both are rupture strategies — one backward, one forward — and both treat the accumulated institutional learning of the past two centuries as either illegitimate or irrelevant. The evolutionary view offers a third way that is neither timid compromise nor splitting the difference. The institutions we have embody real accumulated knowledge — imperfectly, incompletely, with capture and corruption and backsliding, but substantially. You don’t discard evolved systems because they are imperfect. You reform them, extend them, correct their failures, and build new state capacity where new problems demand it. This is how science progresses, how common law develops, how medicine advances. There is no reason institutional governance should be different.

Don’t overthrow government. Evolve it.

Notes

This essay is the fifth in a series developing the theoretical foundations of equitable taxation. Previous essays: “Why Are Capitalists Anti-Capitalist?” establishes that capitalism requires active institutional construction and maintenance. “Taxation Is a Public Service” develops the functional finance argument for what taxation actually accomplishes in a monetary sovereign. “The Benefit IS the Ability-to-Pay” reunites the benefit principle and the ability-to-pay principle, arguing they were always two expressions of the same underlying moral claim. “Capitalism’s Empirical Test: The US Fails, Czech Republic Passes?” examines the cross-country evidence on whether capitalism delivers on its promise of broad-based accumulation. This essay establishes the foundation that the others assumed: what the government was legitimately doing in the first place.

This essay uses household saving data from the OECD Distributional National Accounts and the BEA/BLS Distribution of Personal Saving to test which countries are delivering the core promise of capitalism: broad accumulation of economic surplus across the income distribution. The United States, commonly regarded as capitalism’s exemplar, shows a saving gradient—the gap between what the top and bottom income groups save—of nearly +100 percentage points, with the bottom 40–60% of households structurally prevented from saving. The Czech Republic, whose egalitarian asset distribution traces to communist-era housing privatization and compressed wage institutions, shows a gradient of +21 percentage points with positive saving at every quintile. The essay argues that this reversal is diagnostic, not paradoxical: capitalism’s outcomes depend substantially on the initial distribution of productive assets and the suppression of economic rent, not on the label attached to the system. Because the mechanisms driving the US pattern—rent extraction, monopsony, and extractive consumer credit—are features of the pre-capitalist institutions capitalism was supposed to replace, the appropriate policy response is therefore not redistribution, the traditional answer, but restorative justice—returning to households and communities the economic surplus that state-backed coercion and rent extraction have transferred upward.

Opening: The Paradox in the Data

Consider two countries.

In the first, the bottom quintile of households spends 165 cents for every dollar of income it earns. The second and third quintiles also spend more than they earn—by 22% and 4% respectively. Only the top two quintiles accumulate capital routinely. This is not a crisis anomaly. It is the stable, structural condition of the household sector, observed in every year from 2004 to 2019, across two recessions and a decade-long expansion.

In the second country, the bottom quintile saves 3% of its income (2.2% on the adjusted basis used in the cross-country table below). The second, third, and fourth quintiles all save between 5% and 9%. The top quintile saves 23%. Every quintile accumulates capital every year.

The first country is the United States. The second is the Czech Republic.

If we define capitalism as an institutional order in which the capacity to accumulate economic surplus is broadly available across the income distribution—not restricted to an elite—then the data pose a direct question: which of these two countries is doing a better job of being capitalist?

The question is not rhetorical. It is a measurement. And the measurement says something important about the gap between the label “capitalist” and the substance the label is supposed to denote.

The Test: What Capitalism Should Produce

In a companion essay, Why Are Capitalists Anti-Capitalist?, I argued that capitalism is best understood as an institutional order characterized by the generalized legal and practical capacity of individuals to accumulate, retain, and deploy economically consequential surplus. The key word is generalized. The goal of this definition is to isolate what was historically novel about capitalism—what distinguished it from prior economic orders. Private property, voluntary exchange, and market coordination are prerequisites of a modern capitalist state, but they do not distinguish capitalism from feudalism or mercantilism, both of which had private property and markets. What was historically revolutionary about capitalism was the extension of accumulation capacity beyond a privileged class—to the craftsman, the farmer, and eventually the wage worker.

Abraham Lincoln put this concisely in his 1859 address to the Wisconsin State Agricultural Society.¹ The “prudent, penniless beginner,” he observed, “labors for wages awhile, saves a surplus with which to buy tools or land for himself; then labors on his own account another while, and at length hires another new beginner to help him.” This is not a description of the very wealthy. It is a description of the institutional pathway that capitalism is supposed to open to ordinary people.

One important operational measure of whether that pathway is open is the Average Propensity to Save (APS): the fraction of income that households save rather than consume. If capitalism is working, APS should be positive—or at least not deeply negative—across the income distribution. Structural dissaving means that a household is spending more than it earns, year after year, sustained by borrowing, asset liquidation, or transfer from others. It is the arithmetic negation of the Lincoln pathway.

APS is a flow measure: it records the fraction of current-period income not consumed. It is therefore a proxy for one specific form of accumulation capacity—income-financed surplus—not for total wealth formation. The channels it does not capture directly include capital gains on housing and equities (which can raise wealth without any positive saving flow), human capital investment (education, skills development), direct equity acquisition through employer stock programs, inherited wealth transfers, and pension wealth accrual—though APS is itself influenced by levels of wealth, since households carrying less debt and more assets face lower prior claims on income. What APS isolates is the Lincoln pathway specifically: whether households can accumulate from earned income, independent of prior asset holdings or asset-price appreciation. Each of these other channels deserves scrutiny in its own right, and a complete account of whether capitalism is delivering on its promise would examine all of them. The most plausible alternative accumulation channel for Q2–Q3 is housing appreciation, which in the US generated large nominal gains for homeowners in 2019–2022 and is real. But it is unavailable to the 35–45% of Q2–Q3 households who rent; it is illiquid and non-deployable in the way tools or equity capital are; and the long-run wealth-share data show the middle quintiles losing ground across the full 1989–2022 period despite intermittent asset-price windfalls. This essay uses APS as one useful lens—not the only one—because it is consistently measured across countries and over time in existing distributional national accounts, and because the pattern it reveals is clear enough to sustain the argument without requiring a composite index.

This essay uses the OECD Experimental Statistics on Household Distributional Accounts (EG DNA) for nine countries and the BEA/BLS Distribution of Personal Saving (NIPA Table 2.9) for the United States to measure APS by income quintile. Both datasets anchor distributional estimates to national accounts aggregates, ensuring cross-country comparability. The US data uses Disposable Personal Income (DPI) as the denominator, which excludes Social Transfers in Kind and is most comparable to the OECD cash-basis measure.

The United States Picture

Table 1 shows US average propensity to save by income quintile for selected years from 2004 to 2022.²

Three features of Table 1 require comment.

The structural persistence of Q2 and Q3 dissaving. Q1 dissaving has a partial lifecycle explanation. The bottom quintile contains disproportionate numbers of young households borrowing to invest in human capital—education, skills, initial homeownership—and older households drawing down retirement savings. Both behaviors are consistent with rational lifecycle planning in a well-functioning economy. Figure 1, which shows average APS by age of household reference person for the United States (2024, BLS Consumer Expenditure Survey), confirms that the lifecycle hump is real: mean APS rises from +2.5% for households under 25, peaks at +30.1% for the 55–64 bracket, then declines as households enter retirement.

The lifecycle profile, however, does not explain away the quintile pattern, for two reasons. First, the CEX chart reports means across the full income distribution within each age bracket: the average 25–34 household earns approximately $100,000 and saves 27%. The Q1 household within that same age bracket earns a fraction of that—and dissaves. Age-bracket means conceal exactly the income-rank variation that the quintile data expose. Second, the lifecycle argument cannot account for Q2 and Q3 dissaving. Both quintiles consist predominantly of prime working-age households—the same demographic the CEX lifecycle profile shows saving at 25–30%. If the age composition of Q1 were the primary driver of its dissaving, Q2 and Q3 should be comfortably positive. They are not. A system in which the median household cannot save is not fulfilling the promise of capitalism, regardless of what it calls itself.

The gradient is large and stable. The difference between Q1 and Q5 ranges from +91 to +112 percentage points across the full series. This is not driven by exceptional Q5 saving: at 31–41%, Q5 saving is high but not extraordinary by international standards. The gradient is wide because of the depth of Q1–Q3 dissaving. And it is stable: two recessions, a financial crisis, a decade of low interest rates, and a global pandemic have each modestly compressed the gradient temporarily, then allowed it to snap back.

The pandemic as a natural experiment. The 2020–2021 stimulus episode is instructive. Two simultaneous forces raised Q1–Q3 saving rates substantially: federal transfers—direct payments, expanded unemployment insurance, student loan forbearance—and the lockdown-induced collapse of spending on restaurants, entertainment, and travel, which suppressed consumption even when households would otherwise have spent. Q3 briefly turned positive in 2020. Even Q1, however, never crossed zero: even at the peak of the largest peacetime transfer program in US history, the bottom quintile spent more than it earned. By 2022, with stimulus expired and inflation eroding real income, the gradient had widened to +104 pp and Q4 had joined Q1–Q3 in negative territory. The structure that stimulus momentarily concealed reasserted itself faster than it had been suppressed. Figure 2 traces the full temporal picture.

What sustains Q1–Q3 consumption above income? Three channels: consumer credit (credit cards, auto loans, payday lending), asset liquidation (drawing down savings, cashing out home equity), and inter-household transfers (remittances, family support) that are invisible to the national accounts. None of these is an accumulation mechanism. All three are instruments of decumulation—ways of consuming capital rather than forming it. Lincoln’s prudent penniless beginner, for three-fifths of American households, is not saving toward tools. He is borrowing against the future to cover the present.

Ten Countries: Where the US Sits

The OECD data cover nine countries across Europe and East Asia; together with the US, the full comparison spans ten countries at varied levels of income, welfare-state design, and economic history. The US is not unique in showing a positive APS gradient—all ten countries do. The rich save more than the poor everywhere and always. What varies is the magnitude of the gradient and where the quintile profiles cross zero. Table 2 shows the pattern.³

Several features of Table 2 merit attention.

The variation is not explained by national income level. Mexico, the poorest country in the sample, has a gradient of +66 pp and four dissaving quintiles. Czech Republic, a high-income European economy, has a gradient of +21 pp and zero dissaving quintiles. Belgium and Canada, both wealthy welfare states, show gradients above +80 pp. Income level predicts neither the gradient width nor the number of dissaving quintiles.

Welfare state generosity does not close the gap. Belgium and the Netherlands both have large, well-funded welfare states with substantial in-kind transfers (public healthcare, housing subsidies, education). Both still show Q1 dissaving of −32% to −40% and gradients of +69 pp to +80 pp. Income redistribution—the transfer of cash and services from higher to lower quintiles—compresses income inequality but does not eliminate structural dissaving at Q1–Q2. The reason is that redistribution operates on income flows, while dissaving is driven by the balance sheet: the stock of assets and debts that households carry into each period. A household with a mortgage absorbing 40% of gross income will dissave even if its cash income is supplemented by transfers, because the debt service is prior to any saving decision.

The US gradient leads the sample. At +98 pp, the US gradient is the widest in the sample on the basis shown, with Canada close behind (+94 pp on the adjusted basis; wider still on a comparable cash basis), and both substantially larger than Belgium (+80 pp), Korea (+66 pp), or Slovenia (+54 pp). The US is at the worst end of a group that already includes countries with deep Q1 dissaving.

Figure 3 shows the full ten-country comparison.

The Czech Republic Counterexample

Czech Republic deserves extended attention because it is not a minor outlier. It is a clean falsification of the claim that structural Q1–Q3 dissaving is an unavoidable feature of advanced market economies.

Czech Q1 saves +2.2% of adjusted income (or +3.0% on the cash basis). Czech Q2 saves +5.2%. Czech Q3 saves +5.5%. The gradient is +21 pp. These are not approximations of zero; they are definitively positive at every quintile in a country with GDP per capita comparable to Portugal and well above Mexico. The Czech Republic’s Gini coefficient (0.242) is the lowest in the ten-country sample, nearly 60 points below Mexico and well below the US and Canada.

Candidate explanations. No single structural factor fully accounts for Czech Republic’s breadth of saving. The available data support two as strong candidates; others are plausible on theoretical grounds and warrant further investigation, but cannot be established from national accounts data alone.

Debt-free homeownership. Between 1989 and 1992, the Czechoslovak and then Czech state privatized the bulk of the state-owned housing stock at nominal prices to sitting residents. This was not a market outcome; it was a political distribution. The effect was to eliminate from Czech Q1–Q3 households the single largest driver of dissaving in the US, Canada, and Belgium: mortgage debt service or rent paid to landlords in higher quintiles. In the US, a bottom-quintile household that rents from a top-quintile landlord transfers economic surplus upward with every monthly payment, before any saving decision is possible. Czech Q1–Q3 households, owning their homes outright, faced no such prior claim on their income.

The legal structure of the privatisation mattered as much as the fact of it. The Czech state conveyed direct, individually registered title to the sitting resident—not a tradeable financial instrument. That structural choice made rapid reconcentration structurally difficult: a housing title requires the full apparatus of property law to transfer, including formal registration, notarized transaction, and a functioning title-registry system. The contrast with the Russian voucher privatisations of industrial enterprises is instructive. Each Russian citizen received a tradeable voucher worth 10,000 roubles; most sold for cash within months to well-connected buyers who aggregated controlling stakes before any institutional infrastructure to prevent reconcentration had formed. Czech housing titles could not be similarly swept up because the transferability mechanism required precisely the Rule of Law infrastructure— title registries, property courts, transaction recording—that the state was simultaneously building. The distribution and the institution reinforced each other; neither was sufficient alone.

Figure 4 quantifies the tenure difference. Among households below 60% of median income in 2019, 49.7% of Czech households own their homes outright (no mortgage), compared with 25.0% in Belgium, 28.1% in Ireland, 17.1% in France, and 11.0% in the Netherlands. The contrast with the broader post-communist cluster is equally striking: Romania (93.5%), Croatia (87.1%), Hungary (77.2%), and Poland (75.4%) all show even higher outright ownership among low-income households, confirming that the Czech result is part of a systematic regional pattern.

The tenure data also provide an important caution against monocausal explanation. Slovenia—also a post-communist country with a similar history of housing privatisation—has a low-income outright ownership rate of 52.1%, marginally higher than Czech Republic’s 49.7%, yet its Q1 APS was −24.6% in 2019. Outright homeownership appears to be a necessary condition for broad saving but not a sufficient one. Something else differentiates the Czech outcome from the Slovenian outcome, and whatever that something is, it matters as much as the tenure structure.

Slovakia would be a sharper test still. Czech Republic and Slovakia were a single country until 1993 and underwent identical housing privatisation programs in 1989–92; the Eurostat data confirm Slovakia’s low-income outright ownership rate of 64.5%, higher than Czech Republic’s 49.7%. The Velvet Divorce itself was driven primarily by differences in industrial base and labour market quality—precisely the wage-compression channel this essay identifies as the second candidate explanation. If Slovak distributional APS data showed worse Q1–Q3 outcomes than Czech Republic’s despite its higher outright ownership rate, it would constitute a near-ideal natural experiment: identical housing conditions, divergent wage structures, cleanly separated by a known historical event. Slovakia is not currently in the OECD Distributional National Accounts sample; adding it would be the most informative single extension of this line of research.⁴

Compressed wages. The most plausible candidate for the additional factor is wage compression. Communist wage institutions produced a flat wage distribution that survived the post-1989 transition, and Czech Republic’s Gini coefficient (0.242) remains one of the lowest in the OECD sample. A household that owns its home outright can still dissave if income falls below the cost of food, utilities, healthcare, and transport. Czech Q1 wages as a fraction of mean wages are higher than in comparator countries, providing a margin above subsistence that is arithmetically unavailable to the US bottom quintile—and, apparently, to Slovenia’s as well.

Other candidates. Several additional factors may contribute, individually or in combination: Czech Republic’s unemployment rate was approximately 2% in 2019 (among the EU’s lowest), keeping Q1 income more wage-driven and less benefit-dependent than in higher-unemployment comparators; Czech Republic retains its own currency, allowing real wages and prices to adjust independently of the eurozone price floor to which Slovenia is bound; and historical differences in consumer debt levels at Q1–Q2 may create different residual prior claims on income. Disentangling these contributions requires microdata analysis beyond the scope of this essay.

The paradox stated explicitly. The country that best satisfies the operational test for capitalism got there through a process that was the opposite of market-driven: a communist-era redistribution of productive assets and a state-suppressed wage distribution. The US, which proclaims itself the capitalist exemplar, fails the same test.

This is not a paradox. It is a confirmation of the theory. Capitalism’s outcomes—broad or concentrated accumulation—depend on the initial distribution of productive assets and the degree to which rent extraction—returns from ownership and monopoly power rather than from productive contribution—has been suppressed. Czech Republic did not achieve broad accumulation through capitalism; it achieved it by starting from a distribution of assets that allowed market processes to function without creating the rent-income spiral that generates structural dissaving in Q1–Q3 elsewhere.

Standard rankings corroborate the equivalence. The Heritage Foundation’s Index of Economic Freedom—the most widely cited single-number measure of market capitalism—rates both countries in its Mostly Free tier, within a few index points of each other.⁵ The World Justice Project’s Rule of Law Index has ranked Czech Republic slightly above the United States in recent editions.⁶ The notable exception is Transparency International’s Corruption Perceptions Index, where the United States scores materially higher—a genuine institutional difference that warrants caution against treating the Czech case as a simple template.⁷

The connection to George and Piketty. This is not a novel observation. Henry George argued in Progress and Poverty⁸ (1879) that land rents—returns to an asset no one produced—capture the surplus generated by the whole community and concentrate it in the hands of whoever holds legal title, regardless of merit or effort. The mechanism he described applies precisely to what drives Q1 dissaving in the US: rent payments transfer income from tenants (disproportionately Q1–Q2) to landlords (disproportionately Q4–Q5) before any saving decision is made. The transfer is not a market failure; it is market success—for the asset owner.

Thomas Piketty’s central result⁹—when the return on capital (r) exceeds the growth rate (g), wealth concentrates geometrically over time—describes the dynamic that follows from an unequal initial distribution. If Q5 saves 33% of income while Q1 dissaves 65%, the asset gap compounds in every period. The APS gradient is not just a snapshot of inequality; it is a lower bound on the rate at which inequality is widening in the current period. Because capital gains accrue almost entirely to households already holding appreciable assets, the APS gradient likely understates the actual divergence in wealth-formation rates: Q4–Q5 benefits from both positive saving flows and asset appreciation, while Q1–Q3 has access to neither. Czech Republic, with its near-flat saving profile, is running a different dynamic: one in which accumulation is broadly distributed, and the Piketty spiral is not yet strongly in play.

A necessary hedge. The Czech Republic is one observation, and many other post-communist countries received similarly “clean” initial asset distributions in 1989–91 and ended up with extreme oligarchy: Russia, Ukraine, Romania. The initial distribution was necessary but not sufficient. Russia’s experience illustrates the failure mode precisely: Russian industrial privatisation used tradeable vouchers rather than registered individual titles. Without an institutional infrastructure of transparent title, functioning property courts, and enforceable transfer records—without, that is, the Rule of Law acting as settlor of accounts—the distributed assets were reconcentrated within years. What Czech Republic also had—and Russia lacked—was that institutional follow-through: a functioning rule of law, relatively low levels of post-transition corruption, and the disciplining pressure of EU accession requirements. The lesson is therefore not “redistribute assets and capitalism will follow.” The lesson is that the initial distribution of assets and the institutional capacity to prevent their subsequent reconcentration through rent and coercion both matter. The first condition Czech Republic got from its communist history; the second it built through its European integration.

The saving record of the other post-communist countries reinforces this point from the opposite direction. Romania (93.5% low-income outright ownership), Croatia (87.1%), and Hungary (77.2%) preserved distributed housing just as Czech Republic did—and yet preliminary estimates from Eurostat Household Budget Survey data suggest their Q1 dissaving is at least as deep as in Belgium or the Netherlands. High outright ownership evidently cannot prevent dissaving at the subsistence margin when absolute income levels are too low and institutional quality is insufficient. The Czech outcome required all three conditions simultaneously: distributed housing, adequate wage floors, and the rule-of-law environment that prevented reconcentration. The first two were communist legacies; the third was built through EU accession.

Diagnosis: What the US Has Instead of Capitalism

The US data do not show capitalism failing to deliver on its promise. They show a system in which market mechanisms coexist with structural features that suppress broad accumulation—features that are not capitalist in the sense defined above but that have been given capitalism’s name by the people who benefit from them.

Four mechanisms are responsible for most of the structural dissaving in Q1–Q3.

Before enumerating them, it is necessary to address the most common objection to structural explanations of this pattern: that Q1–Q3 dissaving reflects cultural preference—consumerism, poor financial discipline, present-biased discount rates—rather than institutional constraint. The objection rests on a category error. A household paying 40% of gross income in rent is not making a consumption choice; it is servicing someone else’s asset, meeting a prior claim on income that was established before any saving decision became possible. Mortgage debt service, consumer credit interest, and non-negotiable healthcare premiums operate the same way. Prior claims are compulsory transfers to asset owners, enforced by the legal system. The relevant distinction is between expenditure that can be reduced by preference (discretionary consumption) and payments that cannot be avoided without forfeiting access to housing, transportation, or healthcare (prior claims). All four mechanisms below operate through prior claims—through the structure of asset ownership and debt, not through the preferences of lower-income households.

Asset concentration and rent flows. When Q4–Q5 owns most of the housing stock, stock market wealth, and business equity, the income from those assets bubbles up to the top quintiles regardless of labor market outcomes. More importantly, the corresponding costs—rent, interest, fees—flow out of lower quintiles as prior claims on income. A household paying 40% of gross income in rent is not making a consumption choice; it is servicing an asset that its counterpart in Czech Republic owns outright. The rent payment is a transfer of economic surplus from Q1–Q2 to Q4–Q5 that occurs upstream of any saving decision.

Income composition at the top. The data show that APS rises sharply with income. The behavioral explanation is that high-income households simply prefer to save more. The structural explanation is different: as income rises through the quintile distribution, it becomes increasingly composed of returns to asset ownership—dividends, capital gains, rent, royalties—rather than returns to labor. These are returns to things, not people. They reflect legal entitlements enforced by the state, not productive contribution. At the extreme top of the income distribution, the bulk of income is economic rent in the technical sense: returns that exceed what would be required to elicit the underlying activity. This income is not capitalism producing its reward for contribution; it is the state enforcing property claims that allow surplus extraction without productive contribution.

Market power and wage suppression. Competitive labor markets in a genuinely capitalist economy would drive wages toward the marginal product of labor. The US labor market, in significant sectors, does not function this way. Monopsony power in labor markets (the buyer’s equivalent of monopoly: a market dominated by a single or small number of employers)—documented by Azar, Marinescu, and Steinbaum, among others¹⁰—allows employers to pay wages below competitive levels, suppressing Q1–Q3 income and therefore their saving capacity. This is not a market outcome; it is a failure of market competition, enabled by regulatory tolerance and the erosion of countervailing institutions (unions, labor standards enforcement).

Consumer credit as surplus extraction. Credit is not inherently non-capitalist. But when Q1–Q3 households sustain above-income consumption through consumer credit at high interest rates, the interest payments constitute a systematic transfer of economic surplus from lower to upper quintiles. The credit market functions as a mechanism for extracting the future income of Q1–Q3 households on behalf of the financial sector—which is disproportionately owned by Q4–Q5. The household is not accumulating capital; it is consuming it in advance, at a price.

These four mechanisms are not features of capitalism. They are features of the systems capitalism should have replaced: feudal rent, aristocratic monopoly, debt peonage. To call a system that runs on these mechanisms “capitalist” is to mistake the label for the substance.

The Policy Implication: Restorative Justice and the Silent Partner

If one sees value in pursuing the promise of capitalism, the appropriate policy response follows from the diagnosis. If non-capitalist rent-extraction mechanisms have suppressed broad accumulation, the remedy is to dismantle those mechanisms—not to impose a preferred distributional pattern on top of them.

The moral foundation: Nozick and Carnegie. Robert Nozick argued,¹¹ correctly, that redistribution to achieve patterned outcomes is morally problematic: it treats the holdings of individuals as resources to be reassigned by the state in pursuit of a social ideal—treating persons as instruments of a collective goal rather than as ends in themselves. The question the APS data forces, however, is whether the concentrated holdings in question were legitimately acquired in the first place.

Andrew Carnegie—not ordinarily cited in debates about economic justice—supplied the missing term. In The Gospel of Wealth¹² (1889), he argued that “wherever great wealth accrues honorably, the people are always silent partners”: the public infrastructure, the stable legal order, the network of workers and customers, and above all the location value created by the simple presence of others all contribute to the returns that accrue to the named owner. The silent partner’s contribution is real; it was never compensated; and it creates a prior equity claim that exists independent of any question about redistribution.

Nozick’s legitimacy condition and Carnegie’s silent partner converge on the same conclusion. Restorative justice is not the state taking from wealth-holders; it is the state returning to the silent partner what the silent partner produced and was never paid. The distinction from redistribution is not semantic. Redistribution is motivated by a desired outcome and has no principled stopping point. Restorative justice is motivated by a prior uncompensated contribution and stops when that contribution has been returned. The state does not impose a pattern; it corrects extractions it was complicit in enabling.

The Rule of Law as fiduciary enforcement. This framing gives the Rule of Law a specific meaning in context. The legal order is not merely a precondition that capitalism requires from outside; it is the mechanism by which the community, acting as equity holder, enforces its own prior claim against those who would appropriate the silent partner’s share. When the state permits oligarchic reconcentration—through rent extraction, regulatory capture, or market power—it is not merely failing a policy objective; it is breaching a fiduciary duty to the beneficiary whose contribution it was charged to protect. The WJP Rule of Law rankings noted above can be read as measuring precisely that fidelity.¹³

Three areas of reform follow from this logic.

1. Asset claims: securing the partner’s equity. The primary driver of Q1–Q3 dissaving is the prior claim that rent and debt service make on income before any saving is possible. When institutional investors acquire residential housing as a financial asset, they are not producing a new good; they are extracting the location value that the surrounding community created. That is the silent partner’s equity, not the investor’s. Policy that limits institutional ownership of single-family housing, strengthens pathways to homeownership, and reforms bankruptcy and debt-collection law does not penalize production; it removes the toll booth between the worker and the Lincoln pathway. The Czech lesson is that the balance sheet is the operative variable: income transfers that leave the asset distribution unchanged do not close the saving gap, as Belgium and the Netherlands demonstrate.

2. Labor market competition: ending monopsony as restitution. Competitive labor markets drive wages toward the marginal product of labor. When monopsony power allows employers to pay wages below competitive levels, the gap between the competitive wage and the actual wage is not the employer’s legitimate return; it is an excess appropriation extracted from the silent partner’s share. Employers are entitled to appropriate the surplus their capital and organization generate—that is why employment exists. The objection is not to appropriation as such but to the excess that market power makes possible at the silent partner’s expense. Antitrust enforcement against monopsony, minimum wage floors set at competitive levels, and the right to collective bargaining are not concessions to anti-market ideology. They restore the boundary between the employer’s legitimate return and the excess that competitive markets would not have permitted.

3. Rent capture: the cleanest accounting. Henry George’s proposal—heavily taxing economic rent while reducing taxes on labor and productive capital—is an appropriate instrument for honoring Carnegie’s standard. Returns to location, natural resources, network effects, intellectual property monopolies, and financial leverage are, by definition, returns that the recipient’s productive contribution did not create. They are the silent partner’s surplus. Taxing economic rent does not penalize production; it recaptures for the community the value that the community created but that the legal system currently assigns to named title-holders.

These reforms are difficult. Regulatory institutions can be captured by the interests they regulate; that is Buchanan’s insight and it is correct. Marxists argue that the mechanisms of exploitation are inherent to capitalism and cannot be reformed away; that argument too has empirical support in the persistence of the US gradient across administrations and reform cycles. But neither objection is an argument for despair. Both name the challenge of building institutions strong enough, and sufficiently resistant to capture, to deliver on capitalism’s promise. Czech Republic built those institutions—imperfectly, partially, with European help. The fact that it is difficult is not a reason to stop trying. It is a description of the work that needs doing.

Restorative justice as ongoing commitment. The Czech case makes one lesson especially clear: restorative justice is not a one-time event. The 1989 housing privatisation was not a permanent solution; it was a starting condition that still required active institutional maintenance. What distinguishes Czech Republic from the post-communist countries that received the same initial asset distribution and then lost it to oligarchy is not what happened in 1989—it is what the state continued to do afterward. Capitalism’s promise, once created, does not sustain itself. The rents that concentrated ownership generates compound continuously, and without a state acting as perpetual settlor of accounts—continuously ensuring that market processes do not reconvert Lincoln’s road into a toll bridge—each generation inherits a narrower pathway than the last. This is not an argument against markets. It is an argument for the institutional vigilance that keeping capitalism capitalist requires.

Conclusion: The Measurement Is the Argument

The definition of capitalism offered here is explicitly normative. It establishes a standard—generalized capacity for accumulation—and asks which countries are meeting it. The fact that no system fully achieves the standard is expected; normative definitions are measuring instruments, not descriptions of current reality. We define a healthy diet normatively and then ask how far short of it Americans fall. The definition is not circular because the gap it measures is real and consequential.

What the APS data show is that the United States is far short of being a leading example of real capitalism.

For 2004–2022, the bottom 40–60% of American households have been structurally prevented from saving. The pandemic stimulus of 2020–2021 temporarily closed the gap, then the gap snapped back—and widened.

The mechanisms responsible are not failures of capitalism; they are the operation of non-capitalist elements—rent extraction, monopsony, asset concentration, consumer credit at extractive rates—that have been allowed to dominate the distribution of accumulation opportunities while claiming capitalism’s name.

The Czech Republic, whose egalitarian asset distribution traces to a communist-era housing privatization and compressed wage institutions, passes the test the United States fails. This is not a recommendation to import Czech institutions. It is a demonstration that the capitalist ideal of generalized accumulation is empirically achievable, and that the path to it runs through the initial distribution of productive assets and the suppression of rent—not through the label attached to the system.

The political implication is uncomfortable for both sides of the conventional debate. Those who call themselves capitalism’s defenders are often defending the rent-extraction mechanisms that suppress broad accumulation. Those who call themselves capitalism’s critics are often attacking a system that does not deserve the name they are attacking. Both groups are, in different ways, complicit in maintaining a system that delivers concentrated accumulation while calling it something else.

Stop defending the label. Start measuring the substance.

For the United States: it is time to do a better job of being capitalist.

Notes

1. 1 Lincoln, A. (1859). “Address to the Wisconsin State Agricultural Society.” Milwaukee, Wisconsin, September 30, 1859.

2. 2 US data are from the BEA/BLS joint Distribution of Personal Saving (NIPA Table 2.9 distributional estimates), 2004–2022, published July 2024. APS is computed as (DPIq−PCEq)/DPIq, where DPI and PCE quintile shares are drawn from the BEA decile-level release and summed to quintiles. Methodology: Gindelsky (2025), Review of Income and Wealth; BEA (2024), The Methodology for Distributing Personal Saving via a Blended Survey–National Accounts Approach. Because DPI excludes Social Transfers in Kind (D63), the closest OECD comparator is the cash basis (B6G), not the adjusted basis (B7G) shown in Table 2; the US gradient would be somewhat wider on a strictly comparable basis. BEA distributed totals do not exactly match NIPA aggregates because some flows cannot be attributed to quintile households; quintile-level APS is internally consistent within the distributed framework. CEX under-reporting may modestly overstate Q5 APS relative to OECD DNA methods.

3. 3 OECD data for nine countries are drawn from the OECD Experimental Statistics on Household Distributional Accounts (dataflow DSD_EGDNA_INC_INC), accessed via the OECD SDMX REST API. The APS shown is the adjusted basis: APSadj=B8G/B7G, where B8G is gross saving and B7G is gross adjusted disposable income including Social Transfers in Kind (D63). Reference years: 2019 for Belgium, Canada, Czech Republic, Korea, Netherlands, and Slovenia; 2022 for France, Ireland, and Mexico. Cross-country comparability caveats: (1) The US DPI concept excludes D63; the closest OECD comparator is the cash basis (B6G), so the US gradient shown is somewhat narrower than a fully comparable figure would be. (2) France’s P41 submission covers cash consumption only (P3), making its adjusted-basis APS not directly comparable to other countries. (3) Quintile equivalisation scales differ across datasets. Despite these caveats, qualitative conclusions are robust: the US gradient of +98 pp would be wider, not narrower, on a fully comparable cash basis; Czech Republic’s positive saving at every quintile is confirmed on both adjusted and cash bases.

4. 4 The Czech–Slovak comparison satisfies the two main requirements for a natural experiment: a common shock (identical housing privatisation in 1989–92, confirmed by the Eurostat tenure data showing comparable low-income outright ownership rates) and a plausibly exogenous divergence (the Velvet Divorce separated two economies with different industrial structures and wage institutions, for reasons—constitutional disagreement, pace of liberalisation—largely unrelated to housing policy). The counterfactual prediction is crisp: if wage compression drives the Czech breadth of saving, Slovakia should show worse Q1–Q3 APS than Czech Republic despite its higher outright homeownership rate (64.5% vs. 49.7%). Slovak quintile-level distributional saving data do not appear to be in the public OECD DNA release. A preliminary estimate combining the Eurostat Household Budget Survey (HBS, 2015) for consumption expenditure with EU-SILC income distributions—two separate surveys joined under strong methodological caveats—places Slovak Q1 APS at approximately −67%, comparable to Belgium (−62%) and the Netherlands (−64%) on the same rough measure. Czech Republic Q1 on the same measure is approximately −7%: still the lowest dissaving rate in the fifteen-country sample, and dramatically better than Slovakia despite Slovakia’s higher outright homeownership rate. The directional prediction of the natural experiment is therefore confirmed in the available data: identical housing privatisation, sharply divergent saving outcomes. Constructing rigorous estimates—via OECD DNA accession for Slovakia or a consistent consumption imputation from EU-SILC microdata—remains the most informative single extension of this line of research.

5. 5 Heritage Foundation, Index of Economic Freedom, published annually.

6. 6 World Justice Project, WJP Rule of Law Index, published annually.

7. 7 Transparency International, Corruption Perceptions Index, published annually.

8. 8 George, H. (1879). Progress and Poverty. D. Appleton & Co.

9. 9 Piketty, T. (2014). Capital in the Twenty-First Century. Harvard University Press.

10. 10 Azar, J., Marinescu, I., & Steinbaum, M. (2022). “Labor Market Concentration.” Journal of Human Resources, 57(S), S167–S199.

11. 11 Nozick, R. (1974). Anarchy, State, and Utopia. Basic Books.

12. 12 Carnegie, A. “The Gospel of Wealth, Part II.” North American Review, 183(604):1096–1106 (Dec. 7, 1906). Originally published as “Wealth,” North American Review, 148(391):653–664 (1889). JSTOR 25105713.

13. 13 The connection between Rule of Law and the silent partner’s equity will be developed more fully in a companion essay, “Carnegie’s Silent Partners.”

Yet Seligman and his colleagues have largely vanished from contemporary tax policy discourse. Today’s debates invoke efficiency, revenue maximization, and optimal tax theory—technical frameworks that developed long after the system was built. The moral and civic foundations that Seligman labored to establish have been quietly dropped from the conversation, even though the structure those foundations produced is still the one we operate within. How did the architects of the income tax become so thoroughly forgotten?

This essay argues that their disappearance is connected to a conceptual error—one that Seligman himself helped create, with the best of intentions.

Every economics textbook presents two rival principles of tax justice. The first is the Benefit Principle: citizens should pay taxes in proportion to the benefits they receive from government. The second is the Ability-to-Pay Principle: citizens should pay in proportion to their capacity to bear the burden, which in practice means the wealthy pay more.¹ Students of economics and the philosophy of taxation learn to contrast these as competing visions of fairness—the first rooted in exchange, the second in obligation.

Seligman and his generation attacked the Benefit Principle as a cramped, transactional idea that reduced citizenship to commerce, and championed Ability-to-Pay as the morally superior foundation for a new fiscal order. Their framing became dominant in academic and policy discourse. But in establishing that framing, they severed the link between what government does for individuals and what those individual citizens owe in return—and that severance has drained the moral content out of tax policy over time. Ability-to-pay, standing alone, offers a thin justification: we take from you because you have what we need and because we can. It positions government and citizen as adversaries in a zero-sum contest over resources. This is precisely the rhetoric that dominates American politics today—and it is the opposite of what Seligman wanted.

The thesis of this essay is that the rivalry between these two principles was always false—an artifact of defining “benefit” too narrowly. Once we properly understand what government actually provides, its proper function, the two principles converge on the same idea. And that convergence, far from reducing taxation to a transaction, restores exactly the civic, relational understanding of fiscal obligation that Seligman and his colleagues were reaching for but could not quite grasp.

Adam Smith’s Overlooked Unity

The supposed rivalry between benefit and ability-to-pay was not always taken for granted. Indeed, for more than a century before Smith, the benefit principle stood essentially unchallenged as the justification for taxation. Thomas Hobbes, in Leviathan (1651), argued that taxes were justified as the price of security—the benefit the sovereign provided in lifting subjects out of the state of nature. John Locke, in his Second Treatise of Government (1690), saw taxation as payment for the protection of property rights, which embodied the individual’s liberty. Ability-to-pay had not yet been articulated as a competing framework. Benefit was simply what taxation was for: it was payment for services rendered. As Justice Oliver Wendell Holmes would later put it, “Taxes are what we pay for civilized society”—a pure benefit principle statement, offered as common sense well after the academic economists had begun to abandon that idea.²

Smith was working within this benefit tradition when he stated his first canon of taxation in The Wealth of Nations (1776). But he did something that Hobbes and Locke had not done: he brought ability-to-pay into the picture and, in a single carefully constructed sentence, declared it identical to benefit:

The subjects of every state ought to contribute towards the support of the government, as nearly as possible, in proportion to their respective abilities; that is, in proportion to the revenue which they respectively enjoy under the protection of the state.

The crucial phrase is “that is.” This is best read not as Smith contrasting two ideas, but rather equating them. A subject’s “respective abilities” (ability-to-pay) simply is the revenue that subject enjoys “under the protection of the state” (benefit received). Smith reinforced the point with an analogy: the expense of government is “like the expense of management to the joint tenants of a great estate, who are all obliged to contribute in proportion to their respective interests in the estate.” Smith viewed us all as shareholders or investors in society. Your interest in society and its government is simultaneously the measure of what you have gained from it and the measure of what you can afford to contribute toward its upkeep. Benefit and ability are not two principles but two descriptions of the same thing.

Smith saw no tension here. But his formulation introduced the vocabulary that made the later conflict possible. By articulating both ideas so clearly and in such close proximity, he gave future readers two phrases that could be pulled apart and treated as alternatives rather than as the equivalence he intended.

The real break came with John Stuart Mill. In his Principles of Political Economy (1848), Mill developed ability-to-pay as a freestanding principle grounded in equal sacrifice theory—the idea that taxation should impose an equal loss of utility on each taxpayer. This was a powerful and influential framework, but it was detached from any account of the benefit government provides to individual tax payers. Mill made it possible to discuss what citizens owe without asking what government does for them. Ability-to-pay became an independent principle, no longer merely a proxy for benefit as it had been in the tradition from Hobbes through Smith.

Seligman’s generation then completed the separation by explicitly attacking the benefit principle as inadequate. But the seeds of the false rivalry were planted the moment readers began treating Smith’s unified sentence as containing two separable ideas.

Lindahl’s Unworkable Framing

The Swedish economist Erik Lindahl, writing in 1919, explicitly argued that there was no necessary contradiction between the principles of benefit and ability-to-pay. Ability to pay, Lindahl suggested, could often serve as a reliable indication of the benefit an individual derives from public expenditure.³

But Lindahl did not leave the convergence as a simple observation. He proposed a mechanism for measuring benefit: personalized “Lindahl prices,” a conceptual auction in which each citizen would reveal their willingness to pay for public goods. In this equilibrium, every individual would consume the same quantity of public goods but face a different price, reflecting their personal valuation of the individually quantifiable benefits of government.

The mechanism was elegant but immediately recognized as impractical. When the only source of information about marginal benefits is the individuals themselves, they have every incentive to understate their valuations—the classic free-rider problem. The conceptual auction requires honest preference revelation, and honest preference revelation is precisely what public goods make impossible.

But the deeper problem with Lindahl’s approach was not merely practical. By framing benefit as the sum of an individual’s valuations of specific, separable government services, Lindahl’s mechanism implicitly reduced the benefit of government to an atomistic accounting exercise. The whole was treated as nothing more than the sum of its parts—my valuation of roads, plus my valuation of courts, plus my valuation of defense. This atomistic framing loses sight of the aggregate benefit provided by the institutional order as a whole, the order within which all economic life takes place. The value of the institutional ecology is not the sum of the values of its individual components, any more than the value of a functioning body is the sum of the values of its individual organs.

Lindahl asked the right question: Can benefit be measured? But his answer—however elegant in theory—did lasting damage to the question itself. By reframing benefit as something that could, or should, be determined through an impractical mechanism of individual preference revelation, Lindahl left later critics reasonably skeptical of the benefit principle’s utility as an implementable idea. The baby went out with the bathwater: when the measurement mechanism was rejected, the convergence of benefit and ability-to-pay was rejected along with it. Yet as Linda Sugin has argued, “Even if everyone agrees that a tax is impractical and impossible to administer, it should still be examined if it embodies the ultimate measure of fairness in taxation. If it is the best ideal, then actual tax systems should be evaluated against it.”⁴ The benefit principle’s measurement difficulties were real, but they were grounds for finding a better proxy—not for abandoning the principle.

Musgrave’s Fateful Separation

Richard Musgrave, arguably the most influential public finance theorist of the twentieth century, came closer than anyone to preserving Smith’s unity—and then deliberately abandoned it.

In The Theory of Public Finance (1959) and later in Public Finance in Theory and Practice (1980, with Peggy Musgrave), Musgrave acknowledged the connection between the two principles. Benefits, the Musgraves wrote, could be “viewed in terms of protection received and are thus related to income which, in turn, is also a measure of ability to pay.” The convergence was right there on the page.

But Musgrave chose not to build on it. Instead, he separated the functions of government into three distinct “branches” of the public budget, each with its own principle for how taxes should be structured: an allocation branch, which would tax according to the benefit principle to fund the efficient provision of public goods; a distribution branch, which would tax according to ability-to-pay to handle the equitable sharing of tax burdens; and a stabilization branch, responsible for maintaining full employment and price stability. Notably, Musgrave assigned the allocation and distribution branches each a principle for distributing the tax burden, but the stabilization branch received only a macroeconomic target—no principle for who should pay. Musgrave’s analytical separation became a major source of the textbook separation that later hardened into orthodoxy and is largely responsible for the “false rivalry” that persists in economics education today.

Yet the stabilization function—the government’s role in maintaining employment and price stability—is itself one of the most significant benefits government provides. A stable currency, manageable inflation, and an economy operating near full employment are preconditions for nearly all private economic activity. Musgrave’s own framework, by assigning stabilization to a separate branch, obscured the fact that stabilization is a massive, continuous benefit of the institutional order—one that is reflected in every citizen’s economic position.

Why did Musgrave separate what Smith had unified? The answer lies in a revealing phrase: “protection received.” When Musgrave described the benefit of government, he framed it in terms of protection—a narrow, minimalist conception that implies government’s role is to stand guard over wealth that individuals generate autonomously. Under this framing, the two principles really do seem to pull in different directions: protection is roughly equal for everyone (or at most proportional to holdings), while ability-to-pay is inherently progressive. The separation seemed necessary because the definition of benefit was too thin to do the work Smith had asked of it.

Murphy and Nagel: The Foundation Without the Conclusion

In The Myth of Ownership (2002), the philosophers Liam Murphy and Thomas Nagel mounted what may be the most powerful philosophical argument about taxation in recent decades. Their central claim is that pretax income has no moral significance. Private property, they argued, is not a natural right that precedes government; it is a legal convention constituted by the institutional order. Without government—without courts, contracts, currency, police, regulatory frameworks—there is no market and no meaningful sense in which either any income or property is “yours.” What we have is ours only because of a comprehensive set of public policies and institutions. To speak of the government “taking” our money through taxation is to misunderstand the nature of property itself.

This is the correct philosophical foundation for the convergence of benefit and ability-to-pay. If the entire system of property and exchange is constituted by the institutional order, then everything an individual accumulates above the baseline of no institutional order—Hobbes’ state of nature—is, in a meaningful sense, enabled by that order. The benefit of government is not some discrete service that can be individually priced—it is the whole framework within which economic life occurs.

Murphy and Nagel came remarkably close to seeing this. Unlike the economists who dismissed the benefit principle’s measurement problem as insoluble, Murphy and Nagel identified the right baseline and the right proxy. “The benefit of government services,” they wrote, “must be understood as the difference between someone’s level of welfare in a non-government world and their welfare with a government in place.” They even proposed that “we can use people’s actual levels of welfare, with government in place, as a rough measure of the benefit conveyed to them by government.”⁵ This is the Hobbesian baseline, correctly identified.

But they concluded that the benefit principle was nonetheless unworkable as a guide to tax design, because translating measured benefit into a fair tax schedule requires knowledge of how the marginal utility of income declines—“how steeply marginal utility of income declines, and of how much the rate of decline varies from person to person.”⁶ Since marginal utility is an unobservable psychological quantity that varies across individuals, the principle could not be implemented. They raised additional objections as well—that the benefit principle provides no guidance on the level of government expenditure, that it does not determine whether taxes should be progressive or regressive, and that strict application would prevent government from aiding the indigent.⁷ But the decisive obstacle was the measurement problem: the benefit principle foundered on the impossibility of interpersonal utility comparisons.

The irony is that this impossibility is a problem with the utilitarian framework Murphy and Nagel inherited, not with the benefit principle itself. The convergence of benefit and ability-to-pay does not require knowledge of anyone’s marginal utility. It requires only the observation that each person’s economic position above the Hobbesian baseline—their actual outcome within the institutional order—is simultaneously the measure of benefit received and the measure of ability to pay. The marginal utility problem arises only if one insists on translating benefit into subjective welfare before constructing a tax schedule. If instead one measures benefit by the observable proxy that Murphy and Nagel themselves identified—actual welfare levels within the institutional order—the obstacle they found decisive dissolves.⁸

Seligman, Fiscal Citizenship, and the Rejection of Benefit

The Progressive Era economists who built the case for the income tax had a different objection to the Benefit Principle—and in some ways a deeper one than the measurement problem.

Edwin Seligman championed what he called the “faculty theory”—his term for ability-to-pay—while allies like Richard Ely and Henry Carter Adams mounted parallel arguments for progressive taxation grounded in a positive conception of the state.⁹¹⁰ All three were trained in the German Historical School and co-founded the American Economic Association in 1885. But while Adams and Ely moved on to other pursuits, Seligman devoted his entire career to public finance and became what Mehrotra calls “the Dean of American taxation.”¹¹ As the legal historian Ajay K. Mehrotra has documented in Making the Modern American Fiscal State (2013), Seligman contended that the benefit theory supported an outmoded, individualistic theory of citizenship—one that reduced the relationship between citizen and state to a commercial transaction. “You pay for what you get” might describe a visit to the grocer; it should not describe the obligations of a citizen’s membership in a political community.

Seligman’s theory of taxation was not individualistic. He conceived of the individual as a member of society, which led to what Mehrotra calls a new concept of “fiscal citizenship.” This was not merely an economic argument—it was a moral one. Seligman, deeply shaped by Felix Adler’s Ethical Culture movement, spoke of developing “the feeling of civic obligation” and believed that taxation should express something about what it means to belong to a political community. The Progressive economists championed ability-to-pay precisely because it promoted an active role for the state and supported what Mehrotra describes as “the reconfiguration of civic identity”—goals the transactional benefit principle could not serve.

Their concern was legitimate. If taxation is merely payment for services rendered, then once you have paid your bill, you have discharged your obligation. You owe nothing further to the community. The benefit principle, as then understood, could not sustain a vision of shared civic responsibility.

But there was an unintended cost to Seligman’s strategy. By discarding the benefit principle entirely in order to escape its transactional implications, Seligman severed the link between what government does and how citizens’ fiscal obligations are distributed—between taxation and the purpose of government. This severance is precisely what Musgrave later formalized with his separate budget branches, and it created the intellectual conditions under which taxation and government expenditure could be discussed as entirely separate topics.

The consequences of that severance have been profound. Ability-to-pay, stripped of any connection to what government provides, becomes an essentially amoral principle: we take from you because you have what we need and because we can. There is no account of why the community may properly distribute fiscal burdens, no story about what government does that justifies the claim. The taxpayer is positioned as the victim; the government as the extracting agent. Every dollar taxed is a dollar confiscated.

This is precisely the rhetoric that dominates American political discourse today: “It’s your money and the government uses coercive power to taking it.” “Taxation is theft!” The adversarial framing—citizen versus state, producer versus parasite—is the opposite of what Seligman wanted. He wanted citizens and government engaged in a shared civic project. But by severing taxation from government’s function, he inadvertently provided the raw material for the adversarial framing he was trying to escape.

The consequences extended beyond rhetoric. Once ability-to-pay was severed from any account of what government does, the remaining questions about taxation became purely technical. How should rates be structured to minimize economic distortion? What is the optimal trade-off between equity and efficiency? How do taxpayers respond to marginal rate changes? These are important questions, but they are engineering questions, not moral ones. They can be answered with mathematics, econometrics, and behavioral models. They do not require—and have no use for—a theory of fiscal citizenship, an organic conception of social relations, or an account of the duties that flow from membership in a political community.

Richard Musgrave’s formalization of the separation made this explicit. His allocation branch became the domain of welfare economics and cost-benefit analysis; his distribution branch became the domain of optimal tax theory, culminating in the work of James Mirrlees, Peter Diamond, and Emmanuel Saez. These are sophisticated and valuable contributions. But they operate entirely within a technical framework that has no place for the moral vocabulary that Seligman, Ely, and Adams spoke. The question “what does justice in taxation require?” was replaced by the question “what rate schedule maximizes social welfare subject to incentive constraints?” The ethical architects of the income tax had designed themselves out of the conversation.

Mirrlees and the Completion of the Displacement

The displacement of moral reasoning by technical reasoning did not happen all at once. Mill introduced equal sacrifice as a formal criterion in 1848, which was already a step toward mathematizing a moral intuition. Edgeworth and Pigou then developed the utilitarian framework of diminishing marginal utility, turning sacrifice theory into a calculus problem. Musgrave’s separation into budget branches in 1959 created the institutional architecture within which technical questions could be pursued independently of moral ones. But it was James Mirrlees’s 1971 paper, “An Exploration in the Theory of Optimum Income Taxation,” that completed the displacement and made the moral vocabulary formally irrelevant.¹²

Mirrlees reframed the question of tax design as a constrained optimization problem. In his framework, the government seeks to redistribute income to maximize some social welfare function, but it cannot directly observe each person’s innate ability—it can only observe income, which is a product of ability and effort. If high earners are taxed heavily, some will choose to work less, producing deadweight loss. The optimal tax schedule is the one that balances gains from redistribution against losses from reduced effort. After Mirrlees, one could derive an entire tax schedule from first principles without ever mentioning justice, fairness, civic obligation, or benefit.

Two features of the Mirrleesian framework matter enormously for the argument of this essay. First, ability is treated as exogenous—innate, endowed, existing prior to and independent of government. The institutional order plays no role in determining anyone’s productive capacity. Government enters the picture only as a redistributor, not as an enabler. Second, the theoretical ideal toward which the framework aspires is endowment taxation: the optimal tax would be levied directly on the income a person is capable of producing, regardless of what they actually earn. The income tax, in this framework, is merely a second-best approximation of the endowment tax the government would impose if it could observe ability directly.

Endowment taxation—taxing people based on their capacity to produce rather than their actual achievement within the institutional order—raises moral objections that go beyond its practical impossibility. Legal scholars have identified serious liberty concerns: John Rawls objected that an endowment tax “would force the more able into those occupations in which earnings were high enough for them to pay off the tax.”¹³ The objection is commonly illustrated by the hypothetical “beachcomber” who opts for leisure over income, but the problem is sharper with more realistic examples. Consider a surgeon who chooses to work for Doctors Without Borders at a fraction of her potential private-practice income. Under endowment taxation, she would be taxed on what she could earn, not what she actually earns. Rawls’s objection is thus not only that people might be forced into occupations they do not prefer—it is also that they would be forced to choose employers and assignments that maximize income rather than social contribution. The tax system would punish her for choosing to serve.

But the deeper problem is not coercion—it is the conception of the person that the framework embeds. When the state looks at a citizen and asks “what is this person capable of producing, and how much can we extract from that capacity to produce?” it treats the citizen’s abilities as a resource available for collective use. The citizen becomes a productive asset to be optimized. This violates what Kant identified as a fundamental moral principle: “Act so that you treat humanity, whether in your own person or in that of another, always as an end and never merely as a means.”¹⁴ Endowment taxation instrumentalizes the citizen—it treats human capacity as a means to the state’s fiscal ends rather than treating the citizen as an end whom government exists to serve.

There is an irony here worth noting. The Mirrleesian framework is understood by its practitioners as the sophisticated, modern, market-friendly approach to taxation within a liberal, democratic, capitalist order. But its foundational aspiration—to tax individuals based on what they are capable of producing for the collective—resonates uncomfortably with the logic of central planning.¹⁵ The framework asks the same question a central planner asks: “What is this person capable of? How can we best deploy that capacity for the social good?” Government was made for man, not man for government—but endowment taxation reverses that relationship.

The benefit-as-ability framework avoids this entirely. It does not ask “what are you capable of producing for the collective?” It asks “what economic position has the institutional order enabled you to achieve?” The direction of the question is reversed. The first treats the individual as a source of value for society. The second treats society as a source of value for the individual—and then asks the individual to contribute to maintaining that source in proportion to how much they have benefited from it. Under endowment taxation, the anti-government instinct is justified: government really is a predatory force extracting value from citizens’ capacities, and the rational response is resistance. Under benefit-as-ability, the civic instinct is justified: government is the people acting collectively, and taxation is how citizens maintain their own institutional order.

Remarkably, the convergence of benefit and ability-to-pay has recently begun to reassert itself even within the Mirrleesian tradition. The Harvard economist Matthew Weinzierl has shown that once one makes a simple modification to the standard setup—allowing individual income-earning ability to be a function of both endowed talent and public goods—benefit-based reasoning re-enters the framework naturally.¹⁶ If your productive capacity is partly a product of the institutional order, then those who have achieved more within that order have realized greater benefit from public goods, and the benefit principle and ability-to-pay converge. The government does not ask what you are capable of—only what you have actually achieved within the institutional order. A surgeon who chooses to work for Doctors Without Borders at a fraction of what she could earn in private practice has high ability but a modest economic position, and under benefit-as-ability owes proportionally less—not because she has used less of the institutional ecology, but because she has converted less of its value into personal economic benefit. Weinzierl identifies this as the classical idea of “benefit-as-ability” and observes that it “was not further explored, while benefit-based and ability-based reasoning were developed as separate ideas.” This essay attempts to explain why that exploration was abandoned—and to join Weinzierl in resuming it.

Weinzierl’s finding means that we need not reject Mirrlees’s mathematics—only what motivated it. The calculations that emerge from a model where ability depends on public goods produce the same tax schedules as those that emerge from the convergence of benefit and ability-to-pay—in a world where everyone achieves according to their full endowed capacity. But the two frameworks differ profoundly in what they ask of the taxing authority and of the state. Under endowment taxation, the government must somehow observe each citizen’s innate capacity and extract accordingly. Under benefit-as-ability, the government need only observe what each citizen has actually achieved—and it bears a responsibility for having enabled that achievement. What changes is not the calculation but the justification—and as we shall see, the justification matters profoundly for the kind of citizens the tax system produces.

As Weinzierl points out, Franklin Roosevelt understood the convergence as governing policy. In 1935, he stated: “With the enactment of the Income Tax Law of 1913, the Federal Government began to apply effectively the widely accepted principle that taxes should be levied in proportion to ability to pay and in proportion to the benefits received. Income was wisely chosen as the measure of benefits and of ability to pay.”¹⁷ Barack Obama echoed the same intuition in 2011: “It’s a basic reflection of our belief that those who’ve benefited most from our way of life can afford to give back a little bit more.”¹⁸ As Weinzierl observes, “Modern tax theorists will find the normative arguments underlying these quotations both familiar and strange.” Familiar because the intuition is sound; strange because the academic tradition abandoned them.

This is why the Progressive Era thinkers have disappeared from contemporary tax discourse. Not because their scholarship was forgotten or their contributions were unimportant, but because the conceptual framework they helped create rendered their deepest concerns irrelevant to the ongoing discussion. Today, Seligman, Ely, and Adams are of interest to historians of the Progressive Era, not to the economists and philosophers who debate tax policy and tax design. The principles they championed—the principles that founded the American income tax—no longer speak to the terms in which taxation is discussed. The false rivalry they helped construct between benefit and ability-to-pay displaced the moral foundations of taxation with a technical apparatus that no longer needs moral foundations at all.

But, acceptance of the unified principle—benefit and ability-to-pay, understood as the same thing—would actually support the collaborative relationship that Seligman envisioned. Citizens and government need not be seen as adversaries. They are partners in a joint project: sustaining and improving the institutional order that makes economic life possible, with each contributing in proportion to how much that order has enabled them to achieve.

The “Protection” and Accounting Problems

The common thread linking most of these missed connections—from Smith through Lindahl, Musgrave, and Seligman—is a too-narrow definition of what government provides.

When Musgrave speaks of “protection received,” when Lindahl proposes pricing individual government services through preference-revealing auctions, when critics argue that the poor benefit more from welfare than the rich—they are all operating with the same minimalist, service-by-service conception of government benefit. Government is imagined as a provider of discrete, individually consumed services: roads, courts, police, defense, schools. The question becomes: how much of each service does each citizen consume? Murphy and Nagel, to their credit, escaped this atomistic framing—they identified the Hobbesian baseline as the right reference point and actual welfare as the right proxy. But even they stumbled on the next step: how to translate aggregate benefit into a specific tax schedule without interpersonal utility comparisons.

This narrow framing produces two fatal consequences for those who remain within it. First, it makes benefit appear unmeasurable. How do you determine one person’s share of national defense, or the value to a specific individual of the court system’s existence? The measurement problem that sank Lindahl’s approach is an artifact of this atomistic conception of benefit.

Second, and more fundamentally, the “protection” framing implies that government’s function is merely protective—standing guard over wealth that individuals generate autonomously through their own effort and talent in the marketplace. This is, at bottom, the libertarian premise: the market creates, government protects. Under this framing, the rich benefit “more” only in the trivial sense that a security guard is more valuable to the person with more property to watch over. The convergence of benefit and ability-to-pay cannot work under this conception because “protection” is too thin a concept to bear the weight.

But what if government’s function is not merely protective? Abraham Lincoln offered a far more expansive definition in his “Fragment on Government” (c. 1854):

“The legitimate object of government, is to do for a community of people, whatever they need to have done, but can not do, at all, or can not, so well do, for themselves—in their separate, and individual capacities.”¹⁹

Under Lincoln’s principle, the function of government is not protection but enablement—creating the conditions under which individuals can accomplish together what they could never accomplish alone. If that is government’s function, then the relevant benefit is not a bundle of discrete services consumed but the entire institutional and regulatory ecology that makes productive life possible. And this changes everything about how we think about the Benefit Principle.

The Hobbesian Correction: Government as Institutional Ecology

Thomas Hobbes understood something that the “protection” framing misses entirely.

In Leviathan (1651), Hobbes described the state of nature—life without government, without the institutional order that constitutes political society. In that condition, there are no property rights, no enforceable contracts, no markets, no currency, no corporations, no banks, no insurance, no patents, no regulated professions, no standardized weights and measures. There is, in Hobbes’s memorable phrase, “no place for Industry; because the fruit thereof is uncertain.” Economic life above bare subsistence is impossible. Crucially, Hobbes argued that in this condition all people are essentially equal—not in talent or strength, but in their inability to achieve security or accumulate wealth. Even the strongest can be overcome by the cunning or by combinations of the weak; no one can reliably hold what they have. The state of nature is a condition of rough equality precisely because it is a condition of universal precariousness. It is the institutional order that makes possible both greater accumulation for all and inequality in outcomes, by creating the conditions under which differences in talent, effort, and luck can compound into vast differences in economic position.

The institutional order does not protect preexisting wealth. It creates the conditions under which wealth can be generated at all. The relevant benefit of government is not a bundle of individually priceable services but the entire institutional ecology: currency and banking regulation, contract law and commercial codes, transportation infrastructure, education systems, public health measures, scientific research, telecommunications standards, bankruptcy law, environmental management, intellectual property regimes, and the thousand other institutions without which a complex modern economy could not function.

Consider a concrete example. A skilled surgeon in the United States might earn several hundred thousand dollars per year. That same surgeon, with the same hands, the same training, and the same work ethic, transported to a failed state without functioning institutions, would earn little to nothing as a surgeon—because there would be no hospitals, no medical supply chains, no malpractice insurance, no credentialing system, no patients with the means to pay. The surgeon’s earning capacity is not a product of individual talent alone — but neither is it a product of the institutional ecology alone. It is the result of individual talent, effort, and skill operating within an institutional ecology that makes that talent economically productive. The surgeon’s years of training, discipline, and expertise are real and deserve recognition. The point is not that institutions replace individual effort, but that individual effort without institutions has no market in which to be rewarded.

This is true of everyone, but it is more true of those with the highest economic position. The subsistence farmer depends relatively little on the institutional ecology; the hedge fund manager depends on it entirely. Without the dense web of financial regulation, contract enforcement, currency stability, telecommunications infrastructure, and legal protections that constitute modern capital markets, the hedge fund manager’s considerable analytical skills would have no market in which to generate returns. You may be poor in any country, but to be a billionaire, you must have the support of powerful government institutions.

The Hobbesian baseline—the state of nature, where everyone’s capacity to generate economic surplus is approximately zero, and equal in that precariousness—provides the reference point that makes measurement possible.

The Convergence Restored

Once the benefit of government is defined as the institutional conditions that enable economic achievement above the Hobbesian baseline, the measurement problem vanishes.

We do not need Lindahl’s conceptual auction, because the real economy has already produced the answer. Each person’s economic position above the baseline—their accumulated capacity to generate and command resources—is the most meaningful measure of the benefit received from the institutional order. The institutional ecology enabled it; without that ecology, it would not exist. And that same economic position—whether derived from labor, investment, inheritance, or luck—is, obviously, the measure of ability to pay.

Of course, different individuals achieve different economic positions within the institutional order, and those differences reflect — in part — real differences in effort, talent, and risk-taking. But they also reflect different degrees of benefit from the institutional ecology: the order that enabled a $50,000 income and the order that enabled a $50,000,000 income are the same order, but the second person’s economic life undoubtedly depends on more of its features, and depends on them more intensely. The measure of benefit enjoyed is not a denial of individual effort — it is a recognition that effort produces greater rewards when it operates within a richer institutional context.

The two principles are not merely correlated, not approximately aligned, not usefully similar. They converge on the same underlying base, even if imperfectly proxied. The benefit the institutional order has conferred upon an individual is measured by that individual’s economic position above the baseline of no institutional order. The individual’s ability to pay is measured by the same thing. They are two descriptions of a single reality, just as Smith originally implied with his “that is.”

No preference revelation is needed. No individual pricing of government services is required. No conceptual auction must be conducted. The market economy, operating within the institutional framework that government sustains, has already done the measuring. The result is observable. The convergence is not a theoretical conjecture—it is a more accurate description of the relationship between the two concepts.

What This Implies

If the Benefit Principle and Ability-to-Pay are genuinely the same thing, several consequences follow.

First, progressive taxation is not merely compassionate, not merely pragmatic, and not merely a concession to political necessity—and, as Obama said, it is “not because we begrudge those who’ve done well—we rightly celebrate their success.”²⁰ It is a direct reflection of the fact that the institutional order has enabled greater economic outcomes for some than for others. Those who have achieved the most within that order have received the most from the institutional ecology, and their greater contribution is not charity or sacrifice—it is proportional participation in the joint project that made their achievement possible. This does not deny that those who have achieved the most have often worked hard, taken risks, and exercised exceptional skill. It does say that their economic position was jointly produced — by their effort and by the institutional order within which that effort was rewarded. Their greater contribution to maintaining that order is not charity or sacrifice—it is proportional participation in the joint project that made their achievement possible.

Second, the age-old debate between these two principles has been a distraction—consuming intellectual energy on a false dichotomy while the real questions went unasked.

Third, and perhaps most importantly, the convergence transforms how we understand the relationship between citizens and government. They need not be adversaries in a zero-sum contest over resources. They are partners in a joint project of sustaining the institutional order that enables everyone’s economic life. Under the functional finance framework developed by Abba Lerner, a currency-issuing government does not need tax revenue in order to spend—it issues the currency.²¹ Taxation serves to manage aggregate demand, maintain the currency’s value, and ensure that the community’s stake in the institutional order is properly recognized. The adversarial framing—“the government is taking your money”—is not just politically toxic. It is economically incoherent. The benefit principle, under functional finance, does not answer the question “how much should each person pay for government services?” It answers the question “how should the burden of maintaining price stability be distributed?”

Fourth, the convergence reveals something about the moral-educational function of the tax system that is almost entirely absent from the academic literature. It is often said that the tax code is a moral document. This is true in two distinct ways: its impacts have moral consequences—who bears the burden, who is helped, who is harmed—and its justification teaches moral lessons about the relationship between individual and community. The second dimension may be more consequential than the first. A tax system that produces a mathematically optimal distribution while teaching citizens that government is a predatory force extracting value from their endowments will eventually undermine the political conditions necessary to sustain that distribution. The moral lesson corrodes the moral outcome.

Motives matter, even when outcomes are identical. A tax justified as extraction from the capable teaches citizens that their abilities are collective property and that the state has a prior claim on their productive capacity. A tax justified as proportional contribution to the institutional order that enabled one’s achievement teaches citizens that they are participants in a shared project whose maintenance benefits everyone. Over time, these different lessons produce different political cultures—one that resists taxation as an imposition and one that accepts it as an expression of what citizenship means.

The contrast between two presidential visions makes the stakes vivid. Abraham Lincoln spoke of “government of the people, by the people, for the people”—a formulation in which government is the people, acting collectively.²² Ronald Reagan declared in his first inaugural that “government is not the solution to our problem; government is the problem”—a formulation in which government is an alien force imposed upon the people.²³ These are not merely rhetorical differences. They reflect fundamentally different theories of the state. Under Lincoln’s conception, taxation is the means by which citizens share the burden of their own institutional order—an act of collective self-governance. Under Reagan’s, taxation is extraction by an external power. The convergence of benefit and ability-to-pay makes sense only under Lincoln’s conception. And it is Lincoln’s conception that the evidence supports. The benefit is the ability-to-pay.

The convergence of benefit and ability-to-pay points toward a healthier understanding: citizens contribute to the maintenance of the institutional ecology in proportion to how much that ecology has enabled them to achieve. Not because the government compels sacrifice, but because participation in the joint project is what citizenship means. This is what Seligman was reaching for when he spoke of “the feeling of civic obligation.” He was right about what he wanted. The path he chose to get there—discarding the benefit principle—led somewhere else entirely. Perhaps it is time to find our way back.

What remains, then, are the further questions that the convergence opens but does not answer. What is the proper measure of the economic position that the institutional order enables? How should the tax code recognize it? And what obligations does the community have toward those whom the institutional order has not yet adequately served? These are questions for another day. But they can only be properly asked once we have set aside the false rivalry and recognized that the two great principles of tax justice were always, as Adam Smith quietly implied, the same idea.

1. Fairness and Tax Policy. 2015. JCX-48-15. Joint Committee on Taxation. https://www.jct.gov/publications/2015/jcx-48-15/.↩︎

2. Holmes reportedly made this statement in a 1904 speech and wrote a version of it in his 1927 dissenting opinion in Compañía General de Tabacos de Filipinas v. Collector of Internal Revenue, 275 U.S. 87.↩︎

3. Lindahl’s argument appears in Die Gerechtigkeit der Besteuerung (1919), translated as “Just Taxation: A Positive Solution” in R. A. Musgrave and A. T. Peacock, eds., Classics in the Theory of Public Finance (London: Macmillan, 1958). The paraphrase cited here—that Lindahl “argued that there did not necessarily exist any contradiction between the principles of benefit and ability to pay because ability to pay could often be taken as a good indication of the benefit derived from public expenditure”—is drawn from the overview of the benefit principle in the Handbook of Public Economics (ScienceDirect), which also notes that “on this point, Lindahl’s argument is reminiscent of Adam Smith’s first maxim of taxation.”↩︎

4. Linda Sugin, “A Philosophical Objection to the Optimal Tax Model,” 64 Tax Law Review 229, 238 (2011).↩︎

5. Liam Murphy and Thomas Nagel, The Myth of Ownership: Taxes and Justice (Oxford University Press, 2002), ch. 2.↩︎

6. Murphy and Nagel, The Myth of Ownership, ch. 2.↩︎

7. These objections are addressed in other essays in this series. Briefly: the benefit principle, properly understood, implies constraints on expenditure (Lincoln’s principle) and inherently supports progressive taxation (because the untaxed Hobbesian baseline creates effective progressivity). As for the supposed incompatibility with welfare provision, Murphy and Nagel argue that “the benefit principle is, in fact, incompatible …with every account of social justice that requires government to provide any kind of income support or welfare provision whatsoever to the destitute.” But this objection dissolves under the institutional ecology framing. If government’s proper function is to create the conditions under which all citizens can flourish, then the need for welfare provision is not a benefit conferred on the poor—it is evidence that the institutional order has failed to achieve its core purpose. Programs that address destitution are remedial measures, not the system’s primary product. They represent the government’s obligation to correct its own shortcomings, not a gift that recipients should be expected to pay for.↩︎

8. The question of how to translate observable economic position into a specific rate structure without relying on interpersonal utility comparisons is addressed in a companion work on Surplus Capacity Theory, which substitutes the observable marginal propensity to consume for the unobservable marginal utility of income.↩︎

9. Ely rejected the benefit principle’s transactional framing in explicitly civic terms: “The citizen pays because he is a citizen, and it is his duty as a citizen to do so. It is one of the consequences which flow from the fact that he is a member of organized society…. Only an anarchist can take any other view.” Richard T. Ely, Taxation in American States and Cities (New York: T. Y. Crowell, 1888), 13. For Ely, taxes were “one-sided transfers” and “the element of reciprocity is excluded”—a direct rejection of the benefit principle’s commercial logic.↩︎

10. Adams emphasized the collective rather than individual character of economic life, arguing that “the science of finance is par excellence the branch of economic science which lays stress upon the organic or collective conception of social relations.” Henry Carter Adams to Edwin R. A. Seligman, December 26, 1907, quoted in Joseph Dorfman, “The Seligman Correspondence II,” 56 Political Science Quarterly 270, 275 (1941). See also Adams, The Science of Finance: An Investigation of Public Expenditures and Public Revenue (New York: H. Holt & Co., 1899), 46–49.↩︎

11. Ajay K. Mehrotra, “Edwin R. A. Seligman and the Beginnings of the U.S. Income Tax,” Tax Notes, November 15, 2005. See also Mehrotra, “Envisioning the Modern American Fiscal State: Progressive-Era Economists and the Intellectual Foundations of the U.S. Income Tax,” 52 UCLA Law Review 1793 (2005).↩︎

12. James A. Mirrlees, “An Exploration in the Theory of Optimum Income Taxation,” Review of Economic Studies 38, no. 2 (1971): 175–208.↩︎

13. John Rawls, “Some Reasons for the Maximin Criterion,” American Economic Review 64, no. 2 (1974): 141–146, at 145. See also Kirk J. Stark, “Enslaving the Beachcomber: Some Thoughts on the Liberty Objections to Endowment Taxation,” 18 Canadian Journal of Law & Jurisprudence 47 (2005); Linda Sugin, “A Philosophical Objection to the Optimal Tax Model,” 64 Tax Law Review 229 (2011); David Hasen, “Liberalism and Ability Taxation,” 85 Texas Law Review 1057 (2007).↩︎

14. Immanuel Kant, Groundwork of the Metaphysics of Morals (1785), second formulation of the categorical imperative.↩︎

15. The resemblance to Marx’s principle is difficult to ignore: “From each according to his ability, to each according to his needs.” Karl Marx, Critique of the Gotha Programme (1875). The Mirrleesian framework operationalizes the first half of this formula within a liberal capitalist order.↩︎

16. Matthew Weinzierl, “Revisiting the Classical View of Benefit-Based Taxation,” Economic Journal 128, no. 612 (2018): F37–F64.↩︎

17. Franklin D. Roosevelt, Message to Congress on Tax Revision, June 19, 1935.↩︎

18. Barack Obama, Remarks on Fiscal Policy, George Washington University, April 13, 2011.↩︎

19. Abraham Lincoln, “Fragment on Government” (c. July 1, 1854), in Collected Works of Abraham Lincoln, ed. Roy P. Basler (New Brunswick, NJ: Rutgers University Press, 1953), vol. 2, 220–221. I explore the implications of Lincoln’s principle for understanding taxation as a service rather than an extraction in a companion essay, “Taxation Is a Service.”↩︎

20. Obama, Remarks on Fiscal Policy, April 13, 2011, quoted above.↩︎

21. Abba P. Lerner, “Functional Finance and the Federal Debt,” Social Research 10, no. 1 (1943): 38–51.↩︎

22. Abraham Lincoln, Gettysburg Address, November 19, 1863.↩︎

23. Ronald Reagan, First Inaugural Address, January 20, 1981.↩︎

What Capitalism Actually Is

Capitalism is not simply “markets.” Markets predate capitalism by millennia. Bazaars in ancient Mesopotamia, trading networks across the Roman Empire, merchant guilds in medieval Europe—none of these were capitalist in any meaningful sense. Capitalism is not simply “private property” either. Feudal lords held vast private estates. What, then, is distinctive about capitalism?

Standard definitions tend to assemble catalogs of observed attributes—markets, wage labor, private property, profit-seeking firms—but such catalog-like definitions produce analytical drift. Nearly all modern economies satisfy some subset of these attributes. No consensus exists on which are essential characteristics. The meaning of “capitalism” shifts with whoever is defining it, and disagreement reduces to taxonomic contestation rather than substantive inquiry. Worse, such definitions cannot distinguish capitalism from the feudal, mercantilist, and oligarchic systems that share many of the same surface attributes—which is precisely the distinction that matters.

A more precise definition, focused on the essence of capitalism’s institutional logic rather than the surface traits of systems called “capitalist”, restores analytical clarity:.

Capitalism is an institutional order characterized by the generalized legal and practical capacity of individuals to accumulate, retain, and deploy economically consequential surplus.

This definition isolates what is genuinely distinctive about capitalism. Each term does specific work. Generalized means the capacity is structurally available to all, not restricted by birth, caste, or sovereign favor to specific classes. Legal and practical means that formal rights without practical access are insufficient—the institutions must make accumulation genuinely possible, not merely nominally permitted. Accumulate captures the creation of surplus beyond subsistence. Retain captures the protection of that surplus from confiscation, inflation, or predatory extraction. And deploy captures the ability to put surplus to productive use—turning wealth into capital through investment, enterprise, or education.

Strictly speaking, “capital” means wealth deployed in production—a pile of gold hidden in a vault is wealth but not capital. The definition captures this distinction through “deploy”: capitalism enables not just the accumulation of providential personal financial buffers but also the productive investment of surplus. The capacity to save is the prerequisite for capital formation. You cannot invest what you do not have.

This is what distinguished capitalism from what came before. Under feudalism, serfs had customary use rights to land but could not sell, mortgage, or freely bequeath it. Guild systems restricted who could practice trades. Usury laws constrained credit. The legal and institutional framework was designed to freeze economic relationships, not enable mobility. Most people could not legally accumulate transferable wealth, let alone deploy it productively. Under mercantilism, the state directed economic activity for sovereign benefit. Economic opportunity depended on royal favor, monopoly charters, and guild membership. Accumulation capacity was structurally narrow. Under slavery, labor itself was someone else’s capital. The laborer had no legal right to the fruits of their own work—the most extreme denial of accumulation capacity imaginable.

Capitalism’s institutional revolution consisted of transferable property rights, open credit markets, legal personhood for commoners, and enforceable contracts between equals—the infrastructure that generalized the capacity to accumulate, retain, and deploy surplus. This was genuinely radical. For the first time in human history, the legal and institutional framework was being constructed to enable ordinary people—not just the wellborn, the politically connected, or the militarily powerful—to build wealth and put it to productive use.

No one articulated this promise more clearly than Abraham Lincoln. In his 1859 Address before the Wisconsin State Agricultural Society, Lincoln laid out the free labor vision in terms that map directly onto the definition above. He rejected the “mud sill” theory—the pro-slavery argument that laborers are permanently fixed in their condition, mere instruments of capital owners. Instead, Lincoln described the trajectory that capitalism’s institutions make possible:

The prudent, penniless beginner in the world, labors for wages awhile, saves a surplus with which to buy tools or land, for himself; then labors on his own account another while, and at length hires another new beginner to help him.

Lincoln called this “free labor—the just and generous, and prosperous system, which opens the way for all—gives hope to all, and energy, and progress, and improvement of condition to all.” At New Haven in 1860, he put it more directly: “I take it that it is best for all to leave each man free to acquire property as fast as he can. Some will get wealthy. I don’t believe in a law to prevent a man from getting rich; it would do more harm than good.” And, he wrote earlier: “There is no permanent class of hired laborers amongst us. Twenty-five years ago, I was a hired laborer. The hired laborer of yesterday, labors on his own account to-day; and will hire others to labor for him to-morrow. Advancement—improvement in condition—is the order of things in a society of equals.”

Lincoln also insisted that free labor required universal education—that “heads and hands should cooperate”—and that educated labor, applied to productive land, could make any community “independent of crowned kings, money kings, and land kings.”

Lincoln went further, expressing an aspiration: “If any continue through life in the condition of the hired laborer, it is not the fault of the system, but because of either a dependent nature which prefers it, or improvidence, folly, or singular misfortune.” This is best understood not as a description of the world Lincoln lived in—structural barriers were very real in 1859, and they remain so today—but as a statement of what capitalism should achieve. A world in which failure to flourish is genuinely attributable to individual choice rather than structural barriers would be a world in which capitalism’s institutional promise had been fully realized. We are far from that world. But Lincoln’s vision is the right one, and it demands, rather than excuses, the institutional infrastructure necessary to make it real. Those who invoke individual responsibility while dismantling the institutions that make individual agency possible are betraying Lincoln’s vision and thwarting capitalism, not fulfilling its promise.

Lincoln’s “prudent, penniless beginner” is a perfect illustration of the definition: a person with the generalized capacity to accumulate (save a surplus), retain (keep it securely), and deploy (buy tools or land)—and a system whose institutions make that trajectory available to all, not just the wellborn. Capital accumulation by ordinary people is the revolutionary promise of capitalism, and Lincoln articulated it as clearly as anyone.

Capitalism Requires Institutions—But Not Unlimited Government

The generalized accumulation capacity described above does not happen naturally. It requires institutional infrastructure. Property rights require enforcement through courts, registries, and police. Contracts require adjudication. Currency requires a sovereign guarantor. Competition requires antitrust enforcement—markets left alone tend toward monopoly, not competition. Broad-based capital accumulation requires education, transportation infrastructure, financial regulation, consumer protection, and labor standards. Karl Polanyi argued that market society was politically constructed and continuously sustained by state action; markets and states are interdependent, not opposing forces. And Dani Rodrik observed, “Capitalism is not self-creating, self-sustaining, self-regulating, or self-stabilizing.” The more effective the capitalism you want, the more sophisticated the institutions you need.

In a modern fiat-currency economy, taxation itself is part of the institutional infrastructure of capitalism. It is an essential tool for maintaining stable prices and managing aggregate demand—a vital public service without which the monetary system that enables both market transactions and the accumulation of wealth would collapse.

Price stability deserves special emphasis. Inflation is the silent destroyer of the capacity to accumulate. A worker who saves diligently for decades can have that surplus wiped out in a moment by monetary instability. Price stability—the protection of the retain function in our definition—is something that individuals absolutely cannot provide for themselves. Like roads and courts, it is a collective action problem requiring institutional coordination through central banking, fiscal policy, and taxation. No individual or private entity can substitute for these functions. Even the most ardent libertarian does not argue that individuals can independently control inflation. And yet without solving this collective problem, the entire project of individual savings and capital accumulation is undermined. The person who says “I don’t need government, I just want to save and invest” is depending on a government function for the very possibility of what they claim to do independently.

But this is not an argument for unlimited government. Lincoln understood both halves of the equation: both the necessarily broad scope of government as well as the proper limits on government. In 1854, he wrote:

The legitimate object of government, is to do for a community of people, whatever they need to have done, but can not do, at all, or can not, so well do, for themselves—in their separate, and individual capacities. In all that the people can individually do as well for themselves, government ought not to interfere.

Lincoln’s framework is precise. Government should do what individuals cannot do for themselves—and no more. But the scope of what individuals cannot do well for themselves is substantial. Lincoln enumerated: “public roads and highways, public schools, charities, pauperism, orphanages, estates of the deceased, and machinery of government itself.” And critically, Lincoln says “can not, so well do”—not “can not do at all.” Government should act not only when private action is impossible but when private action is inadequate. This is a much broader mandate than libertarians typically admit.

Even contract enforcement and the prevention of wrongs require collective institutional action. Lincoln’s quiet conclusion is devastating for the anti-institutionalists: “if all men were just, there still would be some, though not so much, need of government.” Even in a world of perfect virtue, institutions are needed—not just to prevent wickedness, but to coordinate collective action that individuals cannot accomplish alone.

Lincoln thus understood both the promise and the prerequisites of capitalism. The 1859 speech describes the goal: a system that “opens the way for all.” The 1854 fragment describes the means: government that does for the people what they cannot do well for themselves. The 1859 aspiration—that failure should be individual, not structural—describes the end state that adequate institutions should produce. Together they define a capitalism that requires robust institutions but resists overreach: government adequate to enable broad-based flourishing, restrained from doing what individuals can manage on their own.

The question was never whether government should exist, but whether it should serve all people or only the powerful. Every item on Lincoln’s list—roads, schools, contract enforcement, the machinery of government—is part of the institutional infrastructure of capitalism. Price stability belongs on that list too. To dismantle these things in capitalism’s name is to misunderstand what capitalism requires.

The Rise and Capture of Capitalism

When capitalism first emerged, it was genuinely radical—a liberation movement. It promised ordinary people a means to escape from feudal hierarchy, guild monopoly, and mercantilist favoritism. This is what Adam Smith celebrated: not the interests of merchants and manufacturers, whom he deeply distrusted, but the possibility that free competition could break the grip of entrenched privilege. Early capitalism delivered on parts of this promise—expanded mobility, new paths to independence, erosion of hereditary economic caste.

But capitalism emerged within feudal and mercantilist societies. It did not replace them cleanly. Feudal power structures, aristocratic privilege, guild monopolies, colonial extraction—these were never fully dismantled. The new institutional logic of capitalism—open markets, transferable property, broad-based accumulation—was layered on top of old power structures, and the old power interests adapted, co-opted, and reshaped the new institutions to serve incumbents.

Some of the new institutional mechanisms—particularly those enabling concentrated industrial wealth—actually reinforced the old hierarchical patterns. The factory owner, running a company town, became the new lord of the manor, with the same practical power over workers’ lives that the feudal lord had once exercised, but without the feudal lord’s customary obligations to his vassals. Where once guilds had enabled tradesmen to organize, the new institutions enabled the suppression of labor unions and thus greatly empowered the elites. By Marx’s time, the system operating under the name “capitalism” had never fully de-feudalized—and had, in important ways, given new force to the very hierarchies that real capitalism would have dismantled. The institutional revolution was in-progress, but incomplete—and the state lacked both the administrative capacity and the democratic mandate to complete it. There was no central banking, no securities regulation, no antitrust law, no income tax. The franchise was restricted to property owners. The very people who most needed capitalism’s institutions had no political voice with which to demand them. New mechanisms of wealth creation were grafted onto old structures of power and privilege, and the old interests—well-represented in the parliaments that excluded workers—proved adept at capturing the new mechanisms for their own essentially anti-capitalist purposes.

Marx brilliantly diagnosed the gap between capitalism’s promises and the reality of those systems called “capitalist”—but drew the wrong conclusion about why the gap existed. He saw a system that promised generalized accumulation but instead delivered concentrated wealth, and he concluded that the concentration was inherent to the system’s logic. His pessimism about reform was understandable: the institutional tools that might have made capitalism’s promise real—central banking, securities regulation, antitrust law, progressive income taxation—barely existed in his time, and some had not yet been imagined. The state that would have wielded such tools was structurally captured by the very interests that benefited from the status quo. The alternative explanation—that the institutions necessary to fulfill capitalism’s promise had never been fully constructed yet could be constructed given sufficient state capacity and democratic accountability—was not so much rejected by Marx as rendered invisible by the world he inhabited.

This misunderstanding has shaped political discourse ever since. Defenders of capitalism celebrate a system that does not fully exist. Critics attack “capitalism” when what they are really describing is capitalism’s capture. The anti-capitalist counter-revolution never announced itself—it called itself capitalism, just as today’s institutional demolitionists call themselves capitalism’s defenders.

The Right’s Error: “Pro-Capitalist” Policies That Undermine Capitalism

Consider the policies most commonly defended in capitalism’s name. Slashing public education reduces human capital formation for ordinary people. Deregulating finance enables predatory extraction rather than productive investment. Weakening antitrust allows monopoly rents that suppress competition. Cutting infrastructure investment reduces the physical substrate of commerce. Opposing labor standards pushes workers below the threshold where capital accumulation is possible. Tax structures that concentrate wealth reverse the broad-based accumulation that defines capitalism. Each of these policies, defended in capitalism’s name, actually moves the economy away from capitalism and toward something older and less democratic.

The deeper problem is that excess income and wealth do not just represent an inequality concern—they actively reduce the state’s capacity to regulate the economic system and preserve capitalism itself. Concentrated wealth captures regulatory institutions, shapes law to protect incumbents, and undermines the very institutional infrastructure that capitalism requires. The concentration of wealth is thus self-reinforcing and self-accelerating unless checked by institutions.

The United Kingdom illustrates this dynamic clearly. The Thatcher reforms of the 1980s, such as the selling of Council Housing to its tenants at a discount, moved the economy toward greater private ownership—but did not broaden accumulation capacity. Quite the opposite: privatization was accompanied by a compression of accumulation breadth that has continued in subsequent decades. Ownership shifted from public to private hands, but the generalized capacity to accumulate, retain, and deploy surplus narrowed rather than widened. The rich got richer, the poor did not. This pattern was repeated in the privatization of Soviet state assets in the 1990s. Although proclaimed as a move to broaden asset ownership, it resulted in a reconcentration of assets in the hands of oligarchs. Privatization does not mechanically imply broadened accumulation. Moving assets from public to private ownership benefits capitalism only if the institutional framework ensures that private ownership is and remains broadly accessible. Otherwise, it merely concentrates existing wealth.

The contrast with the post-war period is instructive. From 1945 to the mid-1970s, the democracies that invested most heavily in capitalism’s institutional infrastructure—progressive taxation, financial regulation, strong labor standards, social insurance, public education—produced the broadest period of shared prosperity in modern history. The Nordic countries have continued to demonstrate this pattern: robust institutional frameworks that produce, rather than impede, broad-based accumulation. If this resembles what is sometimes called social democracy, that is not an objection—it is evidence that social democracies have been, in practice, closer to real capitalism than the systems that more loudly claim the name.

But much of the post-war progress has since been systematically dismantled. Beginning in the late 1970s, the neoliberal turn—deregulation, tax cuts for the wealthy, weakening of unions, erosion of public investment—reversed the broadening of accumulation capacity and reconcentrated wealth, repeating the pattern of Marx’s era: incumbent elites capturing the institutional framework and reshaping it to serve their own interests, all while claiming to defend or strengthen capitalism.

It is often said that “conservatives” are those who seek to preserve and strengthen institutions. They would seem to be the natural defenders of capitalism, since capitalism requires stable and strong institutions. Yet today’s self-described conservatives are leading the charge to dismantle institutions. Lincoln’s own framework condemns this: public roads, public schools, contract enforcement, the machinery of government—these are things people “can not, so well do, for themselves.” Dismantling them is not limiting government to its legitimate object; it is preventing government from fulfilling it.

If Not Capitalism, Then What? A Conceptual Map

To clarify what happens when capitalism’s institutions decay, it helps to distinguish between two fundamental dimensions of any economic system:

1. Ownership of Productive Assets: Are the primary means of production (factories, land, infrastructure) held privately or collectively by the state or community?

2. Breadth of Accumulation Capacity: Is the ability to accumulate, retain, and deploy surplus generalized across the population, or is it restricted to a narrow elite?

As this essay argues, the first dimension distinguishes private-property systems from socialist ones, while the second dimension distinguishes real capitalism from its imposters. By plotting these two independent axes, we can create a conceptual map of different economic regimes (see Figure 1).

This framework reveals that stripping away the institutions that enable broad-based accumulation does not produce “freer” capitalism. Instead, it pushes a society down the vertical axis, from the top-left quadrant (Democratic Capitalism) to the bottom-left (Oligarchic “Capitalism”). This bottom-left quadrant is home to a variety of non-capitalist systems that often claim the capitalist label while violating its core institutional logic:

• Neofeudalism: When hereditary wealth transmission is unchecked, dynastic control of resources hardens, and economic caste becomes self-perpetuating. Wealth and power are determined by birth, with the tech billionaire or the heir to an industrial fortune serving as the new lord of the manor. This is a system of private ownership but narrow, elite-restricted accumulation.

• Crony Capitalism / Corporatism: When regulatory capture allows incumbent firms to write the rules, suppress competition, and extract rents through political connection. The forms of capitalism are preserved, but accumulation capacity is restricted to the politically connected, not generalized.

• Fascism: When concentrated private wealth merges with state power in a nationalist framework, suppressing labor organization and independent institutions while maintaining the facade of private enterprise. It destroys capitalism’s institutional prerequisites for broad accumulation, concentrating power and wealth in a symbiotic elite of state and corporate actors.

All of these systems feature private ownership of productive assets, which leads both their defenders and their critics to mislabel them as “capitalism.” But in each case, the generalized capacity for accumulation has been extinguished in favor of elite-restricted access. They are not capitalism as properly understood. The quadrant helps clarify why: they occupy a different institutional space. The debate we should be having is not just about moving left or right on the ownership axis, but about how to build and maintain the institutions that keep us in the top-left quadrant, realizing the promise of a truly democratic and inclusive capitalism.

The Left’s Error: Accepting the False Definition

Critics who accept that capitalism is exploitation, inequality, and environmental destruction make the mirror-image mistake. By treating these as inherent, definitional features of capitalism—as if they were laws of nature—they foreclose the possibility of reform and progress.

If capitalism necessarily produces these outcomes, the only option is revolutionary overthrow—historically the most costly, dangerous, and unreliable path to social improvement. But if these outcomes result from inadequate institutional development—capitalism without sufficient capitalist institutions—then the better path forward is evolutionary: build better institutions. Evolution is almost always more achievable, less destructive, and more durable than attempted revolution.

The irony is that by accepting the libertarian’s definition of capitalism—unregulated markets, minimal state—the left unwittingly strengthens the case for radicalism over reform. When you tell people the system is inherently rotten, you do not inspire institutional improvement. You inspire either despair or a desire for destructive upheaval.

The left is self-defeating in another way. Its focus on outcomes over institutions means it neglects the very institutional infrastructure needed to produce the outcomes it seeks. If the left learned to love capitalism—properly understood—it could more easily achieve its goals through institutional evolution than through the revolutionary transformation it sometimes dreams of.

Both errors—the right’s and the left’s—serve the interests of those who benefit from the current dysfunctional arrangements.

There Is No Inherent Conflict Between Capitalism and the Social Safety Net

Many programs commonly described as “socialist” are, in fact, neither socialist nor anti-capitalist. This confusion rests on a category error. Capitalism and socialism are not opposite ends of a single axis—they concern different dimensions of institutional design. Capitalism concerns the breadth of accumulation capacity: how widely is the ability to accumulate, retain, and deploy surplus distributed? Socialism concerns ownership structure: are productive assets held privately or collectively? These are analytically independent questions. A society can have broad accumulation capacity and collective risk-sharing simultaneously. Labeling social insurance “socialist” is a category error, not a political analysis.

Social Security, unemployment insurance, public health insurance, public education—none of these restrict individuals’ ability to accumulate capital. They do not collectivize ownership of productive assets. In fact, they enhance accumulation capacity. A worker who is not one illness away from bankruptcy is a worker who can save, invest, and take productive risks. Social safety net programs provide the floor of security from which broad-based capital accumulation becomes possible.

Lincoln’s “prudent, penniless beginner” needed tools and land. Today’s beginner needs different, but analogous, prerequisites. The modern equivalent of “tools and land” is not just physical capital, but human and financial capital: education without crushing debt, healthcare that doesn’t lead to bankruptcy, and access to a stable financial system for savings and credit. The institutional charge remains the same, even if the specific policies required to meet it have evolved.

The Shared Mistake and Why It Persists

Both sides accept the same false equation: capitalism equals whatever economic arrangements currently exist, or existed during industrialization. This is why the debate is so sterile and so frustrating. The right says “capitalism means unregulated markets, and that’s good.” The left says “capitalism means exploitation, and that’s bad.” Both accept the same definition. Both are wrong.

The confusion has deep historical roots. Capitalism and laissez-faire rhetoric emerged together in opposition to mercantilism, and Smith’s critique of mercantilist state intervention has been persistently misread as a critique of all institutional structure. The Cold War hardened this into a binary—capitalism equals less government, socialism equals more government—that obscures rather than illuminates.

Those who benefit from concentrated wealth have obvious incentive to maintain the confusion. If “capitalism” is just a floating signifier attached to whatever economic system currently exists, then defending capitalism means defending the position of those who it has benefited most—whether or not that system benefits others. Meanwhile, critics of the status quo have their own strong incentive to accept the false definition. If capitalism is the problem, the solution is dramatic and morally clear—which is more satisfying than the patient, incremental work of institutional reform.

The ultimate test of any economic system is whether it serves the flourishing of the individuals who live within it. Government is made to serve people, not people to serve government. Capitalism, properly understood, takes individual flourishing as its direct aim—the collective benefit emerges from millions of people successfully building the capacity to lead independent lives. Systems that reverse this priority—subordinating individuals to an abstracted common good or to a powerful elite—historically produce neither the common good nor the individual flourishing they promise.

But “properly understood” is doing essential work in that last paragraph. Individual flourishing does not emerge from unregulated self-interest any more than a garden emerges from untended soil. It requires the institutional infrastructure—education, law, regulation, public investment—that enables self-interest to become productive rather than predatory. The collective benefit is real, but it is a product of institutional design, not of serendipity. Adam Smith’s insight that private interest can serve public benefit was always conditional—dependent on the institutional framework within which that self-interest operates. Those who cite Smith’s “invisible hand” in defense of unregulated markets are making the same error as those who cite Marx in defense of revolutionary overthrow. They extract a conditional claim from its context and treat it as unconditional.

The Debate We Should Have: Institutional Design

Once we clear away the definitional confusion, the productive question emerges: which institutional arrangements best fulfill capitalism’s promise—the generalized capacity to accumulate, retain, and deploy surplus?

The practical path forward is to use markets for what they do well—distributed coordination, incentive alignment, innovation—while building institutions to correct what they do poorly: distributional justice, externalities, public goods, preventing concentration. This is capitalism—real capitalism—not a compromise between capitalism and something else.

Every institutional system requires ongoing maintenance. The gap between principle and practice is permanent and universal—no system has ever fully realized its own ideals, and none ever will. The relevant question is not whether the gap exists—it always does—but whether the principles are worth the cost of their maintenance. We do not abandon democracy because every democracy has corruption. We do not abandon medicine because every doctor sometimes makes errors. We recognize that complex institutional achievements require continuous effort, vigilance, and reform. The same is true of capitalism.

A government committed to capitalism should feel obligated to ensure that all of its citizens can accumulate savings and flourish without undue dependence on either state or patron. The goal is not to eliminate super-normal returns and inequality of outcomes entirely—some are not only inevitable but desirable, since they motivate productivity and innovation. But there is a point at which returns are so excessive that they no longer serve a societal purpose—when they cease to incentivize and begin merely to concentrate.

The precise point at which returns cease to incentivize and begin merely to concentrate is a legitimate and important question, one that requires its own careful analysis. But the difficulty of drawing a precise line does not negate the existence of the problem. We do not refuse to treat fever because body temperature exists on a continuum. We recognize that while 99°F is a minor variance, 105°F indicates a life-threatening crisis requiring immediate intervention. The fact that we cannot specify an exact dollar threshold does not mean we cannot recognize when accumulation has clearly exceeded any incentive function and has begun to distort the institutional framework itself.

Beyond the inequality concern, excess wealth actively undermines the state’s capacity to regulate the economic system and preserve capitalism itself. Concentrated wealth captures institutions, shapes law, and erodes the very infrastructure that broad-based accumulation requires. This is how capitalism is destroyed from within: not by its critics, but by its supposed beneficiaries whose excess destabilizes the institutions that made their accumulation possible.

Institutions should progressively resist further excess—not by confiscating private property, but by calibrating the system to discourage returns beyond their useful social function. A need to resort to confiscation would itself indicate a failure of government and society to preserve capitalism by other, less drastic means. It may sometimes be necessary, but only as a last resort—evidence that earlier institutional safeguards failed.

We have the institutional tools to address excess: progressive taxation, antitrust enforcement, financial regulation, public investment. We should use them and see if they provide a sufficient solution to the problems that have long been identified before concluding that more radical measures are needed.

The question is not “how much government?” but “which institutions, designed how, serving whom?” Or, in Lincoln’s terms: what are the things that people “can not, so well do, for themselves”—and are we doing them?

Closing

Those who genuinely believe in capitalism’s promise—that ordinary people can and should be able to build lives of independence through their own productive effort—should be capitalism’s most vigorous institutional defenders. And those who rightly identify the excesses of our current system should recognize that reform intended to better achieve real capitalism, not revolution, is both the more practical and the more radical path. Evolution is almost always more achievable, less destructive, and more durable than revolution.

Capitalism requires institutions that open the way forward for all. Government’s legitimate object is to do for people what they cannot do well for themselves—and the infrastructure of a just, capitalist economy is precisely such a thing. To defend capitalism is to defend the institutions that make it real.

What would institutionally serious “real capitalism” look like? Future essays will explore the principles that should guide institutional design—including how tax systems can serve, rather than undermine, capitalism’s core commitment to broad-based flourishing.

Americans hate paying taxes. That seems natural, even obvious. Taxes feel like the government taking your money—money you earned, money that belongs to you—to fund its operations. If that’s the way you understand taxes, the natural response isn’t to ask how to design them well. It’s to ask how to minimize them, avoid them, or make them go away entirely.

From that starting point, the debate about taxation naturally becomes one of damage control: how much the government should be allowed to take, how to limit it, and who should pay.

But what if that entire framing is wrong?

What if taxation isn’t primarily the government taking something from you at all—but the government providing a service you benefit from, whether you notice it or not?

This isn’t a semantic trick. It follows from some basic facts about how modern money works—facts economists have understood for decades, even if they rarely spell out the implications in plain language.

Why a Currency‑Issuing Government Taxes

The U.S. government issues its own currency. That means it does not need to “get” dollars from taxpayers before it can spend them. It doesn’t need tax revenue to fund its operations. That’s not why our government taxes.

It can be useful to think about the economy as though it were a bathtub. Government spending turns on the faucet, adding dollars to the economy. Taxes act as the drain, pulling dollars back out.

Neither the point nor the analogy are new. In 1943, economist Abba Lerner argued that for a government like ours, the real economic constraint on spending is not revenue but inflation—the risk of pushing total spending beyond what the economy can actually produce. Three years later, Beardsley Ruml, then chair of the New York Fed, put it even more bluntly: under monetary sovereignty, the primary purpose of taxation is “to stabilize the purchasing power of the dollar.” Today, most economists accept the basic operational fact that a government that issues its own currency can always make payments in that currency. What matters is not whether the government can spend, but what happens after it does: specifically, whether its spending will push total demand beyond the economy's capacity and cause inflation.

That distinction matters. It forces us to rethink what federal taxes are for—and to recognize that under a fiat currency system, taxation’s purpose is fundamentally different than it was under the old gold standard. If taxation’s primary function is to control inflation—to prevent your money from losing value—then taxation isn’t about “taking your money,” it is about providing you with a service. It’s protecting the purchasing power of every dollar you hold.

At this point, a reasonable reader might ask: If this is so obvious, why don’t economists say it this way?

Part of the answer is professional habit. Over time, economists have become careful—sometimes overly so—about separating technical analysis from explicit value judgments. They are comfortable saying that taxation serves to help control inflation, because that can be presented as a descriptive, technical claim about macroeconomic stability.

But economists are much more hesitant to take the next, logically equivalent step and say what that implies: that taxation is a service—something we rely on to protect the value of money and the functioning of the economy. Saying it that way sounds normative. It sounds like a statement about what taxation is for, not just what it does. And that is precisely the move today’s economists are trained to avoid making explicit.

What Taxes Really Do

When the government taxes, it reduces private purchasing power—opening the drain and pulling dollars out of the economy before demand spills over into inflation. Households have less purchasing power. Businesses have fewer dollars chasing workers, inputs, and customers. Demand is reduced.

That’s not a side effect of taxation. It’s the point of taxation. It’s the service that taxation provides—and it’s a really important service. It protects the most fundamental economic tool we all rely on: stable money.

Inflation happens when total demand outruns the economy’s real capacity—when too much money is chasing too few goods and services. Taxes help prevent such overruns by pulling purchasing power out of the economy.

Other tools matter too. Interest rates, regulations, supply constraints, and productivity all play a role in inflation. But taxes are a uniquely direct tool. They work by reducing how much we, collectively, are trying to buy.

Seen this way, taxation is not about “funding” the government. It’s about making the economy function without blowing itself apart.

Inflation Is a Tax—a Very Bad Tax

If the government were to keep spending without withdrawing enough purchasing power from the economy, the result would be inflation. Prices would rise. Each dollar would buy less.

When inflation reduces purchasing power, it functions a lot like a tax—really, two taxes at once. One tax falls on wealth, as cash and fixed savings lose value. The other tax falls on income, as wages and salaries fail to keep up with rising prices. If prices rise ten percent, both your savings and your paycheck buy roughly ten percent less.

But unlike an explicit tax, inflation is badly designed.

Inflation hits hardest those least able to protect themselves: people whose wealth is held mostly in cash or who have fixed incomes, workers whose wages lag behind prices, households living paycheck to paycheck. It is invisible, unlegislated, and indifferent to fairness.

If we care about equity—and policymakers constantly say they do—relying on inflation to do the work that taxes could do more openly, intentionally, and equitably is hard to defend.

Taxes as Insurance or Dental Hygiene

A useful way to think about taxation is as a kind of insurance premium. You don’t buy insurance because you expect your house to burn down tomorrow. You buy it because the risk, if it materializes, is catastrophic—and because paying a predictable premium is far cheaper than facing catastrophic loss.

Taxes work the same way. By giving up some purchasing power in advance, we insure ourselves against runaway inflation later. We pay a visible, structured price to avoid a hidden, chaotic one.

Inflation is what happens when we try to go uninsured.

There’s another way to think about it. Paying taxes is also a bit like going to the dentist. You don’t only go because something has already gone terribly wrong. You go for regular cleanings to prevent problems, to preserve the usefulness of your teeth, and to avoid far more painful and expensive interventions later.

Taxes work the same way. They are routine maintenance for the monetary system—visible, ongoing costs paid to keep money functioning as a reliable tool and store of value, rather than waiting for instability to force far more disruptive corrections.

The Real Question (And Why We’re Not Asking It)

This reframing reveals something important: most of our political debate about taxation is focused on the wrong question.

Turn on cable news or listen to any budget debate in Congress. The question everyone argues about is: “Should taxes be higher or lower?”

Conservatives argue for lower taxes, treating any increase as an encroachment on freedom and a drag on growth. Progressives argue for higher taxes, treating them as necessary to fund vital programs and reduce inequality. Both sides treat the aggregate level of taxation as a fundamental ideological battleground—as if the total amount we tax is primarily a question about political values and the proper role of government.

But once you understand what taxation actually does, this framing starts to look confused. The aggregate level of taxation is not primarily an ideological question; it is a technical one, determined by the macroeconomic conditions we face.

Tax too little, and inflation erodes everyone’s purchasing power—the wealthy and the working poor alike. Tax too much, and you suppress economic activity unnecessarily, creating unemployment. The right amount to tax is simply whatever amount is needed to prevent inflation while maintaining full employment, given current government spending and the economy’s productive capacity.

This doesn’t mean politics disappears. It just means we are arguing over the wrong thing.

The real question is not “How much revenue does the government need?” A currency‑issuing government doesn’t have a revenue requirement in the same way households do. Once the technical level of taxation is determined by the needs of the economy, the genuine political and moral choices begin:

• From whom should that purchasing power be withdrawn?

• How can the “drain” be designed to ensure fairness?

• Whose claims on real resources should give way when the economy is under strain?

These are moral and political questions, and they are deeply contested. They are about power, equity, and the kind of society we want to build. They are also where the technical and moral meet.

A tax on those already struggling to buy groceries is a technically efficient way to reduce total demand, but it is a moral catastrophe. By contrast, because billionaires usually spend a much smaller fraction of their income, a tax on them is a ‘slower’ drain on demand. If the goal of the service is to cool the economy without causing hardship, technical realities suggest that we should tax the top at much higher rates—not as punishment or because of envy, but as sound design. Tax policy isn’t just about deciding how much to tax—it’s about designing a fair mechanism for distributing the cost.

By obsessing over ‘high vs. low’ in the abstract, we avoid these uncomfortable, unavoidable questions about distribution. We treat as an ideological choice what is actually a technical requirement, and in doing so, we fail to have an honest conversation about the service that taxation actually provides.

Taxation Is a Public Service

By deliberately withdrawing purchasing power from the economy, taxes make it possible to sustain high employment and stable prices at the same time. They prevent inflation from becoming the default, unlegislated tax that quietly erodes wealth from the bottom up.

The real choice is not between “high taxes” and “low taxes.” It’s between explicit, intentionally designed taxation—and implicit, regressive inflation.

If something must be paid to preserve price stability—and it must—the only real question is how, and by whom.

That is not an argument against taxes. It’s an argument for finally being honest about what they do.

None of this is a new discovery.

The descriptive foundations have been around for decades. Abba Lerner laid them out in the 1940s. Beardsley Ruml said the quiet part out loud shortly after. More recently, scholars associated with Modern Monetary Theory have spelled out the mechanics in detail.

What’s new isn’t the analysis. What’s new is simply stating the obvious implication that usually goes unstated: if taxation performs a real and valuable service—if it helps protect the value of money and keep the economy usable for everyone—then it isn’t just a necessary evil.

It’s a service.

And like other services we rely on, it’s something we should evaluate seriously, design carefully, and—when it does its job well—recognize as beneficial rather than resent as confiscation. Taxation is a public service. The sooner we treat it as one, the sooner we can have honest and useful debates about how to design and distribute it fairly.

Below, I provide the current list of actions that I think we should take. I am, of course, well-aware that these will not be universally or easily accepted. I am also aware that for these proposals to be fully understood, I will need to document each of them more completely. I hope to do so. Nonetheless, my hope is that what I have proposed here will at least motivate some discussion. I invite you, I encourage you, if you find anything here that you either support, oppose, or believe could be improved or should be modified, please comment on this post or send an email to let me know. Also, if there is something missing, please feel free to suggest it. Thanks in advance for your time and attention.

Summary

The United States government is undergoing a profound stress test. In my view, the constitutional “guardrails” designed to ensure a balance of power have proven insufficient to prevent a dangerous concentration of authority in the executive branch and a corresponding decline in the efficacy of the legislature. This imbalance, coupled with a growing public perception that our economic and political systems are fundamentally unfair, has led to a crisis of governance characterized by executive overreach, legislative gridlock, and eroding public trust.

This document outlines a package of integrated proposals I have developed to address these structural failings. They are not partisan in nature; rather, they are foundational reforms aimed at restoring a functional balance of power. This blueprint is designed to achieve several interconnected goals:

• To re-empower Congress as the nation’s primary lawmaker by expanding the House, reforming the filibuster, and providing it with the institutional capacity to write the detailed laws and regulations needed to govern a complex modern society.

• To clarify and constrain the President’s role to that of the nation’s chief executive, charged with faithfully executing the laws passed by Congress, while preserving the ability to act decisively in a genuine crisis.

• To enhance the integrity and independence of the judiciary by separating administrative adjudication from the executive branch and implementing long-overdue ethics and structural reforms for the Supreme Court.

• To ensure the transparency and accountability of public officials through robust, real-time disclosure of their financial activities.

• To restore public faith in our economic system by making the tax code demonstrably fairer and guaranteeing the nation’s commitment to its most important social insurance program.

This is a proposal for renewing the promise of a government of laws, not of men.

Reclaiming Legislative Authority: Restoring the First Branch

The cornerstone of this reform package is the revitalization of Congress, the branch of government designed to be closest to the people.

Resize and Empower the House of Representatives

The House of Representatives is frozen at a size set in 1929, leading to massive constituencies that diminish representation. I propose that Congress pass a new apportionment act based on the “Cube Root Rule,” a political science standard that would increase the House size from 435 to approximately 700 members based on the current population.

Impact: This would create smaller, more responsive districts, making it more difficult to gerrymander. Crucially, it would re-balance the Electoral College by diluting the disproportionate weight of the two Senate electoral votes given to small states, ensuring the outcome more closely reflects the national popular will without a constitutional amendment.

Restore Congress as the Primary Rulemaker

To reclaim its exclusive constitutional power to legislate, Congress must be empowered to write the detailed regulations that implement its laws. This can be achieved through a new, constitutionally sound process.

• Creation of a Legislative Technical Service (LTS): I propose the creation of a new, independent, and non-partisan legislative branch agency, staffed by the technical experts (scientists, economists, lawyers) who currently write regulations within executive agencies. The LTS’s sole mission will be to provide Congress with the in-house expertise to draft detailed, effective regulations.

• Regulations Passed as Law: All such regulations will be passed as “Joint Resolutions of Regulatory Interpretation.” This ensures they are approved by both the House and the Senate and presented to the President for signature or veto, satisfying the requirements of Article I, Section 7.

• Provision for Executive Action: To ensure governmental agility, the President will be granted statutory authority to issue temporary regulations in response to urgent national needs for which Congress has not yet legislated. These temporary regulations will automatically expire after 60 days unless formally approved by Congress through a Joint Resolution.

Impact: This model restores Congress as the primary author of law and policy, enhances the quality of legislation, and ends unconstitutional delegation, while still providing a crucial safety valve for decisive executive action in a crisis.

Reform the Senate Filibuster

To address legislative gridlock, I propose the Senate reform its rules with a “slow lane” approach: if a cloture vote on a bill fails to achieve a 60-vote majority, the bill is not killed. Instead, a final period of 30 hours of debate is triggered, after which the bill must receive a final up-or-down vote that passes with a simple majority.

Impact: This would eliminate the minority’s ability to completely block legislation, guaranteeing that every bill gets a final vote, while still preserving the Senate’s tradition of extended debate and deliberation.

Abolish the Debt Ceiling to End Self-Imposed Crises

The statutory debt limit is a relic of World War I, originally designed to make it easier for the Treasury to issue bonds and pay the bills that Congress had authorized. Today, its function is the exact opposite. It serves no useful purpose and has been transformed into a political weapon used to threaten a catastrophic default on our national obligations, enabling a minority to interfere with the legally authorized operations of the government. I propose that Congress pass a law to abolish the debt ceiling entirely.

Impact: This would align the U.S. with the practice of nearly every other major economy. It would allow the Treasury to pay the bills that Congress has already incurred, ending the perpetual cycle of manufactured crises that create global financial uncertainty and increase borrowing costs for taxpayers. Debates about the proper level of debt and spending would continue to happen where they belong: during the annual budget and appropriations process

Creating an Institutional Response to the Judiciary

Congress often fails to act when federal courts interpret statutes in ways that diverge from legislative intent, or when new constitutional rulings create legal voids that only Congress can fill. This inertia effectively cedes legislative authority to the courts by default. To address this, I propose the creation of a new, non-partisan division within the Congressional Research Service (CRS) to be called the “Congressional Office of Statutory and Constitutional Review” (COSCR).

The COSCR’s mandate would be to track major federal court decisions and publish an annual “Judicial Impact Report.” This report would trigger a two-track, action-forcing mechanism:

• For cases of statutory interpretation, where the Court has interpreted a law passed by Congress, the relevant committee must report a “Joint Resolution of Congressional Position.” This would force a vote on whether Congress concurs with or dissents from the Court’s interpretation and intends to pursue clarifying legislation.

• For cases of constitutional interpretation that impact the functioning of existing federal law (such as when a right previously protected by the Court is overturned), the committee must report a “Joint Resolution of Legislative Intent.” This would force Congress to vote on a resolution stating whether it intends to amend the affected statutes to conform with the new ruling, or whether it will defer taking legislative action.

Impact: This refined mechanism compels Congress to engage with judicial decisions in a constitutionally appropriate manner. It forces members to take public responsibility for the state of the law without directly challenging the Supreme Court’s authority as the final arbiter of the Constitution. It ensures that the final word on our nation’s laws rests with the elected legislature, not with the judiciary by default.

Defining and Constraining Executive Power

To re-balance the system, the role of the President must be returned to its constitutional boundaries through both constitutional amendment and targeted statutes.

A Constitutional Amendment to Clarify Presidential Authority

The ambiguity of Article II’s Vesting Clause is the source of claims to unlimited, inherent executive power. I propose a constitutional amendment to replace the first sentence of Article II, Section 1 with the following:

“A President of the United States of America shall be vested with the authority to take Care that the Laws be faithfully executed, and to exercise the other powers enumerated in this Constitution.”

Impact: This amendment would definitively end the debate over unenumerated presidential powers. It would clarify that the President’s primary role is to execute the laws passed by Congress, supplemented only by the specific powers explicitly granted in the Constitution.

Reclaim War Powers

To restore Congress’s constitutional role in matters of war and peace, I propose the repeal of the War Powers Resolution of 1973 and its replacement with a new “War Powers and Authorization Act.” This law would require the President to receive affirmative, prior authorization from Congress before introducing armed forces into hostilities, except in the case of repelling a direct attack on the United States. To enforce this, the Act would include an automatic funding cutoff for any unauthorized military deployment after 60 days.

Reform Emergency Powers

To prevent the abuse of emergency declarations, I propose a fundamental reform of the National Emergencies Act of 1976. The new law would require any emergency declared by the President to automatically expire after 30 days unless Congress votes to extend it. When declaring an emergency, the President must specify which statutory powers he intends to use. Any congressional extension would be limited in duration to no more than one year, or sixty days after the beginning of a new Congress, whichever is longer.

Restore the Senate’s Role of Advice and Consent

To prevent the President from bypassing the Senate’s constitutional role, I propose the “Advice and Consent Restoration Act.” This statute would amend the Federal Vacancies Reform Act to limit the tenure of any “acting” official in a Senate-confirmable role to a non-extendable 120 days. Furthermore, it would explicitly prohibit any individual who has been formally nominated for a position and rejected by the Senate from being appointed as the “acting” official for that same role.

Fortifying the Democratic Process

Ensure Full Suffrage for All Citizens

A foundational principle of democracy is that all citizens should have a voice in electing their leaders. Currently, millions of U.S. citizens residing in territories and some living abroad are denied the right to vote for President. This is inconsistent with our nation’s democratic ideals. To correct this, I propose a new constitutional amendment to guarantee the right of all citizens to be represented in the Electoral College.

Section 1. For purposes of representation in the election of the President and Vice President, the United States shall include the territories. Congress shall have the power to implement this article by appropriate legislation, including provisions for citizens residing outside the United States.
Section 2. A number of electors of President and Vice President equal to the whole number of Senators and Representatives in Congress to which the residents of territories would be entitled if the territories were a single State, but in no event more than the least populous State; they shall be in addition to those appointed by the States, but they shall be considered, for the purposes of the election of President and Vice President, to be electors appointed by a State; and they shall meet in the District and perform such duties as provided by the twelfth article of amendment.

Impact: This amendment would enfranchise millions of American citizens, fulfilling the promise of a government representative of all its people and enhancing the legitimacy of presidential elections.

Creating a Modern, Secure, and Humane Immigration System

Our current immigration system is broken. It is economically inefficient, failing to provide legal pathways that meet our country’s labor needs, and it is inhumane to the millions of long-term residents who live in the shadows. To restore the rule of law, I propose that Congress pass a comprehensive reform bill that addresses this problem from all sides. The bill would have four core components:

• A Conditional Pathway to Citizenship: Create a fair and achievable, but earned, pathway to citizenship for undocumented immigrants who entered the country prior to January 1, 2025. To be eligible, they must meet stringent requirements, including long-term residency, passing thorough criminal background checks that disqualify anyone convicted of a serious crime, and passing a civics and English proficiency test similar to the U.S. Naturalization Test.

• Serious Enforcement Against Unlawful Employers: Significantly increase the civil and criminal penalties for employers who knowingly hire undocumented workers, turning off the "jobs magnet" that encourages illegal immigration.

• A Reliable and Mandatory Verification System: To make employer enforcement fair and effective, phase in a mandatory, nationwide E-Verify system. This system must be made simple, fast, reliable, and free for all employers, providing them with a "safe harbor" from prosecution if they use it in good faith to verify the legal status of all new hires.

• New Guest Worker Programs: To meet the demonstrated needs of our economy, especially in agriculture and other non-agricultural sectors with labor shortages, create new, flexible temporary guest worker visa programs. These programs would allow foreign workers to enter the country legally for a defined period to fill essential jobs, but would not represent a pathway to citizenship, thereby reducing the incentive for illegal immigration driven by economic need.

Impact: This balanced, four-part approach creates an immigration system that is both humane and secure. It provides a permanent solution for the existing undocumented population while restoring the integrity of our laws by creating legal channels for temporary work and making it much more difficult for employers to hire undocumented workers in the future. This would create a more orderly, manageable, and fair system for all.

Reinforcing Judicial Independence and the Rule of Law

To ensure the laws are applied fairly, the administrative judiciary must be separated from executive branch influence, and our core legal institutions must be protected from political interference.

Establish an Independent Administrative Judiciary

I propose a constitutional amendment to move the corps of Administrative Law Judges (ALJs) out of the executive and into the judiciary, creating a new tier of specialized, term-limited courts. The proposed amendment to Article III would read:

Section 4.
Clause 1: Congress shall have the power to establish tribunals within the Judicial Branch presided over by Administrative Adjudicators to hear cases and controversies arising from the execution of the laws of the United States, as defined by law.
Clause 2: The requirement in Section 1 of this Article that Judges hold their Offices during good Behaviour shall not apply to Administrative Adjudicators. Congress shall have the power to set the term of office, manner of appointment, and qualifications for such adjudicators.
Clause 3: The right of appeal from the decisions of these tribunals to a Court established under Section 1 of this Article shall not be abridged.

Impact: This reform would create a legitimate, independent judiciary for regulatory disputes, enhancing the separation of powers while the guaranteed right of appeal ensures consistency and oversight.

Implement Supreme Court Term Limits

To de-escalate the political toxicity of Supreme Court confirmations and restore the Court’s legitimacy, I propose a statutory reform to end life tenure in active service. The law would establish a single, non-renewable 18-year term for Supreme Court Justices. To comply with Article III, after 18 years of active service, a Justice would be designated a “Senior Justice,” retaining their office and salary while being available to serve on lower appellate courts. In the event of an unexpected vacancy on the Court, the most recently designated Senior Justice would temporarily return to active status to ensure the Court always has nine members.

Impact: This system creates a regular, predictable appointment process—one new Justice in the first and third year of every presidential term. It makes the Court more responsive to the nation’s democratic evolution over time and lowers the political stakes of any single confirmation battle.

Enact a Binding Code of Ethics for the Supreme Court

The nine Justices of the Supreme Court are the only judges in the federal system not bound by a formal code of ethical conduct. To restore public confidence, I propose that Congress pass the “Supreme Court Ethics, Recusal, and Transparency Act.” This law would require the Judicial Conference to establish a binding code of conduct for the Justices, including clear standards for recusal, gift and travel transparency, and a requirement for written justification when a Justice declines to recuse themselves from a case where a conflict has been raised.

Codify Protections for the Department of Justice

To prevent the politicization of justice, I propose that Congress pass the “Department of Justice Independence Act.” This law would establish statutory “for cause” removal protections for the Attorney General and the Director of the FBI. “For cause” would be defined to include misconduct or dereliction of duty, while explicitly excluding a refusal to interfere in an investigation for political reasons.

Ensuring Transparency and Accountability for Public Officials

A healthy republic requires an informed electorate and confidence that officials are working in the public interest.

Mandate Comprehensive Financial Transparency

I propose a new statute, the “Federal Candidate and Officeholder Tax Transparency Act,” to ensure continuous financial disclosure.

• Initial and Annual Disclosure: Upon filing with the FEC, a candidate’s six preceding years of tax returns will be published within 90 days. An annual disclosure of the prior year’s return will then be required for every year the individual remains an active candidate or serves in office.

• Right to Review and Correct: A 15-day window will be provided for candidates/officeholders to review redacted documents and correct any IRS clerical errors before publication.

• Plain-Language Summary: The IRS will be required to publish a standardized, one-page summary of key data (Total Income, Taxes Paid, Effective Tax Rate, Charitable Contributions) for each return.

• Tiered Audits and Secure Portal: A full, manual audit will be reserved for high-level officeholders and major candidates, while all returns will be published on a single, secure, official government portal.

Impact: This system creates robust, continuous transparency while ensuring procedural fairness and providing clear, accessible information to voters.

Require Pre-Trade Disclosure for Members of Congress

To prevent conflicts of interest, I propose that Congress be subject to the same kind of prophylactic ethics rules that already apply to corporate executives with access to privileged information. A new law would require members of Congress, their spouses, and senior staff to publicly file a “Notice of Intended Transaction” 30 days prior to making a trade in any individual stock or security, a direct application of the principles behind the SEC’s Rule 10b5-1.

Impact: This proactive “sunlight” approach forces accountability before a transaction occurs and creates a powerful deterrent against even the appearance of a conflict of interest.

Promoting Economic Fairness and Fiscal Responsibility

Restoring trust in government also requires addressing the perception that our economic system is no longer fair.

Guarantee Social Security and Restore Public Confidence

I propose that Congress pass the “Social Security Guarantee Act.” This act would reform the program’s financing by eliminating the automatic benefit reduction provision, while explicitly authorizing the Treasury to make payments from the general fund to cover any shortfall. The Social Security Trust Fund would be retained as an important accounting tool to track the program’s finances and as a powerful symbol of the “earned right” nature of the benefits.

Impact: This reform would end the recurring political crises and public anxiety over the program’s future, making the federal government’s commitment to pay scheduled benefits explicit and ironclad.

Create a More Progressive and Fairer Tax System

To restore public confidence in the tax code’s fairness, I propose a series of reforms to ensure that the wealthiest Americans and most profitable corporations pay a fair share.

• More Progressive Personal Income Tax: Add new tax brackets for extremely high incomes to ensure that the marginal tax rate on annual personal income above $25 million (in 2025 dollars) is no less than 90%.

• Cap the Capital Gains Preference: Limit the amount of investment income eligible for the preferential long-term capital gains rate to $5 million per year. Any capital gains income above this cap would be taxed at the higher ordinary income rates.

• Limit the “Step-Up in Basis”: Limit the “step-up in basis” that allows capital gains to be passed on to heirs tax-free. I propose a $5 million lifetime exemption per person on capital gains transferred at death. Any gains above this amount would be taxed.

• Close the Carried Interest Loophole: Pass a law to close the loophole that allows private equity managers to classify their labor income as capital gains, ensuring they pay the same ordinary income tax rates as other high-earning professionals.

Impact: This suite of targeted tax reforms ensures that at the highest levels of income and wealth, income from wealth is no longer treated more favorably than income from work. This would increase tax fairness, generate significant revenue, and reduce extreme economic inequality.

Promoting Fair Competition and Rewarding True Value Creation

A healthy market economy depends on a tax system that rewards genuine innovation and productive enterprise. The current flat corporate tax fails to distinguish between the normal profits earned by a business in a competitive environment and the supernormal profits that can result from monopoly power, government-granted privileges, or other market distortions. To promote true competition, I propose that the nation begin a serious debate on replacing the flat tax with a progressive surtax on supernormal corporate profits. Such an “excess profits tax” would be indexed to a company’s rate of return on capital in excess of some legislated “normal” return. While I recognize that the precise definitions and mechanisms require rigorous expert analysis, the principle itself provides a clear path forward. To that end, I propose that Congress formally direct the Congressional Budget Office and the Joint Committee on Taxation to study this proposal and report to the American people with specific, workable models.

Impact: Initiating this debate would signal a commitment to strengthening free-market principles. It would begin the process of designing a more efficient tax system that encourages true competition and rewards earned success over the extraction of economic rents, thereby ensuring a more dynamic and trusted capitalist system.

Restore the Purpose of Tax-Advantaged Retirement Accounts

To address the issue of “mega-IRAs” being used as tax shelters, I propose that Congress establish a cumulative cap on the total value of an individual’s tax-advantaged retirement accounts at $5 million (indexed to inflation). If an individual’s aggregate balance exceeds this cap, the excess amount would be subject to a Required Minimum Distribution (RMD). To ensure fairness, this RMD would be phased in over a five-year period, with a minimum of 20% of the initial excess amount required to be withdrawn each year. These withdrawals would be taxed as ordinary income.

Impact: This highly targeted reform would affect only a tiny fraction of the wealthiest savers, restoring the original purpose of retirement accounts while generating significant tax revenue in a fair manner.

Mandate a “Single Price” System for All Healthcare Services

The current healthcare market is broken, characterized by opaque, discriminatory pricing that favors the largest players and punishes the vulnerable. To restore sanity and fairness to healthcare pricing, I propose the “Healthcare Fair Price Act.” This law would not set prices, but would require that any given provider — be it a hospital, clinic, or doctor — must set a single, transparent, and published price for each specific service, drug, or medical device they provide. That price must be charged to every payer, whether they are a large insurance company, a small insurer, or an uninsured individual paying out-of-pocket. Price discrimination would be illegal.

Impact: This reform is not a government takeover of healthcare; it is a market-based rule to ensure true competition. It would have several profound effects:

• It would create unprecedented price transparency, allowing consumers to shop for care based on price and quality for the first time.

• It would level the playing field between large and small insurance companies, forcing them to compete based on their own internal efficiency, customer service, and ability to manage care, rather than their ability to use market muscle to get secret discounts.

• It would end the practice of price-gouging the uninsured, ensuring they are charged the same rate as the most powerful insurer.

• Crucially, it would be a major step toward making a future single-payer system more achievable. By creating a single, transparent price for every medical service in the country, it would dramatically simplify the eventual transition to a system where the government acts as the single payer for those established prices.

Ending the “Hidden Tax” of Uncompensated Care

Under federal law (EMTALA), hospitals are required to provide emergency care to everyone, regardless of their ability to pay. While this is a moral and legal imperative, the cost of this uncompensated care is not paid for by the government. Instead, hospitals cover these costs by overcharging their privately insured patients. This “cost-shift” functions as a hidden and inefficient tax on everyone with health insurance. I propose that Congress create a new federal program to directly reimburse hospitals for the documented costs of providing EMTALA-mandated treatment to the uninsured.

Impact: This reform would make the cost of our nation’s healthcare safety net transparent and honestly funded by the public through taxes, rather than hidden in the inflated hospital bills and insurance premiums of private citizens. It would eliminate a primary justification hospitals use for charging high prices, which, when combined with other reforms, should lead to lower healthcare costs for all Americans.

Enforcing the “Community Benefit” Standard for Non-Profit Hospitals

Non-profit hospitals receive billions of dollars in tax exemptions in exchange for providing a “community benefit.” However, this standard is poorly enforced, and many non-profits use their market power to charge exorbitant prices. To ensure this public subsidy is truly earned, I propose that Congress pass the “Non-Profit Hospital Accountability Act.” This law would establish a clear, bright-line rule for maintaining tax-exempt status: over a given fiscal year, the average price a hospital charges to private payers across all services may not exceed 250% of the average Medicare reimbursement rate for those same services. The penalty for violating this standard would be the revocation of the hospital’s tax-exempt status.

Impact: This reform uses a powerful, existing enforcement mechanism to curb systemic price-gouging by non-profit monopolies. By using an average, it gives hospitals the flexibility to manage the complex finances of different medical services while still ensuring that their overall pricing structure serves the community rather than simply maximizing revenue.

Proposals I Have Not Made

To provide a clearer understanding of this document’s perspective, it is useful to outline several major reform proposals, often associated either with modern conservatives or with liberals, for which I am not advocating. While I don’t accept these proposal as they stand, I am still often considering doing so or trying to find better solutions to the problems addressed.

Proposals Often Made by Conservatives

A Balanced Budget Amendment

This would amend the Constitution to require that federal spending not exceed revenue in any fiscal year.

Impact: Advocates for this argue that it would enforce fiscal discipline, end deficit spending, reduce the national debt, and lift an immoral burden from future generations.

Rebuttal: This would eliminate the government’s ability to use fiscal policy to fight recessions (through automatic stabilizers and stimulus spending), potentially turning minor downturns into major depressions. It would also force massive, immediate cuts to popular programs like Social Security, Medicare, and national defense.

Term Limits for Members of Congress

This would amend the Constitution to limit the number of terms a member of the House or Senate could serve.

Impact: Advocates for this argue that it would ensure a constant influx of “citizen legislators,” break up entrenched power, reduce the influence of lobbyists, and make Congress more responsive to the people.

Rebuttal: This would create a legislature of perpetual amateurs, shifting power from elected representatives to unelected, permanent staff and lobbyists who would possess all the institutional knowledge. It also infringes on the right of voters to choose whomever they deem most qualified to represent them.

Repeal of the 17th Amendment

This would amend the Constitution to return the power of appointing U.S. Senators to the state legislatures.

Impact: Advocates for this argue that it would restore the principle of federalism by making Senators directly accountable to state governments, giving states a powerful check on federal overreach.

Rebuttal: This would be a profoundly anti-democratic step, disenfranchising millions of voters and returning the selection of Senators to state legislatures, which are often heavily gerrymandered and prone to corruption and partisan gridlock.

A Federal “Regulatory Budget”

This would enact a law to cap the total economic cost that federal agencies can impose on the public through new regulations each year.

Impact: Advocates for this argue that it would make the hidden costs of regulation transparent and force a public debate on the trade-offs, creating an incentive for agencies to find the most efficient, least burdensome regulations.

Rebuttal: It is practically impossible to accurately quantify the costs (and especially the benefits, like cleaner air or safer products) of regulations in advance. This would create a system that prioritizes easily measured compliance costs over harder-to-measure public benefits, leading to weaker protections for health, safety, and the environment.

Converting Federal Programs into Block Grants

This would convert large federal social programs (like Medicaid) into block grants, sending lump sums of money to states to run their own programs with broad flexibility.

Impact: Advocates for this argue that it would shrink the federal bureaucracy and allow states to act as “laboratories of democracy,” creating innovative and more efficient programs tailored to their local needs.

Rebuttal: This would likely lead to a “race to the bottom,” as states, facing budget pressures, would have a strong incentive to cut benefits and services for their most vulnerable populations. It would also eliminate the national standards that ensure a basic level of support for all citizens, regardless of where they live.

A National “Right-to-Work” Law

This would enact a federal law stating that no person can be forced to pay union dues as a condition of employment.

Impact: Advocates for this argue that it would protect individual workers’ freedom of association and end the system of “compulsory unionism,” making labor markets more flexible and promoting economic growth.

Rebuttal: This is designed to weaken unions by allowing “free riders” — workers who benefit from a union’s collective bargaining without contributing to its costs. This would erode the power of organized labor, leading to lower wages, weaker benefits, and less safe working conditions for all workers.

Enact National School Choice

This would enact a law allowing federal education funds to follow a student to any accredited school of their choice—public, private, charter, or homeschool—rather than being sent directly to public school districts.

Impact: Advocates for this argue that it would introduce competition into the K-12 education system, breaking the “government monopoly” and forcing all schools to improve. It would empower parents, especially those in low-income areas with failing public schools, to choose the best educational environment for their children.

Rebuttal: This would drain critical funding from public schools, which are obligated to serve all students, and subsidize private and religious schools that are not accountable to the public. It would likely lead to greater stratification and inequality in education, leaving the most disadvantaged students behind in under-resourced public schools.

Abolish Redundant Federal Departments

This would propose the systematic elimination of entire cabinet-level departments whose functions are seen as unconstitutional or better left to the states and the private sector. Common targets include the Department of Education, the Department of Commerce, and the Department of Energy.

Impact: Advocates for this argue that it would dramatically reduce the size, cost, and scope of the federal government, returning power to the states and the people as the Framers intended. It would eliminate redundant programs and burdensome regulations that stifle economic growth.

Rebuttal: This would eliminate critical national functions. Abolishing the Department of Energy would gut oversight of the nation’s nuclear arsenal; eliminating the Department of Commerce would harm our ability to conduct the census and manage oceanic resources; and eliminating the Department of Education would harm students with disabilities and those in low-income communities who rely on federal programs and protections.

Require a Supermajority to Raise Taxes

This would amend the Constitution to require a two-thirds or three-fifths supermajority vote in both houses of Congress to pass any new federal tax or increase any existing tax rate.

Impact: Advocates for this argue that it would create a powerful and permanent check on the government’s ability to raise taxes, protecting taxpayers from the shifting winds of politics and forcing Congress to focus on cutting spending rather than raising revenue to solve fiscal problems.

Rebuttal: This would give a small minority of legislators a permanent veto over the fiscal policy of the entire country, leading to chronic gridlock and making it impossible to respond to national crises or address the national debt in a balanced way. It would force the government to rely almost exclusively on spending cuts, disproportionately harming the most vulnerable.

Proposals Often Made by Liberals

Enact Public Financing of Elections

This would fundamentally reform campaign finance by creating a system of public funding for all federal elections. One common model provides candidates with a 6-to-1 match for small-dollar donations (e.g., up to $200), funded by a surcharge on corporate lawbreaking penalties. This would be paired with a constitutional amendment to overturn judicial precedents like Citizens United, allowing Congress to regulate independent expenditures by corporate entities.

Impact: Advocates for this argue that it would empower ordinary citizens, reduce the influence of wealthy megadonors and special interests, and free candidates from the endless cycle of fundraising.

Rebuttal: This infringes on the First Amendment’s protection of political speech. It would also force taxpayers to subsidize the campaigns of candidates they find repugnant and could entrench the power of incumbent parties.

Establish Universal Pre-Kindergarten and Child Care

This would create a new, federally-funded social program to guarantee access to high-quality, affordable child care and pre-kindergarten for every family in America. The program would cap what most families pay for child care at a small percentage of their income.

Impact: Advocates for this argue that it is a critical investment in early childhood development, would enable millions of parents (especially women) to join and remain in the workforce, and would significantly reduce both gender and economic inequality.

Rebuttal: This would be an enormously expensive new federal entitlement, creating a massive new bureaucracy and placing a heavy burden on taxpayers. It could also crowd out existing private and faith-based child care providers. In any case, the proper source for the bulk of school funding is, and should remain, the states, not the federal government.

Provide Widespread Student Debt Relief and Affordable Higher Education

This reform would address the student debt crisis through a two-pronged approach: a one-time cancellation of a significant amount of federal student loan debt (e.g., $50,000 per borrower) and making public colleges, universities, and trade schools tuition-free for most families.

Impact: Advocates for this argue that it would be a powerful economic stimulus, freeing up a generation from a crushing debt burden and allowing them to buy homes, start families, and launch businesses. It would treat higher education as a public good, essential for a modern economy.

Rebuttal: Conservatives would argue that this is a regressive policy that disproportionately benefits higher-earning college graduates at the expense of taxpayers who did not attend college. They would also believe that it is fundamentally unfair to those who have already paid off their student loans or chose more affordable educational paths. Neither of these arguments is compelling. [TBD – The correct solution requires more thought.]

Pass Comprehensive Criminal Justice Reform

This would enact a major federal law to address mass incarceration and police misconduct. Key provisions would include ending cash bail for non-violent offenses, banning chokeholds and no-knock warrants at the federal level, creating a national registry for police misconduct, and reforming mandatory minimum sentencing laws.

Impact: Advocates for this argue that it would reduce the profound racial disparities in the justice system, make communities safer by focusing resources on serious crime, and create a more just and humane system for all.

Rebuttal: [TBD]

Strengthen Labor and the Right to Organize

This would pass a law, often modeled on the PRO Act (Protecting the Right to Organize Act), to make it easier for workers to form and join unions. It would strengthen penalties on companies that engage in union-busting, ban “captive audience” meetings, and override state-level “right-to-work” laws.

Impact: Advocates for this argue that it would reverse decades of declining worker power, allowing employees to bargain for better wages, benefits, and working conditions. This would be a powerful tool for reducing economic inequality.

Rebuttal: [TBD]

Assessment and Anticipated Objections

This package of reforms, if enacted, would fundamentally alter the American system of government, creating a new equilibrium designed to restore public trust. The overarching impact would be to forge a government that is more representative, transparent, ethical, and deliberative. Power would shift decisively from the executive to a larger, more capable, and more responsive legislature. The President’s role would be clarified as the nation’s chief administrator, while the judiciary’s independence and integrity would be reinforced. Key economic policies, from taxation to social insurance, would be reformed to enhance public confidence in their fairness and stability by ensuring that the benefits of our economy are more broadly shared and that income from immense wealth is not treated more favorably than income from work.

However, such a profound restructuring would face significant and powerful objections. Critics would argue that these changes would create a government that is inefficient and prone to gridlock, a threat to national security due to a loss of executive energy, and one that risks economic disruption. I disagree.

Ultimately, in my view, this framework represents a deliberate balancing of interests. It trades the speed and discretion of a powerful executive for the deliberation, consensus-building, and accountability of a strong legislature. It is based on the belief that rebuilding the public’s trust in the fairness, integrity, and representativeness of their government is the most critical task facing the Republic, even if it comes at the cost of some efficiency.

I recognize that the political challenges to enacting such a comprehensive package are monumental. However, the value of this framework is not measured solely by its immediate feasibility. It is intended to serve as a “North Star”—a coherent vision of a re-balanced constitutional order. Knowing what an ideal solution looks like is essential for evaluating the merit of the more incremental, “next best” reforms that may be proposed. Furthermore, laying out the full scope of these interconnected problems and their potential solutions serves a vital public purpose: to stimulate a more ambitious national conversation and motivate others to engage in the ongoing work of perfecting our Union.

A potentially more significant, and frequently overlooked, aspect of tax history is the historical evolution of tax bracket thresholds and their interplay with economic factors. Adjustments to these brackets, particularly when analyzed in real (inflation-adjusted) terms, can have an impact on taxpayer liability equivalent to, or even greater than, changes in statutory rates. Even when bracket thresholds and rates appear to be stable from year to year, inflation can lead to 'bracket creep,' effectively increasing an individual's real tax burden as nominal income growth pushes them into higher tax brackets. During periods of inflation, stable tax policy is essentially a policy of constant tax increases.

Presenting historical tax brackets in nominal dollars, rather than their equivalent worth today (real dollars), can make it genuinely challenging for most people to understand their true historical impact.

Consider this common scenario: An individual today might be aware that the top federal income tax rate in 2024 applied to incomes exceeding roughly $609,350. If they then learn that in 1913, the first year of our modern income tax system, the top rate was levied on incomes over $500,000, their initial reaction might be that the thresholds aren't dramatically different. However, their perspective on historical policy shifts would likely change significantly upon discovering that $500,000 in 1913 had the purchasing power of about $16 million in 2024 dollars. The 2024 top bracket threshold, rather than being about 20% higher than the 1913’s, is actually only about 4% of 1913’s threshold. Today’s reader might also be surprised to discover that in several past years, the top marginal rate only applied to incomes that were greater than $100 million in today’s dollars!

Unfortunately, presenting both brackets and rates in a single chart is difficult. Nonetheless, I’ve attempted to produce such a chart which you’ll find below with some explanation following.2

In the chart, whose y-axis has a log scale, the brackets for each year (as many as 56) are shown by horizontal black lines and the tax rate for each bracket is indicated by its color, according to the legend on the right.

Low-End Taxes have remained low while Top Taxes have fallen dramatically

What should be immediately apparent is that the tax rates that have applied to single filers with incomes similar to today’s median income, about $57,000, have not changed dramatically over time. What has changed is that taxes on wealthier filers have dropped dramatically. While those with annual incomes greater than $1 million once paid much higher taxes than others, and those with incomes of $10 million or $100 million paid even higher taxes, today everyone with an income over $609,350 pays a rate once only paid by those with much, much lower incomes.

Bracket-Creep is quite visible

The chart above shows the effect inflation once had on real bracket thresholds. As can be seen by comparing that inflation-adjusted chart to the un-adjusted chart below, it is clear that even when nominal dollar brackets didn’t change, “Bracket-Creep” due to inflation tended to drive brackets constantly lower from year to year between 1940 and 1985. During those years, even if one’s real income was stable, the applicable tax rate grew over time due to inflation. Since 1985, the bracket thresholds have been annually adjusted for inflation.

War-time once led to higher taxes

During the periods corresponding to World War I (late 1910s) and especially World War II (early-mid 1940s) rates increased dramatically across many income levels and so did the number of brackets. The high WWII rates persisted until the tax cuts proposed by President Kennedy were adopted in the Revenue Act of 1964.

The sheer complexity of the mid-century system is striking

The density of lines and color gradations from roughly the 1940s to the 1960s visually screams "complexity!" This wasn't just about high top rates; it was about a highly articulated system of brackets with many small steps. Given this complexity, it isn’t surprising that tax simplification became a recurring political theme.

Periods of extreme volatility vs. relative stability in tax structure:

The graph highlights distinct eras of tax policy. The period from the 1930s through the 1970s appears particularly turbulent in terms of both rates (color shifts) and bracket structures (lines moving). The early years (1913-1920s) also show rapid changes. In contrast, the period from the late 1980s/early 1990s through to today appears much more stable in terms of both bracket and rate structure. Of course, this Post-Reagan era, reflects the impact of the Republican’s almost absolute refusal to consider tax increases as well as their focus on tax cuts, particularly for the wealthy, whether or not there is any evidence that such tax cuts will either generate increased revenue or be offset by spending cuts. The result of this refusal to raise taxes, codified via the Taxpayer Protection Pledge, has resulted in a rapid increase in the national debt — a harbinger of future tax increases.

While I believe that the chart above more clearly illustrates key aspects of income tax history than do others I have found, I recognize that it is a bit of a mess — and may be challenging for the color blind... It is useful to compress a great deal of information in a single image, but sometimes, one can go too far. Thus, I’m hoping that if anyone knows of a more effective presentation of this information that they will let me know of it.

I haven’t found much evidence of previous attempts to display the history of both brackets and rates in a single image while also considering the impact of inflation. I’ve also found very little evidence of detailed academic studies that present the history of brackets. Almost all the studies I have found, other than a few that analyze the Bracket-Creep problem, are focused almost exclusively on rates, not on brackets. However, one study that I have found to be very useful was: Tracey M. Roberts’s 2014 article, Brackets: A Historical Perspective.3 If this subject interests you, I think you’ll find Roberts’s discussion worth the time to read.

Let me know in the comments if this presentation is at all useful to you. Also, please use the comments to make any suggestions for improving the chart above, or for an alternative, superior presentation.