The Number That Moved by Standing Still
How the fixed $500 Statute of Frauds threshold lost 91.6% of its real value while its text never changed — and what a nominal boundary in an inflating unit of account really captures
Keywords: Statute of Frauds; UCC § 2-201; nominal threshold; inflation erosion; CPI-U; real value; indexation; payment finality; documentation cost; transaction costs; value axis
Abstract. Section 2-201 of the Uniform Commercial Code makes an oral contract for the sale of goods unenforceable at or above $500, a figure fixed in the 1952 Official Text and unchanged since. Using the operative statutory text (verified at Cornell LII) and the U.S. Bureau of Labor Statistics CPI-U annual-average series (verified), this essay computes that the real value of the fixed $500 threshold has fallen by 91.6% since 1952—to $42.24 in 1952 purchasing power, against the $5,918.66 that constant real value would require in 2024. The erosion is expressed as an identity, R(t)=$500×CPI(1952)/CPI(t), carrying no econometric vulnerability because nothing is estimated. The direction of the resulting expansion of the writing requirement’s coverage is shown to be monotone and distribution-free; its magnitude is explicitly not claimed, because it would require a distribution of real transaction values not measured here. Locating the threshold on the value-and-identifiability axis developed in two prior essays, the essay argues via Fox (1996) that the Statute and money’s currency are one evidentiary instrument calibrated to different axis positions; via Kahn and Roberds (2009) that recorded payment is dissolving the rule’s documentation-cost premise from below as inflation drags it down from above; via Schwartz and Scott (2011) and Levmore (1987) that the threshold addresses a fabrication hazard whose real placement inflation has silently moved; and via Coase (1937) that the concentrated-cost, diffuse-benefit structure of correction predicts the observed non-correction. What the erosion captures is reliance: the enforceability of ordinary oral bargains, withdrawn by the accident of a number that moved by standing still.
I. A number that has not moved
Section 2-201 of the Uniform Commercial Code, the Statute of Frauds for the sale of goods, makes an oral contract for goods unenforceable above a stated price. The stated price is $500. The operative clause makes a contract for goods priced at “$500 or more” unenforceable without a signed writing (UCC § 2-201(1)). That figure appears in the 1952 Official Text of the Code and it is still $500 today. The nominal number has not moved in seventy-two years.
The doctrinal ancestor is older. The English Statute of Frauds of 1677, An Act for the Prevention of Frauds and Perjuries, first imposed a writing requirement on certain classes of promise; its sale-of-goods provision was denominated in pounds sterling, not dollars, and the $500 figure is not drawn from it. The lineage is real but the number is not inherited from 1677. The number is a twentieth-century American figure, fixed in the 1952 Official Text, and the claim advanced in this essay concerns that figure and its behavior since 1952.
What the figure does is draw a legal boundary on a value axis. Below $500, an oral agreement stands; a handshake binds. At or above $500, the law withdraws its enforcement machinery unless the parties reduce the bargain to a signed writing. The threshold is not a penalty and not a tax; it is a switch between two legal regimes for the same act of exchange. The premise of the Statute is that above some value, the risk of a fabricated contract, of perjury about a bargain that never happened, is grave enough to justify requiring documentary evidence, while below that value the cost of demanding a writing for every trivial trade exceeds the fraud it would prevent. The threshold is where the drafters placed the line between those two costs.
A nominal dollar line placed once and never adjusted does not stay where it was placed. Prices rise; the real value of a fixed nominal figure falls. The boundary the drafters set at a particular point in the real distribution of transaction values slides, year by year, down that distribution, capturing ever smaller real transactions inside the writing requirement. This is not a claim about anyone’s intent. It is arithmetic, and this essay computes it exactly from the official price index, then asks what the computed erosion means for the theory of why the threshold exists at all. The connective tissue is a framework I have developed across two prior essays: that legal rules governing exchange are positioned on an axis of value and identifiability, and that a rule fixed in nominal terms is a rule whose position on that axis is silently moved by inflation while the text stays still.
II. What must be true, and what I have verified
The quantified claim in this essay rests on two primary-source facts and one arithmetic operation. I state each, its source, and its status, before using any of them, because a number used without its provenance is not evidence.
Fact 1: the threshold is $500 and dates from the 1952 Official Text. Source: the operative text of UCC § 2-201(1), read in full from the Legal Information Institute at Cornell Law School (https://www.law.cornell.edu/ucc/2/2-201). Status: verified against the primary text, and corroborated by state codifications carrying the identical figure. I am not relying on a secondary characterization of the statute; I read the section.
Fact 2: the annual-average Consumer Price Index for All Urban Consumers (CPI-U, U.S. city average, all items, 1982–84 = 100) was 26.5 in 1952 and 313.689 in 2024. Source: the U.S. Bureau of Labor Statistics historical CPI-U table (series CUUR0000SA0), the official price index of the United States. Status: verified from the BLS series. The 1952 figure is the published annual average; the 2024 figure is the published annual average. I use annual averages rather than single-month readings because the annual average is the standard basis for year-over-year real comparison and removes within-year seasonal variation; this choice is stated, not hidden, and the essay’s conclusion does not depend on it, since any two comparable CPI-U readings produce the same order of magnitude.
The operation. Real erosion of a fixed nominal figure between two dates is the ratio of the price indices. The CPI-U rose from 26.5 to 313.689, a ratio of 11.8373. Therefore a threshold that was $500 in 1952 would need to be $500 × 11.8373 = $5,918.66 in 2024 to occupy the same real position it occupied when it was written. Equivalently, the $500 line as it stands in 2024 captures a transaction whose real value, in 1952 purchasing power, is $500 × (26.5 / 313.689) = $42.24. The fixed threshold has lost 1 − (26.5 / 313.689) = 91.6% of its real value since the Official Text fixed it. These are not estimates. They are the direct arithmetic of two verified index values, and they are reproducible by anyone with the BLS series.
One limitation is stated plainly rather than buried. The CPI-U is itself a constructed index, not a directly observed price; it embeds the BLS basket, weighting, and methodology, and those have changed over the period. I do not claim the CPI-U is the true and only measure of the price level. I claim it is the official measure, that it is the standard instrument for exactly this kind of real-value comparison, and that the erosion it records is so large—an order of magnitude—that no defensible alternative price index would overturn the qualitative result. A reader who prefers the PCE deflator or a chained index will get a different second decimal and the same conclusion: the threshold has lost the great bulk of its real value. Where the essay’s argument turns on the magnitude, I rely on the order of magnitude, which is robust to the choice of index; where it turns on the exact figure, I give the CPI-U figure and name it as such.
One point of provenance must be stated precisely, because the anchor year determines every computed figure and an imprecise anchor would infect the arithmetic. The $500 figure is the one carried in the 1952 Official Text of the Uniform Commercial Code and adopted by the states over the following decade and a half as they enacted Article 2. I anchor the computation on 1952, the year of the Official Text, and use the 1952 CPI-U annual average of 26.5 accordingly. I do not anchor on the earlier Uniform Sales Act, whose text I have not read and whose sale-of-goods figure I therefore cannot represent; to compute erosion from a date whose statutory figure I had not verified would be to build on an unchecked premise. The anchor is 1952 because 1952 is the date for which I have verified both the statutory figure and the price index. A reader who wished to anchor on a state’s specific adoption year—which varied—would shift the base CPI-U by the few points separating 1952 from the mid-1960s and would obtain a real erosion of the same order, but I do not present those variants as computed results because I have anchored, and verified, on one date. The discipline is that every number in this essay traces to a date and a source I checked, and the anchor is the year I could check.
III. The evidentiary root: why the Statute and money’s finality are the same instrument
The Statute of Frauds and the currency of money are usually filed in different drawers—one in contract, one in property and payments—but the framework of this series shows they are two settings of a single instrument, and Fox’s (1996) reconstruction of money’s currency supplies the proof. Both rules are responses to the same problem: the reliability of proof about a past transaction. Both draw a line keyed to that reliability. And both were positioned, originally, in real terms that inflation has since moved.
Fox showed that money acquired its defining legal attribute—currency, the rule that an honest taker for value gets a fresh title extinguishing all prior claims—in two successive rationales, the first purely evidentiary. Coined money had, in the language of the early cases, no earmark; one coin could not be distinguished from another, so a former owner could never prove that specific coins were his once they mixed with another’s holding. The courts drew the consequence that no action for the specific recovery of loose coin would lie, because the proof was impossible. Money’s finality began as an evidentiary surrender: where proof of identity was impossible, the law stopped entertaining claims that depended on it. Fox’s later chapters show the rule hardening into doctrine even after paper money made individual instruments identifiable, because the exchange function demanded that identifiability be treated as legally inert.
The Statute of Frauds is the same evidentiary logic applied at the opposite pole. Where money’s currency says proof of a coin’s provenance is so unreliable that we will not entertain claims resting on it, the Statute says proof of an oral bargain is so unreliable, above a certain value, that we will not entertain claims resting on it without a writing. Both rules identify a domain in which oral or possessory proof is too weak to bear the weight placed on it, and both respond by refusing enforcement that would depend on that weak proof. The difference is only the axis position. Money sits at the extreme where proof of identity is impossible and the law abandons the claim entirely; the Statute sits in the interior where proof is merely unreliable above a value, and the law demands documentary reinforcement. They are the same instrument—an evidentiary threshold keyed to the reliability of proof—calibrated to different points on the axis this series has been mapping.
Seeing them as one instrument sharpens what the $500 erosion means. Fox’s currency rule was calibrated to a physical fact—the indistinguishability of coin—that did not drift; the coin of 1758 was as indistinguishable as the coin of 1958, so the rule’s real content was stable even as its doctrinal rationale shifted. The Statute’s threshold, by contrast, was calibrated to a value, and value is denominated in the very unit whose real worth inflation erodes. The Statute made the mistake money’s currency did not: it pegged an evidentiary line to a nominal quantity rather than to a physical or structural fact, and so its real content decays where money’s did not. The comparison isolates the source of the pathology. It is not that evidentiary thresholds are inherently unstable; money’s is stable. It is that thresholds keyed to nominal value, in an inflating unit, are unstable, and the Statute’s is keyed to nominal value. The erosion is the price of denominating an evidentiary judgment in a moving unit of account.
IV. The erosion, computed
The arithmetic of the previous section can be carried across the whole period, and the result is a curve, not a point. Taking the CPI-U annual average in each year and dividing $500 by the ratio of that year’s index to the 1952 index gives the real value, in constant 1952 purchasing power, of the still-nominal $500 threshold. The figures below are computed directly from the BLS CPI-U series; every value is the arithmetic of two published index numbers.
In 1952 the threshold was, by construction, worth $500 in 1952 dollars. By 1960 it had fallen to $448. By 1970, to $341. By 1980, after the inflation of the 1970s, to $161. By 1990, to $101. By 2000, to $77. By 2010, to $61. By 2020, to $51. By 2024, to $42.24. The line the drafters drew at $500 now sits, in the money of the year it was written, at roughly forty-two dollars. A transaction that in 1952 the Statute deliberately left outside the writing requirement—a forty-dollar handshake, beneath the drafters’ threshold—is today, if its nominal price reaches $500, inside the writing requirement, even though in real terms it is the same forty-dollar transaction the drafters chose to exempt.
The shape of the curve matters as much as its endpoints. It is a decay: monotone downward under any positive inflation, steepest where inflation was fastest, and—this is the feature that defeats the intuition that the erosion is a manageable drift—never self-correcting. A nominal threshold does not merely fall; it falls at a compounding rate. Over the seventy-two years from 1952 to 2024 the CPI-U compounded at 3.49% per year, and 3.49% per year for seventy-two years is not a modest adjustment but a factor of nearly twelve. The threshold did not lose a manageable fraction of its value. It lost 91.6% of it. What was set as a serious commercial line—$500 in an economy where that bought what roughly $5,900 buys now—has become a line that catches the sale of a used appliance, a mid-range phone, a set of tires. The writing requirement, conceived to apply to consequential bargains, now applies to the ordinary transactions of daily commercial life, and it does so not because anyone decided it should but because the number stood still while the currency moved beneath it.
V. The compounding dynamics, made explicit
The erosion identity R(t) = H · P(t₀)/P(t) conceals a dynamic worth drawing out, because the rate at which the real threshold falls is not constant and its non-constancy is what makes the phenomenon durable. Differentiating the real value with respect to time, the proportional rate of decline of R(t) equals the instantaneous inflation rate: dR/R = −dP/P. The real threshold falls fastest exactly when inflation is fastest, and it falls even in low-inflation years, never recovering ground, because there is no deflationary mechanism in the rule to push R back up. The 1970s, when CPI-U inflation ran into double digits, account for a disproportionate share of the total erosion: R(t) fell from $341 in 1970 to $161 in 1980, losing nearly half its remaining real value in a single decade. But the low-inflation decades did their work too. From 2010 to 2024, with inflation mostly modest, R(t) still fell from $61 to $42, a further third of the remainder. The decay is relentless precisely because it is multiplicative: each year multiplies the real threshold by the reciprocal of that year’s price ratio, and a product of numbers each slightly below one converges toward zero without ever reversing.
This multiplicative structure defeats the intuition that a threshold set “high enough” is safe for a long time. A threshold set at $500 in an economy where that was a serious sum feels durable; the drafters could reasonably have thought it would govern consequential transactions for a generation. But multiplicative decay has no safe distance. At the realized CPI-U compound rate of 3.49% per year, the real threshold halves roughly every twenty years—the rule-of-72 applied to 3.49 gives about 20.6 years to halving—so $500 becomes $250 in real terms by the early 1970s, $125 by the early 1990s, $62 by the early 2010s, and continues down. Three halvings and the threshold has lost seven-eighths of its value; the drafters’ generation-long horizon was, in real terms, a couple of decades. No nominal figure, however generously set, escapes this, because the halving time depends only on the inflation rate, not on the starting value. Setting the threshold higher buys a fixed number of years of headroom, not permanence; the decay rate is invariant to the level.
The dynamic also quantifies the cost of the correction that was not made. Had the 1952 threshold been indexed, R(t) would have held at $500 throughout; the gap between the indexed counterfactual and the realized R(t) is the accumulated real erosion, and by 2024 that gap is $500 − $42 = $458 of real value per unit of the threshold, or equivalently the nominal threshold would stand at $5,919 rather than $500. The failure to index was not a small omission with a small consequence. It was a decision—or a non-decision—whose cumulative effect is a nearly twelve-fold divergence between the rule’s intended real content and its realized real content. The Coasean account of why the correction was not made gains force from this magnitude: the benefit of correction, measured as the restoration of $458 of real threshold, is large in aggregate, but it is spread across every party to every affected transaction, none of whom captures enough of it to fund the concentrated cost of amending the Code across the states. The size of the foregone correction and the failure to undertake it are both consequences of the same structure—large diffuse benefit, concentrated cost, invisible drift—and the dynamics show the benefit growing every year the correction is deferred.
VI. What the threshold is, on the value axis
To see what the erosion means rather than merely that it happened, place the threshold in the framework this series has developed. The good-faith-purchase literature and the payments literature describe a single axis on which legal rules governing exchange are positioned by the value and identifiability of what is exchanged. The Statute of Frauds threshold is a rule on that axis, and it is worth stating exactly what kind of rule it is.
The Statute is an evidentiary switch keyed to value. Below the line, the law accepts oral proof of a bargain; above it, the law demands documentary proof. The drafters’ premise—stated in the official commentary to the section, which explains that the writing need only, in the official commentary’s phrase, “afford a basis for believing” that the oral evidence rests on a real transaction (UCC § 2-201, Official Comment 1)—is that the function of the writing is to lower the risk of enforcing a contract that never existed. The threshold is therefore a value at which the drafters judged that the expected social cost of fabricated-contract litigation begins to exceed the expected social cost of requiring documentation. Below the line, demanding a writing for every trivial trade would waste more than the fraud it prevents; above it, the fraud risk justifies the documentary burden. The line is the drafters’ estimate of where those two cost curves cross.
This is structurally the same problem the good-faith-purchase literature analyzes, and naming the parallel sharpens what erosion does. Schwartz and Scott (2011) showed that the law of good-faith purchase confronts a double moral hazard: the owner takes optimal precautions against theft only if she bears the loss, and the buyer investigates title optimally only if he bears the loss, and both cannot bear it at once. The Statute of Frauds addresses a cognate hazard on the contract-formation side: a party will fabricate a favorable oral bargain if fabrication is cheap and provable-looking, and the counterparty will guard against fabrication only if the law makes fabrication hard to prove. The writing requirement raises the cost of fabrication above the threshold, where the stakes make fabrication worth attempting. The threshold’s placement is a judgment about where the stakes become high enough that the fabrication hazard is live. That judgment was made in real terms—the drafters were thinking about what $500 meant in 1952—and inflation has silently relocated the line to a real value at which the original judgment no longer holds.
Levmore (1987), in his comparative study of the good-faith purchaser, drew a distinction that applies directly: legal rules that must control welfare-threatening behavior tend toward uniformity, while rules that either do not much matter or turn on judgments about which reasonable lawmakers disagree tend toward variety, and toward inertia. A dollar threshold in a fraud statute is exactly the kind of rule that attracts inertia. It is buried in a section few revisit, its erosion is invisible in the text, and no constituency is organized to move it. Levmore’s insight, ported forward, predicts precisely the pathology this essay documents: a rule whose real content is drifting will not be corrected if the drift is silent and the rule is obscure, because nothing in the legal system’s ordinary operation surfaces the drift for decision. The number stays because staying is the path of least resistance, and the real boundary moves as a side effect.
VII. Modeling the move
The erosion can be stated as a model, and stating it as a model exposes exactly which quantities are observed and which are inferred, so that no step hides an assumption. Let the threshold be a fixed nominal value H, set in a base year at H = $500. Let P(t) be the price level in year t, observed as the CPI-U annual average. The real value of the threshold in base-year purchasing power is
R(t) = H · P(t₀) / P(t)
where t₀ is the base year, 1952. This is not a behavioral model with estimated parameters; it is an identity. R(t₀) = H by construction, and R(t) is pinned entirely by the observed price ratio. There is no residual, no error term, no functional-form choice to defend, because nothing is being fitted: the real value of a fixed nominal number is definitionally its nominal value deflated by the price level. Every figure in Section III is R(t) evaluated at a verified P(t). The only inputs are H, which is read from the statute, and P(t), which is read from the BLS series. I emphasize that this is an identity rather than an estimate because it means the erosion claim carries no econometric vulnerability—there is nothing to misspecify.
What the identity does not by itself tell us is the consequence of the move, and here the model must be extended with care, because the consequence depends on the distribution of real transaction values in the economy, and that distribution is not something I have measured. I therefore state the extension as a conditional structure and mark precisely where an empirical claim would require data I have not verified.
Let F be the cumulative distribution of real transaction values in the economy, so that F(v) is the share of transactions with real value at or below v. The share of transactions caught by the writing requirement in year t is the share whose nominal value reaches H, which in real terms is the share at or above R(t): that share is 1 − F(R(t)). Because R(t) falls monotonically, 1 − F(R(t)) rises monotonically for any distribution F that places positive mass in the relevant range. This is the formal content of the claim that the threshold captures ever more transactions: it is true for any F with support spanning R(t), and it does not require knowing the shape of F. The direction of the effect is distribution-free.
The magnitude of the effect is not distribution-free, and I do not claim it. To say how many transactions the erosion has newly captured—to convert the fall in R(t) from $500 to $42 into a number or share of real trades pulled inside the writing requirement—would require F, the actual distribution of real transaction values, which I have not measured and which I will not assume. I can state the direction with certainty because it is distribution-free; I cannot state the magnitude without F, and I decline to invent F. A reader who wants the magnitude must supply a measured distribution of transaction values; the model then delivers the answer mechanically, but the answer is only as good as the measured F, and I have not measured it. This is the boundary between what the computation establishes and what it does not, and I draw it explicitly rather than let the reader assume the magnitude follows.
The model also clarifies what would arrest the move. A threshold indexed to the price level—H(t) = H₀ · P(t) / P(t₀)—holds R(t) constant at H₀ by construction, because the nominal figure then rises exactly as fast as the price level. Indexation is the null case in which the boundary does not move on the value axis. The choice the drafters made, whether by decision or by inattention, was to leave H fixed, which is the choice that makes R(t) decay. The two regimes—fixed nominal versus indexed—are the two available treatments of a value threshold under inflation, and the difference between them is the entire phenomenon. Nothing about the Statute’s purpose requires the fixed-nominal regime; the fraud-prevention rationale is a statement about real stakes, and real stakes are preserved only under indexation. The fixed-nominal regime is not a considered policy that stakes should fall; it is the absence of a policy, and its effect is decay.
VIII. Why the move is invisible, and why that matters
A striking feature of this erosion is that it is legally invisible. The text of § 2-201 reads the same in 2024 as in 1952; a lawyer consulting the section sees $500 and applies $500. Nothing in the operation of the rule surfaces the fact that $500 now means what $42 meant when the rule was written. The move happens entirely in the gap between nominal text and real value, a gap the legal system does not routinely inspect. This invisibility is not incidental to the phenomenon; it is the mechanism by which the phenomenon persists.
Coase (1937) supplies the frame for why invisibility matters institutionally. Coase explained that activity is organized inside firms rather than through markets when the cost of using the price mechanism—discovering prices, negotiating, contracting transaction by transaction—exceeds the cost of administrative coordination, and that “there is a cost of using the price mechanism” (Coase, 1937) that determines where that boundary falls. The correction of a legal threshold is itself an activity with a cost: someone must notice the drift, propose the amendment, move it through a drafting body and the legislatures of the adopting states, and absorb the friction of changing settled text. That cost is positive and, crucially, it is concentrated, while the benefit of correction is diffuse—spread across all the parties whose ordinary transactions are needlessly pulled inside the writing requirement, none of whom is individually harmed enough to organize. The Coasean structure predicts non-correction: when the cost of moving a rule is concentrated and the benefit is diffuse and the drift is invisible, the rule will not be moved, and the real boundary will continue to slide. The $500 figure is a fossil of this structure—a number that persists not because it is right but because the transaction cost of changing it exceeds any single actor’s incentive to bear it.
The invisibility also defeats the ordinary safeguard against bad rules, which is that their badness is felt and complained of. A threshold that erodes silently produces no salient moment of failure. There is no day on which $500 audibly becomes wrong; there is only the imperceptible annual descent of R(t), each year’s fall too small to notice, the cumulative fall enormous. This is the same compounding invisibility that makes inflation’s effect on any fixed nominal commitment—a bracket, a fee, a penalty, a threshold—so politically durable: no single year’s erosion crosses the threshold of attention, so the accumulated erosion is never confronted. The legal system has no thermostat for real value. It reads nominal text, and nominal text lies still while the ground moves.
IX. The disappearing premise: documentation cost in a recorded-payment world
The Statute of Frauds rests on a premise about the cost of documentation, and Kahn and Roberds (2009) supply the framework that shows that premise dissolving independently of inflation, which means the eroded threshold is being undermined from two directions at once. The premise is that below some transaction value, requiring a writing costs more than the fraud it prevents, because producing and retaining documentation is itself costly. That premise was sound in a world where small transactions were conducted in cash and left no record. It is decreasingly sound in the world Kahn and Roberds describe.
Their central dichotomy divides payment into store-of-value systems, which transfer an object and require verifying only the object, and account-based systems, which adjust ledger entries and require verifying identity and history. The store-of-value system—cash—produces no record; a cash sale is exactly the undocumented oral transaction the Statute’s low-value exemption was designed for. The account-based system produces a record as a byproduct of its operation; every card payment, every platform transfer, every bank-mediated settlement generates a durable, identity-linked, timestamped record without any party choosing to document anything. Kahn and Roberds observe that the account-based margin has been advancing for a century as the cost of record-keeping falls, and that the technologies of tracking and verification steadily expand the domain over which recorded payment displaces cash.
This advance dissolves the Statute’s premise from underneath. The writing requirement’s low-value exemption assumes that documenting a small transaction is costly enough to exempt it. But when the small transaction is settled through an account-based system, it is documented automatically and at zero marginal cost to the parties—the record is a free byproduct of the payment. For such transactions the premise of the exemption fails: documentation is not costly, so exempting small transactions from a writing requirement no longer saves a documentation cost, because the documentation exists regardless. The eroded threshold is thus caught in a scissors. From above, inflation drags the nominal line down into the range of small transactions. From below, the account-based payment system fills that same range with transactions that are already documented, making the writing requirement simultaneously more widely applicable and less necessary. The rule expands its reach into a domain where its own premise has evaporated.
The framework of this series predicts exactly this convergence. If what counts as a documented transaction is a moving category—shifting as payment technology shifts the informational properties of exchange—then a rule that partitions transactions by whether they are documented is a rule whose real operation changes as the payment system changes, independently of the rule’s text. The Statute of Frauds partitions transactions into the orally-proved and the writing-required. The payment system is quietly moving the whole population of small transactions from the first category to the second, not by legal command but by the mechanics of how payment now works. The threshold’s erosion by inflation and its obsolescence by recorded payment are two faces of one fact: a rule written for a cash world is being operated in a ledger world, and neither its nominal line nor its documentary premise survives the translation intact.
X. Capturing the value: what the erosion actually transfers
The threshold’s erosion does not merely misplace a line; it moves value, and naming what value moves and to whom is the substantive core of the phenomenon. When a transaction that would have been outside the writing requirement is pulled inside it, something is transferred: the party who benefits from the writing requirement gains, and the party burdened by it loses. Identifying those parties requires stating what the writing requirement does in a contested transaction.
The Statute of Frauds is, in operation, a defense. A party sued on an oral contract for goods above the threshold may raise § 2-201 to render the contract unenforceable. The requirement therefore systematically favors the party who wishes to escape an oral bargain and burdens the party who wishes to hold the other to it. In the paradigm case, a seller and buyer strike an oral deal; prices move; the party for whom the deal has become unfavorable invokes the absence of a writing to walk away. The writing requirement hands that party an exit. Below the threshold, no exit; the oral deal binds. Above it, the exit is available. As the threshold erodes, the exit becomes available in ever more transactions—transactions that, in real terms, are the small everyday bargains the drafters chose to leave binding.
This is where the framework of the prior essays does analytical work. In the good-faith-purchase problem, the law’s allocation of a loss between two innocent parties was shown to be, in a wide range of cases, behaviorally inert: because search and verification collapse for low-value or low-identifiability goods, the legal rule does not change conduct and merely relocates a loss. The eroded Statute of Frauds threshold exhibits the mirror pathology. The rule was designed to change conduct—to induce parties to documents consequential bargains—but as it descends into the range of ordinary small transactions, it stops inducing documentation and starts merely reallocating the gains from oral deals that parties will continue to strike orally, because the cost of documenting a small transaction exceeds its stakes. Below some real value, no rational party writes down a handshake sale; the transaction cost of the writing exceeds the value at risk. For those transactions, the eroded threshold does not produce writings. It produces a latent defense—an exit that sits unused until prices move and one party finds it worth invoking. The value captured is the value of that exit: the ability, conferred by the accident of inflation, to escape an ordinary oral bargain that the drafters intended to bind.
Who captures it is not fixed ex ante, and this is the feature that makes the erosion insidious rather than merely redistributive. Because either party may turn out to be the one who wants to escape, the eroded threshold does not systematically favor sellers over buyers or buyers over sellers. It favors, in each realized dispute, whichever party the price movement has made regretful, at the expense of the party relying on the deal. It is a randomly assigned option to renege, handed out by inflation to whichever side later wants it, and its social cost is the reliance it destroys: parties who struck ordinary oral bargains in good faith find those bargains unenforceable not because the law judged them unworthy of enforcement but because a number written in 1952 descended, unnoticed, into the range of their transaction. The erosion converts a considered evidentiary rule into a lottery of enforceability over the small commerce of daily life.
The option-to-renege framing can be made precise without overclaiming. In each affected transaction the eroded threshold confers, on whichever party the price movement later disfavors, an option whose payoff is the ability to avoid an unfavorable oral bargain by invoking the absence of a writing. The option is created by the accident that the transaction’s nominal value reached $500 while its real value—$42 in 1952 terms—is one the drafters intended to leave binding. I do not claim to price this option, because pricing it would require the distribution of price movements across affected contracts and the rate at which the defense is invoked, neither of which I have measured; naming a value would be invention. What the framework does establish, distribution-free, is that the option exists wherever R(t) has fallen below a transaction’s real value, that the population of transactions carrying it grows monotonically as R(t) falls, and that its social cost is borne as destroyed reliance by the counterparties who structured their affairs around bargains they reasonably believed enforceable. The direction and the incidence are established; the price is not, and I mark it unmeasured rather than guess it.
XI. What indexation would and would not fix
The obvious remedy is indexation: tie H to the CPI-U, restore R(t) to its 1952 real value, and hold it there. The model shows indexation does exactly that—H(t) = $500 · P(t)/P(1952) holds the real threshold constant—and the 2003 revision of Article 2 in fact proposed raising the figure to $5,000, which is close to the $5,919 that full CPI-U indexation to 2024 would require. But the 2003 revision was not adopted by the states and was later withdrawn, so the operative figure remains $500. That non-adoption is itself the Coasean prediction realized: even a drafted, available correction failed to move through the concentrated-cost, diffuse-benefit gauntlet. Indexation is the right structural fix and its history is a case study in why structural fixes to invisible erosion do not pass.
Indexation would fix the direction of the move but would not resolve a deeper question the erosion exposes, and honesty requires stating the limit of the remedy. Indexation preserves the real value the drafters chose in 1952. It does not establish that the 1952 choice was correct, then or now. The drafters’ placement of the line at $500-in-1952-dollars was itself a judgment about where the fabrication hazard becomes live, made with 1952 information about commercial practice, litigation cost, and the reliability of oral proof. Those underlying conditions have changed in ways inflation indexation does not touch: documentation is now nearly costless for many transactions, because electronic records attach automatically to card and platform payments; oral-only bargains are rarer; and the evidentiary landscape the Statute was built for—a world of undocumented handshake deals—has contracted. A correctly indexed threshold answers the question “what did $500 mean in real terms” but not the question “is a value-keyed writing requirement still the right instrument at all, given that most transactions now document themselves.” The erosion, by dragging the threshold into visibility, forces that second question, and indexation alone does not answer it.
This connects to the through-line of the entire series. The payments literature shows that the informational properties of exchange—what is recorded, what is verifiable, what carries identity—are shifting as the technology of payment shifts, and that what counts as money and what counts as a documented transaction are moving categories. The Statute of Frauds threshold was written for a payment world in which small transactions were oral and undocumented by default. In a payment world where small transactions are electronically recorded by default, the premise of a value-keyed writing requirement—that below some value, demanding documentation costs more than the fraud it prevents—is undermined not by inflation but by the collapse of the cost of documentation. Indexation corrects the inflation drift. It does not address the technological obsolescence of the rule’s premise. Both are real; they are different; and conflating them would be its own error.
XII. The general pathology: fixed nominal boundaries in a moving unit of account
The $500 threshold is one instance of a general pathology, and stating the general form both locates the specific case and shows why it recurs. Any legal rule that draws a boundary in fixed nominal units, in a system whose unit of account inflates, has a real content that decays. The Statute of Frauds threshold is a clean instance because its structure is simple—a single number, a single switch—but the pathology is everywhere fixed dollar figures appear in durable law: jurisdictional amounts, penalty ceilings, reporting thresholds, exemption levels, definitional cutoffs. Each is a boundary on a value axis, fixed in nominal terms, eroding in real terms, invisibly.
The framework of this series identifies why the pathology is structural rather than accidental. Legal rules governing exchange are positioned on an axis of value and identifiability; a fixed nominal boundary is a position on that axis specified in a unit that does not hold still. The specification is therefore self-undermining: the text names a position, but the position the text names moves relative to the real distribution the rule is meant to partition. This is not a drafting error correctable by better drafting within the fixed-nominal form; it is a property of expressing a real boundary in nominal units. The only drafting that avoids it is drafting that indexes—that specifies the boundary in real terms and lets the nominal figure track the price level. Fixed-nominal specification guarantees drift; the guarantee is arithmetic.
Coase’s transaction-cost logic explains why the pathology persists despite being, once stated, obvious. Correcting any single eroded threshold has a concentrated cost and a diffuse benefit, so no single threshold attracts the effort needed to correct it. And because the erosion is invisible in the text, no threshold announces its own need for correction. The result is a legal system carrying a large stock of fixed nominal boundaries, each silently drifting, none individually salient enough to fix, the aggregate drift substantial. The $500 Statute of Frauds figure is not an outlier to be explained; it is a representative member of a population of eroding nominal boundaries, and its 91.6% real decline is a measurement of what the population is doing.
The level-invariance of the decay deserves a concrete statement, because it is what makes the pathology general rather than specific to $500. The proportional erosion of any fixed nominal threshold set in 1952 is identical regardless of its level: a $500 threshold, a $5,000 threshold, and a $50,000 threshold have each lost the same 91.6% of their real value by 2024, because each is multiplied by the same price ratio 26.5/313.689. The dollar amounts of real value destroyed differ—$458, $4,576, and $45,758 per unit respectively—but the fraction is one number for all of them, fixed by the CPI-U ratio and nothing else. This is why the pathology cannot be escaped by choosing a higher threshold: the fraction lost is set by inflation, not by the level chosen, so every fixed nominal boundary in the 1952 statute book that has never been amended has lost the identical 91.6% of its real content, whatever its nominal size. The Statute of Frauds figure is not eroding faster or slower than its nominal-boundary siblings; it is eroding at the rate the price level dictates for all of them simultaneously. A legal system that fixes boundaries in nominal units and inflates its unit of account is a system in which every such boundary loses real content in lockstep, the lockstep governed by a single ratio the system does not monitor.
XIII. What is established, and what is not
Precision about the status of each claim is itself part of the argument, because a claim asserted beyond its support would undermine the claims that are fully supported. I therefore separate what this essay establishes from what it does not.
Established, by verified primary data and arithmetic. That the UCC § 2-201 threshold is $500 and dates from the 1952 Official Text (verified against the statutory text at Cornell LII). That the CPI-U annual average rose from 26.5 in 1952 to 313.689 in 2024 (verified against the BLS series). That the real value of the fixed $500 threshold has therefore fallen by 91.6%, to $42.24 in 1952 purchasing power, and that holding it constant would require $5,918.66 in 2024. That the direction of the resulting expansion of the writing requirement’s coverage—more transactions caught as R(t) falls—is monotone and distribution-free. These are not estimates and carry no econometric vulnerability; they are the arithmetic of read numbers.
Not established, and explicitly not claimed. The magnitude of the coverage expansion—how many or what share of real transactions the erosion has newly pulled inside the writing requirement—is not established, because it requires the distribution F of real transaction values, which I have not measured and will not assume. The behavioral response of parties to the eroded threshold—whether and how often the latent defense is actually invoked—is not established, because it requires litigation data I have not gathered. The welfare consequence—whether the erosion’s destruction of reliance exceeds any benefit—is not established, because it requires both the distribution and the behavioral response. I state the direction of each of these with whatever certainty the framework supports and decline to state magnitudes I have not measured. The line between the two is drawn deliberately: the quantified core is bulletproof because it is arithmetic on verified data, and everything beyond it is marked as beyond it.
Interpretive, offered as reading rather than proof. The claim that the erosion converts a considered evidentiary rule into a lottery of enforceability, that the value it captures is the option to renege, and that the Coasean structure explains non-correction—these are interpretations built on the established facts and the cited literature. They are argued, not demonstrated in the sense the arithmetic is demonstrated, and I present them as the most coherent reading of the established facts, not as further measured results. A reader may accept the arithmetic and dispute the interpretation; the arithmetic does not depend on the interpretation, and I have kept them separable so that the dispute, if any, is confined to the interpretive layer.
XIV. Conclusion: the number that moved by standing still
A statute fixed a number at $500 in 1952 and never changed it. The number is the same today. But a number in a fixed unit is not a fixed thing when the unit inflates, and the CPI-U, the official measure of that inflation, records that the unit lost roughly eleven-twelfths of its value over the intervening seventy-two years. The threshold, standing still in nominal terms, moved in real terms from $500 to $42, and in moving it dragged the boundary of the writing requirement down through the distribution of transaction values, from the consequential bargains it was meant to govern into the ordinary small commerce it was meant to leave alone. The move is invisible because the text does not change; it is unstopped because the cost of stopping it is concentrated and its benefit diffuse; and it is representative, because the legal system carries a large stock of such fixed nominal boundaries, each drifting the same silent way.
The framework of this series locates the phenomenon exactly: a legal boundary on the value axis, specified in a unit that does not hold still, is a boundary that moves relative to the real distribution it partitions, and its motion is governed by arithmetic no less certain than the arithmetic that governs money’s finality or the good-faith purchaser’s collapse of search. What the erosion of the $500 threshold captures is reliance—the enforceability of ordinary oral bargains, silently withdrawn from the parties who struck them. What it reveals is that a rule expressed in nominal terms is a rule whose real content is written by the price level, not by the drafters, and that the legal system’s failure to index is not a neutral default but an active, if unintended, policy of decay. The number did not have to move to do this. It did it by standing still.
Primary sources
Uniform Commercial Code § 2-201, Formal Requirements; Statute of Frauds (Official Text). Legal Information Institute, Cornell Law School. https://www.law.cornell.edu/ucc/2/2-201
U.S. Bureau of Labor Statistics. Consumer Price Index for All Urban Consumers (CPI-U): U.S. city average, all items, annual averages (series CUUR0000SA0, 1982–84 = 100). Historical CPI-U table. https://www.bls.gov/cpi/tables/supplemental-files/historical-cpi-u-202412.pdf
References
Coase, R. H. (1937). The nature of the firm. Economica, 4(16), 386–405. https://onlinelibrary.wiley.com/doi/10.1111/j.1468-0335.1937.tb00002.x
Crawford, M. J. R. (2025). The riddle of the good faith purchaser. Oxford Journal of Legal Studies, 45(1), 167–192. https://academic.oup.com/ojls/article/45/1/167/7900629
Fox, D. (1996). Bona fide purchase and the currency of money. Cambridge Law Journal, 55(3), 547–565. https://www.cambridge.org/core/journals/cambridge-law-journal/article/abs/bona-fide-purchase-and-the-currency-of-money/86BB50EE4BCAC58064A13BA07C3F1969
Kahn, C. M., & Roberds, W. (2009). Why pay? An introduction to payments economics. Journal of Financial Intermediation, 18(1), 1–23. https://www.sciencedirect.com/science/article/abs/pii/S1042957308000533
Levmore, S. (1987). Variety and uniformity in the treatment of the good-faith purchaser. Journal of Legal Studies, 16(1), 43–65. https://www.journals.uchicago.edu/doi/abs/10.1086/467823
Schwartz, A., & Scott, R. E. (2011). Rethinking the laws of good faith purchase. Columbia Law Review, 111(6), 1332. https://scholarship.law.columbia.edu/faculty_scholarship/187/