Anchoring speculation to the metal in the vault

Cap every speculator's open exposure at a fixed multiple of the physical inventory it verifiably holds. Speculators become storers — and larger stocks mean fewer, smaller price spikes.

Net open speculative exposure in a storable commodity should be capped at a fixed multiple ρ ≥ 1 of the verified physical inventory held by the same participant. The rule joins two traditions: the theory of storage, which predicts and empirically supports a negative relationship between inventory and spot-price volatility (Working 1949; Symeonidis et al. 2012; Geman and Smith 2013), and the real-bills doctrine, which held that claims should rest on the goods they represent (Smith 1776; Mints 1945). Under the rule, aggregate speculative open interest is bounded by ρ times verified stocks, and the upper tail of volatility generated by claim expansion is truncated. A direct volatility effect, separate from the inventory channel, requires that variance depend on speculative exposure conditional on inventory, a condition the storage literature does not establish and the speculation literature contests (Irwin and Sanders 2011; Cheng and Xiong 2014). The rule raises the capital required for an unbacked position by a factor of 1 + 1/(ρm), where m is the initial-margin ratio; at margins of 5–15% and ρ ∈ [1,5], this ranges from roughly 2.3× to 21×. The rule has never been applied. Applied to COMEX as of October 2026, at ρ = 1 it would require managed money to hold 72–96% of all deliverable gold ($45–59 billion), and the present silver book exceeds deliverable silver outright. At ρ = 5, the figures are $9–12 billion of gold and up to $1.5 billion of silver. These sums are not a cost in the ordinary sense: they move capital from paper claims into physical stocks, turning speculators into storers. Because larger stocks mean fewer and smaller price spikes, the rule's benefit does not depend on whether speculation itself moves prices, only on whether the metal speculators must hold adds to the stock of the good.

Introduction

Commodity spot-price volatility has two conceptually distinct sources. The first is change in the quantity of the physical good: harvests, outages, demand shocks, and the inventory adjustments they induce. The second is change in the volume of financial claims written against the good, which in derivatives markets is not mechanically bounded by the quantity of the good itself. A regulatory constraint can bind the second source to the first without suppressing the information the first carries.

The theory of storage gives the first source a precise structure. Inventory buffers shocks: when stocks are ample, a shock is absorbed by drawing down or adding to storage, and prices move little. As stocks approach zero, storage can no longer smooth, convenience yields rise, the futures curve moves into backwardation, and spot prices become volatile (Working 1949; Brennan 1958; Deaton and Laroque 1992). This relationship is among the most robust in empirical commodity economics.

The second source is contested. Whether financial participation in commodity futures amplifies spot volatility has been debated intensely since the 2007–2008 price spike, with influential studies reaching opposing conclusions (Masters 2008; Irwin and Sanders 2011; Tang and Xiong 2012; Cheng and Xiong 2014). The rule does not depend on settling that debate. If claim expansion against fixed inventory contributes to volatility, the narrowest remedy removes that contribution and leaves inventory-driven volatility intact.

Fig. 1 — The rule binds the second source to the first.

The real-bills doctrine held that bank credit should expand only against short-term commercial paper representing goods in production or sale (Smith 1776). The doctrine failed as a monetary rule for reasons established by Mints (1945). One narrow component survives those critiques: the requirement that a claim rest on a verifiable quantity of the good it references. The rule applies that component not to money but to residual speculative exposure in a single commodity.

The resulting rule requires that a participant's net open speculative exposure not exceed ρ times that participant's verified inventory, with bona fide hedges exempt.

Related Literature

2.1 The theory of storage

Kaldor (1939) and Working (1949) established that the price of storage—the spread between futures and spot prices—is a function of inventory. Working's supply-of-storage curve implies that the spread approaches full carry when inventories are large and can become negative when they are small, producing backwardation. Brennan (1958) formalised the convenience yield as a declining, convex function of stocks, and Fama and French (1987) found that the basis behaves as the theory predicts across a broad set of commodities.

Because inventory cannot be negative, storage smooths prices only when stocks are positive; a stock-out removes the buffer and transmits shocks to price in full. Deaton and Laroque (1992, 1996) showed that a competitive storage model generates the skewness and occasional explosive spikes observed in commodity prices. Cafiero et al. (2011) demonstrated that a finer-grid estimation of the same model matches observed price autocorrelation, resolving an earlier empirical difficulty.

Direct tests confirm the inventory–volatility link. Using physical inventory data on 21 commodities over 1993–2011, Symeonidis et al. (2012) find that volatility declines with inventory for the majority of markets, more strongly when the curve is in backwardation, and that the result is robust to alternative inventory measures and to the 2000s commodity boom. Geman and Smith (2013) obtain the same result for the six LME base metals (aluminium, copper, lead, nickel, tin, and zinc), and find that adding Shanghai Futures Exchange inventory strengthens the relationship. Further, Gorton et al. (2013) show that inventory levels predict both the basis and futures risk premia. The relation σ² = f(I), with f′ < 0, is established.

2.2 Financial speculation and commodity prices

The rule targets the second source of volatility, and here the evidence divides. Masters (2008) argued in congressional testimony that index investment inflated prices in 2007–2008. Tang and Xiong (2012) documented that commodity returns became more correlated with one another and with equities as index investment grew, consistent with a financialisation channel. Against this, Irwin and Sanders (2011) found little evidence that index positions predicted futures returns or volatility; Büyükşahin and Harris (2011) found that speculative positions in crude oil followed rather than led prices; and Hamilton (2009) attributed the 2007–2008 oil price run-up chiefly to strong demand meeting stagnant production, while Kilian and Murphy (2014) likewise find a primary role for fundamental demand rather than speculative trading. Cheng and Xiong (2014) conclude that financial trading affects commodity markets mainly through risk sharing and information, with effects on price levels harder to establish.

A futures contract does not itself remove a physical unit from the market. For pure claim expansion to move spot prices, it must affect inventory decisions, convenience yields, or the information on which physical traders act (Pindyck 2001). The rule is therefore a constraint on a channel of uncertain magnitude, not one already shown to be large.

2.3 The real-bills doctrine

Smith (1776, Book II, Ch. 2) argued that banks discounting real bills—claims arising from completed sales of goods—would issue neither too much nor too little paper. Thornton (1802) identified the central flaw: because the nominal value of bills rises with prices, credit indexed to bills can expand without limit in an inflation. Mints (1945), who coined the term, judged the doctrine wholly mistaken: a bank confined to discounting real bills, but free to set the discount rate, can alter the purchasing power of its notes without limit, leaving the price level indeterminate. Because the quantity of eligible paper rises with prices, the constraint is endogenous and does not bind. Humphrey (1982) traces the doctrine's influence on Federal Reserve policy. Sargent and Wallace (1982) offered a qualified rehabilitation. In economies where the quantity theorists' predictions hold, the price level fluctuates more under a real-bills regime of unrestricted intermediation, yet the Pareto criterion does not support the legal restrictions the quantity theory recommends.

“The constraint is measured in physical units, not nominal value, so Thornton's inflationary feedback cannot arise.”

The rule borrows only the narrow principle that a claim should rest on the goods it references, and avoids the elements Mints criticised. The constraint is measured in physical units, not nominal value, so Thornton's inflationary feedback cannot arise. And the rule does not classify claims by purpose; it uses the existing regulatory distinction between bona fide hedging and other positions, and constrains only the latter.

2.4 Position limits and covering requirements

U.S. federal position limits are set by the CFTC final rule published on 14 January 2021 and effective 15 March 2021, which extended federal coverage from 9 to 25 core referenced futures contracts. Spot-month limits are set at or below 25% of estimated deliverable supply; non-spot-month federal limits apply only to the nine legacy agricultural contracts; bona fide hedges are exempt (CFTC 2021). In every case the limit is a market-wide quantity applied identically to each participant, not a function of the participant's own holdings. Position limits under Article 57 of MiFID II (European Union 2014) follow the same structure, with baselines of 25% of deliverable supply in the spot month and 25% of open interest in other months; Directive (EU) 2021/338 has since confined them to agricultural and significant or critical contracts (European Union 2021).

The closer precedents lie outside commodities. In U.S. equities, Rule 203(b) of Regulation SHO requires a broker to locate borrowable shares before effecting a short sale, tying a short claim to an identifiable underlying security (SEC 2004). In the European Union, Regulation (EU) No 236/2012 permits a sovereign credit default swap position only where the holder has exposure to the underlying debt or a correlated risk (European Union 2012). Both condition a derivative position on the holder's relationship to the underlying, which is the logic extended here. No commodity regime caps speculative exposure as a participant-specific multiple of that participant's verified physical inventory. Hedge exemptions are participant-specific but run in the opposite direction: they relax limits for those with physical exposure rather than make physical holdings a precondition for speculation.

2.5 From delivery to differences: how the line moved, 1905–2000

The distinction the rule draws is old. In Board of Trade v. Christie Grain (1905), the Supreme Court upheld futures traded on the Chicago Board of Trade against the charge that they were wagers. Justice Holmes held that contracts settled by paying differences “stand on different ground from purchases made merely with the expectation” of set-off, because the parties remain bound to deliver: “A set-off is, in legal effect, a delivery.” The bucket shops the Court distinguished were places of “the pretended buying and selling of grain” with no “intention of receiving and paying for the property so bought.” The line ran between a binding claim on the good and a wager on its price. State gambling and bucket-shop laws enforced it.

Fig. 2 — Sources: Christie Grain, 198 U.S. 236; CFTC Act of 1974; Purcell and Valdez, National Agricultural Law Center; Blake 2010, Institutional Investor; CFMA of 2000; CRS RS20560.

Taken together, these measures removed the state-law backstop that once separated claims from wagers, and let financial price exposure qualify as hedging. Index swap assets rose from about $50 billion in 2004 to $125 billion at the end of 2007 (Blake 2010). The rule restores the older line by a different route: it defines the right to speculate by physical holdings, as Christie Grain defined a legitimate future by the obligation to deliver.

The Rule

3.1 Statement

Consider a storable commodity c. Index participants by i ∈ N and dates by t. Let I i,t ≥ 0 be the quantity of c, in physical units, held by i at t, verified under Section 3.4 and unencumbered by any other position. Let E i,t be the net open speculative exposure of i in c, in physical-unit equivalents, across futures, delta-adjusted options, swaps, and other referenced contracts, excluding positions that qualify as bona fide hedges. The rule requires, for all i and t,

The absolute value is essential: a net short (E i,t < 0) would otherwise satisfy the constraint at any inventory, including zero, leaving short speculation unbound. The floor ρ ≥ 1 ensures that a participant may always speculate up to the full quantity it holds; a multiple below one would require collateral in excess of full physical backing.

3.2 Exposure and exemptions

Exposure is measured net and delta-adjusted across all referenced contracts in the same commodity, and aggregated across affiliated entities under existing CFTC aggregation standards. Bona fide hedges—positions offsetting price risk on physical inventory, anticipated production, merchandising commitments, or anticipated requirements—are excluded from E i,t . Commercial hedging is untouched; only residual speculative exposure is bound.

The interaction between the exemption and the constraint produces an asymmetry. A participant holding a speculative long of size E must hold at least E/ρ units of verified inventory; the inventory backs the long rather than offsetting it, so a backed speculative long is doubly long in price risk. On the short side, a short held against one's own inventory is ordinarily a bona fide hedge and exempt, while a participant without inventory may hold no speculative short at all. The rule therefore operates on the short side as a prohibition on uncovered shorts, closely analogous to the Regulation SHO locate requirement.

3.3 Eligible inventory

Inventory counts toward I i,t only if three conditions hold.

The single-use condition underpins the aggregate bound in Proposition 1. If a warrant could be lent and counted by both lender and borrower, admissible exposure would no longer be tied to physical stocks.

3.4 Verification and enforcement

Verification relies on existing infrastructure: exchange-registered warehouse receipts, LME warrants, approved delivery points, and third-party custodial statements. Inventory held outside approved facilities counts only on audited attestation at a frequency the exchange sets. A central registry keyed to unique receipt identifiers enforces single use.

A participant whose exposure exceeds ρI i,t —because inventory falls, exposure rises, or a hedge exemption lapses—must reduce exposure or acquire and verify additional inventory within a cure period. Persistent breaches are handled as position-limit violations are today. Because the bound moves with the participant's own inventory, the rule is self-adjusting: a participant that builds stocks gains speculative capacity automatically.

Analytical Framework

Let V t denote total verified, unencumbered inventory of c at t, and H t open interest attributable to bona fide hedges.

4.1 Aggregate bound

The bound is attained only if every holder of verified inventory exhausts its speculative capacity. Because V t counts only verified, unencumbered stocks, it is generally smaller than total physical inventory, and the bound is correspondingly tight.

Corollary 1. Holding V t fixed, aggregate speculative exposure cannot rise above ρV t . Any expansion of speculative claims beyond that level requires a prior increase in verified inventory.

4.2 Volatility

If spot variance depended on inventory alone, σ t ² = f(I t ), restricting E at a given I would leave variance unchanged. A direct volatility effect therefore requires an additional condition.

The first condition is the storage result of Section 2.1. The second is the contested claim of Section 2.2; neither Symeonidis et al. (2012) nor Geman and Smith (2013) test it, because neither conditions on speculative exposure.

The upper tail of volatility generated by claim expansion against fixed inventory is truncated precisely when Assumption A holds. Remaining volatility tracks changes in inventory and in underlying supply and demand.

Indirect channel. The rule can lower volatility even if ∂g/∂E = 0, through inventory. Participants wishing to speculate must hold stocks, raising private demand for storage. If aggregate inventory rises from I₀ to I₁ > I₀, variance falls from f(I₀) to f(I₁) by the storage result alone. The channel is not automatic: in competitive equilibrium aggregate stocks are set by intertemporal arbitrage, and inventory held by speculators to satisfy the rule may displace inventory commercial storers would otherwise hold. The indirect effect is positive to the extent that crowding-out is incomplete.

4.3 Capital required

The multiplier ranges from 2.3× to 21× and declines in both ρ and m, but remains substantial throughout: at ρ = 3 it still lies between 3.2× and 7.7×. Capital committed to inventory is not consumed; it is held in an asset that earns the convenience yield, incurs storage cost, and bears price risk. What the holder gives up is the opportunity cost of that capital plus net storage cost.

Corollary 2. A participant already holding verified inventory I ≥ E/ρ needs no additional capital.

What the rule would do today

No exchange or regulator has ever applied this rule, and there is no estimate of its effects. What it would demand of the market as it stands can be computed for COMEX gold and silver, where exchange inventory and positioning are both published, from the CFTC Disaggregated Commitments of Traders report of 13 October 2026 (CFTC 2026), the CME warehouse report of 14 October 2026 (CME Group 2026), and spot prices of 16 October 2026 (gold $4,168/oz, silver $57.84/oz).

5.1 Method

The CFTC managed money category is treated as speculative exposure, and managed-money traders are assumed to hold no verified exchange inventory. Funds of this kind do not ordinarily hold warrants, but the assumption cannot be checked from public data. Per-trader positions are not published, so ∑ i |E i,t | is bracketed: the category's net position is a lower bound, and its gross position (long plus short, excluding spreads) an upper bound.

Total open interest: gold 396,345 contracts, silver 108,433. Sources: CFTC Disaggregated COT, 13 Oct 2026; CME depository statistics via The Vault Report, 14 Oct 2026; prices from Fortune (gold, silver), 16 Oct 2026.

5.2 Results

The minimum feasible multiple is the binding fact. By Proposition 1, aggregate speculative exposure cannot exceed ρ times verified stocks. The gross managed-money book in silver is 129.9 million ounces against 103.2 million registered. At ρ = 1, the current silver book could not be carried even if speculators owned every deliverable ounce on COMEX; it must shrink by at least a fifth. In gold, the gross book equals 96% of registered stocks: at ρ = 1, it fits only if speculators own almost all of it.

A participant holding no inventory must acquire |E|/ρ ounces to keep a position of size E.

Black rule = 100% of registered stocks. Bars run from the net (lower) to the gross (upper) managed-money position. Silver at ρ = 1 crosses the line.

At ρ = 1, hedge funds and CTAs in COMEX gold would hold between $45 billion and $59 billion of metal—between 72% and 96% of all deliverable gold on the exchange—in place of positions now carried on a few percent of margin. In silver, the upper bound exceeds the deliverable stock outright. At ρ = 5, the figures are $9–12 billion of gold and $0.3–1.5 billion of silver.

5.3 What the numbers mean

These sums are not spent. They are moved. Capital that today holds a claim on gold would instead hold gold. The rule turns speculators into storers, and storage is the activity the literature of Section 2.1 identifies as the market's buffer against shocks. Kaldor (1939) and Working (1949) treated the speculator who carries stocks as the agent of intertemporal smoothing; the rule makes every speculator that agent. Speculation does not disappear. It must rest on the good.

The adjustment runs through price. New demand for metal raises spot prices while positions are rebuilt, and the higher price is the signal that draws metal from outside the exchange, from recycling, and from new production into verified storage. Verified stocks grow until the rule is satisfied. Where the metal comes from decides the long-run result. Metal drawn into storage from consumption or new supply raises the stock of the good available to absorb future shocks. Metal merely relabelled from commercial to speculative holders does not. Silver at ρ = 1 marks the limit: the present book exceeds deliverable stocks, so either the book shrinks or the stocks grow.

When the funds exit, the stocks don't vanish with them; the metal stays in the warehouse.

A fabricator who used to rebid every job when copper jumped on futures now locks the metal for the quarter. Before, a quote on a switchgear job carrying $65,000 of copper and $25,000 of profit could lose a quarter of that profit to a 10 percent move driven by fund buying, with not a pound of copper leaving the warehouse. So every job was rebid, and every bid padded. Under the rule, the funds that drove those moves hold copper under title. The fabricator buys the quarter's copper once, prices the job once, and stops charging customers for the risk.

5.4 Long-run effects

Constrained participants respond along three margins: reducing speculative exposure, acquiring and verifying inventory, or migrating to unregulated venues or substitute commodities. The second margin connects to competitive storage. If crowding-out is incomplete, aggregate inventory rises, and competitive storage models imply that higher mean inventory reduces the frequency and amplitude of price spikes because stock-outs become rarer (Deaton and Laroque 1992, 1996; Cafiero et al. 2011; Wright 2011). The function f is unchanged; the market spends more time in its low-volatility region.

First, the ratio of speculative open interest to verified stocks falls and is bounded above by ρ (Proposition 1). Second, verified inventory rises among constrained participants, while the change in aggregate inventory depends on crowding-out. Third, under Assumption A, the contribution of claim expansion to volatility falls, most strongly when inventory is low and the constraint binds (Proposition 2). Fourth, speculative liquidity declines when the constraint binds, and because binding is most likely when inventory is low, the decline coincides with periods in which hedgers most need counterparties.

Whether speculation itself moves spot prices is unsettled. Studies of speculative flows under current rules mostly find little direct effect (Irwin and Sanders 2011; Büyükşahin and Harris 2011), but none tests a rule of this kind. The inventory channel does not depend on that question. If the rule raises aggregate stocks, volatility falls by the storage result alone.

5.5 Crude oil

The gap between claims and goods is far wider in oil than in metals. NYMEX WTI open interest on 13 October 2026 was 1,900,179 contracts of 1,000 barrels each, or 1.90 billion barrels (CFTC 2026). Crude stocks at Cushing, Oklahoma, the contract's delivery point, were about 26.6 million barrels in the week ending 9 October (EIA 2026). Paper claims stand at roughly 71 times the oil at the delivery hub, against 2.7 times for gold and 5.3 times for silver.

Each square ≈ 26.6M bbl, one Cushing. Sources: CFTC Disaggregated COT, petroleum, 13 Oct 2026; EIA weekly data via Oklahoma Energy Today and Rigzone, week ending 9 Oct 2026. NYMEX only, not ICE WTI.

Measured against the delivery hub, the managed-money net position alone exceeds every barrel at Cushing, and the gross book would require ρ above 11. Measured against all U.S. commercial crude, the gross book would fit at ρ = 1, but most of that oil is neither at Cushing nor committed to delivery. Oil is therefore the market in which the rule would bind hardest and in which the distance between claims and deliverable goods is greatest.

Oil also supplies the clearest recent evidence of what that distance costs. On 20 April 2020, the May WTI contract fell about $55 in a day and settled at −$37.63 a barrel. Holders of long positions who could neither store nor take delivery had to sell before expiry as storage at Cushing filled. The CFTC staff report on the episode recorded open interest of about 630,000 contracts, against 400,000 to 500,000 at the previous year's peak, with the peak arriving only about ten days before expiry; it set out the facts without drawing conclusions (CFTC 2020). A participant required to hold verified, delivery-conforming inventory against its position has, by construction, arranged storage for it and cannot be forced to sell for lack of it. Whether the rule would have prevented the collapse cannot be shown, but it would have removed the class of position that drove it: long claims held by participants with no means of receiving the oil.

The contingency was insurance against moves in paper, not oil.

A trucking dispatcher stops adding a 15 percent contingency to every diesel quote. Fuel is about a fifth of a carrier's cost per mile, and truckload carriers run on margins under 1 percent, so a diesel spike that never touched the barrels at Cushing could wipe out a quarter's earnings. When every claim on crude rests on crude held in storage, diesel stops spiking on positioning, and the padding comes off every load.

Why Fewer Spikes Matter

The rule's purpose is fewer price spikes, and the gain from fewer spikes extends well beyond commodity markets. Smoother price paths produce a more stable price level, slower and less volatile inflation, and greater economic predictability. The result is not merely a quieter statistical series. Relative prices convey clearer information about scarcity and opportunity, households and firms can plan without hedging against abrupt cost jumps, and the real value of wages and savings is less exposed to unexpected erosion.

6.1 Stability as an outcome

Monthly price changes always contain noise from weather, seasonal patterns, one-off disruptions, and measurement. When those movements remain small and largely offsetting, the underlying rate of inflation stays low and steady. Businesses can commit to multi-year investment plans and contracts without large contingency margins for input-cost surges, and consumers can treat wage increases as real gains rather than as offsets against the next spike. Lower uncertainty reduces the risk premium attached to long-lived projects and to household balance sheets. Frequent spikes reverse each of these effects: they erode the real value of cash and fixed-income savings, force repeated renegotiation of wages and contracts, and shorten the planning horizon.

The two principal sources of smoother price paths are productivity growth and supply stability. On the first, Marto and McCollum (2026) study 85 U.S. industries from 2006 to mid-2025. After removing industry and time effects, they find that relative producer-price inflation falls to 3.7 percentage points below the cross-industry average four quarters after the onset of a productivity boom and stays below it for about six quarters. Across all industry-quarters, a one-percentage-point increase in residual productivity growth is associated with a 0.22-point decline in relative producer-price inflation. Their cases include computer and electronic products around the arrival of the smartphone, data processing during the adoption of cloud computing, mining during the shale-drilling efficiency gains, and health care during the adoption of electronic health records; in each, relative producer-price inflation fell by between 1.2 and 6.7 points below the average. The long-run pattern across the consumer basket is consistent. Over roughly four decades, prices of goods such as apparel, toys, car parts, and computer equipment have risen less than the CPI, reflecting quality improvements and manufacturing productivity, while prices of services with limited scope for productivity gains, such as education and medical care, have risen more (Federal Reserve Bank of St. Louis 2025).

6.2 Stability as a stabiliser

Supply stability is the channel the rule addresses. Shortages and disruptions produce the large positive jumps that appear when capacity or stocks are temporarily exhausted, and in commodity markets these are precisely the stock-out episodes of Section 2.1. Recent microevidence shows why such jumps matter beyond their own size. Firms adjust prices in a state-dependent way: the probability of a price change rises with the gap between a firm's price and its optimal price. In Belgian manufacturing microdata, the quarterly frequency of price adjustment was stable at about 0.29 before the pandemic and peaked at about 0.63 in 2022, a surge driven mainly by intermediate input costs, and cost pass-through was markedly stronger for firms facing large cost gaps than for those facing small ones (Gagliardone et al. 2026). Across advanced economies, the frequency of price changes more than doubled during the recent inflation surge, and state-dependent pricing implies a steeper Phillips curve when inflation is high (Karadi et al. 2025).

The implication is that small, frequent input-cost movements are absorbed with limited pass-through, while large ones trigger widespread repricing and propagate. High and volatile inflation thus makes the economy more sensitive to the next cost shock: firms reprice more often, and a given cost disturbance passes more fully and more quickly into inflation. Low and stable inflation is therefore not only an outcome but a stabiliser. The 1970s and early 1980s supply the historical demonstration, as cost shocks that might have been transitory under low inflation produced large, self-reinforcing swings.

6.3 The link to the rule

Commodity inputs are among the most common sources of the large cost shocks that tip firms into widespread repricing. By raising the stock of the good held against claims on it, the rule shrinks the shocks that reach downstream prices and keeps more of them within the range absorbed without general repricing. Its aim is not a one-time change in the price level. A one-time gain in productivity or supply lowers marginal cost and places downward pressure on prices, after which inflation returns to its previous path. A persistent reduction in the frequency and size of input-price spikes instead lowers the volatility of inflation itself.

With the spikes gone, shelf prices move when costs move.

Grocery buyers extend their contracts on packaged goods instead of rewriting them after every paper-driven oil move. Freight, packaging and processing all carry an energy cost, and each oil spike used to reopen supplier terms. With the spikes gone, the contracts run for the season, and shelf prices move when costs move.

The extra physical sitting under title absorbs the shock that used to hit the price first.

A sustained rise in productivity growth raises permanent income and can stimulate demand; without a corresponding monetary-policy response, inflationary pressure can appear, and if demand rises before supply gains are realised, inflation can increase even as capacity expands. The aggregate result therefore depends on the scale and breadth of the gains and on a policy response that keeps expectations anchored. Within those limits, fewer spikes preserve the informational content of relative prices, reduce the real cost of uncertainty, and allow efficiency gains to appear as higher real purchasing power rather than as a continuous series of nominal offsets.

Tradeoffs and Objections

The capital the rule requires is not a deadweight loss. It buys a real asset that remains in the economy as a buffer stock. What speculators give up is leverage: a position that required 2.3× to 21× less capital now requires ownership, and the holder bears storage charges and financing on the metal. The rule requires a receipt registry, single-use enforcement, attestation of off-exchange stocks, and continuous monitoring of each participant's E/I ratio. And while hedge positions are exempt, hedgers need counterparties, and speculators supply much of that liquidity. Keynes (1930) and Hicks (1939) argued that speculators earn a risk premium precisely for absorbing hedgers' risk; restricting them may raise hedging costs. This cost is procyclical, since the constraint binds most when inventory is low and hedging demand is highest.

The liquidity objection assumes that hedgers need speculators without physical exposure on the other side. In agriculture, the record suggests otherwise. In 2016, just over 2 percent of U.S. farms used futures or options; among corn and soybean farms the share was 11.8 percent, and use was higher among farms carrying debt (Prager et al. 2020). Far more farms managed price risk through marketing contracts with buyers: about 156,000 farms in all, including some 61,000 corn growers and 73,000 soybean growers. The usual counterparty to such a contract is an elevator, merchant or processor that holds grain. That counterparty already satisfies the rule, because it owns the good. The farmer's hedge does not require a counterparty betting on price; it requires one holding stock, and the rule produces more of them.

In oil and gas, hedging demand is partly created by lenders. Producers borrow against their reserves under reserve-based lending (RBL), a market with hundreds of billions of dollars in commitments (Byun and Monteith 2026). RBL agreements may require the borrower, as a condition of closing or soon after, to hedge a minimum share of projected production, with volume and tenor set by the borrowing-base credit sought (Rabalais and Einbinder 2014). Practitioners describe the requirement as unusual in conventional U.S. facilities, applied mainly to highly leveraged borrowers or where hedges are needed to reach the borrowing base (Fox, Gonsoulin and Price 2014). It is standard in the newer securitisations of producing reserves, which typically require hedges on 75 to 85 percent of production (Byun and Monteith 2026). The swaps are typically written by the lending banks themselves and secured by the same oil and gas properties as the loan (Rabalais and Einbinder 2014). The hedging demand that speculative liquidity is said to serve is thus, in part, demand the financial system writes into its own loan contracts.

The same contracts already apply the principle of this rule to producers. RBL agreements typically cap hedges at a share of anticipated production, historically 80 to 90 percent of proved developed producing reserves in the United States, expressly to avoid “speculative hedges” and “commodity hedges with notional volumes in excess of actual physical production” (Fox, Gonsoulin and Price 2014; Rabalais and Einbinder 2014; OCC 2018). Lenders will not let a producer hold price claims beyond the oil it actually has. The rule asks the same of everyone else.

“Lenders will not let a producer hold price claims beyond the oil it actually has. The rule asks the same of everyone else.”

Finally, deliverable-supply spot-month limits already address congestion and manipulation near expiry. The rule's marginal contribution lies outside the spot month and in markets where open interest is large relative to visible stocks, and its case rests on that margin.

Empirical Agenda

A welfare calculation needs four estimates.

If ∂g/∂E is indistinguishable from zero, the rule's case rests on the indirect channel alone, and the crowding-out estimate determines whether that channel exists.

Conclusion

The rule caps each participant's net speculative exposure at ρ times its own verified, unencumbered physical inventory, exempting bona fide hedges. It takes from the real-bills doctrine only the principle that claims rest on the goods they reference, measured in physical rather than nominal units to avoid the doctrine's classic failure. Aggregate speculative exposure is bounded by ρ times verified stocks, so claims cannot multiply against fixed inventory. The rule truncates the volatility generated by claim expansion if variance rises with exposure conditional on inventory. And a pure paper position requires 1 + 1/(ρm) times its present capital, between roughly 2.3× and 21× at observed margins, with the difference held as metal. The rule has never been tried. Applied to COMEX today, at ρ = 1 managed money would hold 72–96% of all deliverable gold, $45–59 billion of metal; the present silver book exceeds deliverable silver outright. At ρ = 5, the figures are $9–12 billion of gold and up to $1.5 billion of silver. In crude oil, NYMEX claims stand at about 71 times the oil at the Cushing delivery hub, and the collapse of April 2020 showed what unbacked longs do when storage runs out.

This is a transfer, not a loss. Capital now holding paper claims would hold the commodity. Speculators become storers, and the theory of storage is unambiguous that larger stocks mean fewer and smaller price spikes. The rule's benefit therefore does not depend on whether speculation itself moves prices. It depends on whether the metal speculators must hold adds to the stock of the good, through new supply and metal drawn out of consumption, rather than reshuffling existing holdings. Prices over time are governed by physical resources, not by the volume of claims written against them. The rule ties financial capital to those resources.

The result is a net benefit to the economy as a whole, and its beneficiaries are not speculators. Producers and processors face steadier input costs and plan capacity against clearer price signals. Hedgers trade against a market whose claims rest on goods. Firms downstream meet fewer of the large cost shocks that trigger general repricing (Section 6), so inflation is lower and less volatile. Households keep more of the real value of wages and savings, and plan with less need to insure against abrupt cost jumps. Even speculators hold a real asset in place of a leveraged claim. The only thing surrendered is leverage: the ability to multiply claims against goods that do not exist. What the economy gains in exchange is a larger stock of the good itself, the buffer on which every holder of a claim, and every user of the commodity, ultimately depends.

The data do not identify the right multiple. They show that ρ = 1 cannot accommodate today's silver market, and that any multiple moves capital at scale from claims into stocks.

Blake, R. (2010). In commodity swaps explosion, a bona fide controversy. Institutional Investor , 26 March 2010.

Board of Trade of the City of Chicago v. Christie Grain & Stock Co., 198 U.S. 236 (1905).

Brennan, M. J. (1958). The supply of storage. American Economic Review , 48(1), 50–72.

Büyükşahin, B., & Harris, J. H. (2011). Do speculators drive crude oil futures prices? Energy Journal , 32(2), 167–202.

Byun, S. J., & Monteith, M. T. (2026). Reserve-based lenders must adapt risk framework as energy producers package best collateral in asset-backed securities. Federal Reserve Bank of Dallas, 15 October 2026.

Cafiero, C., Bobenrieth, E. S. A., Bobenrieth, J. R. A., & Wright, B. D. (2011). The empirical relevance of the competitive storage model. Journal of Econometrics , 162(1), 44–54.

Cheng, I.-H., & Xiong, W. (2014). Financialization of commodity markets. Annual Review of Financial Economics , 6, 419–441.

CME Group (2026). COMEX metal depository statistics, 14 October 2026, as reported by The Vault Report.

Commodity Futures Modernization Act of 2000, Pub. L. 106-554 (see CRS RS20560).

Commodity Futures Trading Commission Act of 1974, Pub. L. 93-463, 88 Stat. 1389.

Commodity Futures Trading Commission (2020). Interim staff report: trading in NYMEX WTI crude oil futures contract leading up to, on, and around April 20, 2020. November 2020.

Commodity Futures Trading Commission (2021). Position limits for derivatives. Final rule, 86 Fed. Reg. 3236, 14 January 2021.

Commodity Futures Trading Commission (2026). Disaggregated Commitments of Traders, futures only, report of 13 October 2026: metals and petroleum.

Deaton, A., & Laroque, G. (1992). On the behaviour of commodity prices. Review of Economic Studies , 59(1), 1–23.

Deaton, A., & Laroque, G. (1996). Competitive storage and commodity price dynamics. Journal of Political Economy , 104(5), 896–923.

Energy Information Administration (2026). Weekly petroleum status report, week ending 9 October 2026, as reported by Oklahoma Energy Today and Rigzone.

European Union (2012). Regulation (EU) No 236/2012 on short selling and certain aspects of credit default swaps.

European Union (2014). Directive 2014/65/EU on markets in financial instruments (MiFID II), Article 57.

European Union (2021). Directive (EU) 2021/338 amending Directive 2014/65/EU (the MiFID II "quick fix").

Fama, E. F., & French, K. R. (1987). Commodity futures prices: Some evidence on forecast power, premiums, and the theory of storage. Journal of Business , 60(1), 55–73.

Federal Reserve Bank of St. Louis (2025). Long-term price trends. FRED Blog, November 2025.

Fox, J., Gonsoulin, D., & Price, K. (2014). A tale of two markets, part 3: reserve-based finance. Oil & Gas Journal , 13 March 2014.

Futures Trading Practices Act of 1992, Pub. L. 102-546.

Gagliardone, L., Gertler, M., Lenzu, S., & Tielens, J. (2026). Micro and macro cost-price dynamics in normal times and during inflation surges. Federal Reserve Bank of New York Staff Report No. 1195.

Geman, H., & Smith, W. O. (2013). Theory of storage, inventory and volatility in the LME base metals. Resources Policy , 38(1), 18–28.

Gorton, G. B., Hayashi, F., & Rouwenhorst, K. G. (2013). The fundamentals of commodity futures returns. Review of Finance , 17(1), 35–105.

Hamilton, J. D. (2009). Causes and consequences of the oil shock of 2007–08. Brookings Papers on Economic Activity , 2009(1), 215–261.

Hicks, J. R. (1939). Value and Capital . Oxford: Clarendon Press.

Humphrey, T. M. (1982). The real bills doctrine. Federal Reserve Bank of Richmond Economic Review , 68(5), September/October, 3–13.

Irwin, S. H., & Sanders, D. R. (2011). Index funds, financialization, and commodity futures markets. Applied Economic Perspectives and Policy , 33(1), 1–31.

Kaldor, N. (1939). Speculation and economic stability. Review of Economic Studies , 7(1), 1–27.

Karadi, P., Nakov, A., Nuño, G., Pasten, E., & Thaler, D. (2025). Why monetary policy should crack down harder during high inflation. VoxEU, 4 May 2025.

Keynes, J. M. (1930). A Treatise on Money , Vol. 2. London: Macmillan.

Kilian, L., & Murphy, D. P. (2014). The role of inventories and speculative trading in the global market for crude oil. Journal of Applied Econometrics , 29(3), 454–478.

Marto, R., & McCollum, M. (2026). Do productivity booms push down prices? Federal Reserve Bank of St. Louis, On the Economy, 24 September 2026.

Masters, M. W. (2008). Testimony before the Committee on Homeland Security and Governmental Affairs, U.S. Senate, 20 May 2008.

Mints, L. W. (1945). A History of Banking Theory in Great Britain and the United States . Chicago: University of Chicago Press.

Office of the Comptroller of the Currency (2018). Oil and gas exploration and production lending. Comptroller's Handbook, version 2.2.

Pindyck, R. S. (2001). The dynamics of commodity spot and futures markets: A primer. Energy Journal , 22(3), 1–29.

Prager, D., Burns, C., Tulman, S., & MacDonald, J. (2020). Farm use of futures, options, and marketing contracts. USDA Economic Research Service, Economic Information Bulletin 219.

Purcell, W. D., & Valdez, R. The Commodity Futures Trading Commission Act of 1974. National Agricultural Law Center.

Rabalais, R., & Einbinder, M. (2014). Oil and gas reserve-based lending. In ICLG: Lending & Secured Finance 2014 , ch. 11, 52–55.

Sargent, T. J., & Wallace, N. (1982). The real-bills doctrine versus the quantity theory: A reconsideration. Journal of Political Economy , 90(6), 1212–1236.

Securities and Exchange Commission (2004). Regulation SHO, 17 C.F.R. § 242.203(b).

Smith, A. (1776). An Inquiry into the Nature and Causes of the Wealth of Nations . London: W. Strahan and T. Cadell.

Symeonidis, L., Prokopczuk, M., Brooks, C., & Lazar, E. (2012). Futures basis, inventory and commodity price volatility: An empirical analysis. Economic Modelling , 29(6), 2651–2663.

Tang, K., & Xiong, W. (2012). Index investment and the financialization of commodities. Financial Analysts Journal , 68(6), 54–74.

Thornton, H. (1802). An Enquiry into the Nature and Effects of the Paper Credit of Great Britain . London: J. Hatchard and F. and C. Rivington.

Working, H. (1949). The theory of price of storage. American Economic Review , 39(6), 1254–1262.

Wright, B. D. (2011). The economics of grain price volatility. Applied Economic Perspectives and Policy , 33(1), 32–58.