Negative Oil Prices, April 2020

Key Takeaways

• The May WTI contract opened the April 20 trading session at $17.73 a barrel on Sunday evening and fell $58.05 to an intraday low of -$40.32 at 2:29 p.m. ET Monday, before settling one minute later at -$37.63, per the CFTC's interim staff report on the day.
• The mechanism was physical delivery, not a collapse in the value of crude itself. Cushing, Oklahoma's 75.8 million barrels of working storage capacity was about 76 percent full by April 17, up from roughly 50 percent in mid-March, and much of what remained was already leased.
• Almost nobody who intended to hold the May contract to expiration was still in it by the afternoon of April 20. Open interest had fallen from a peak of 634,727 contracts on April 2 to 108,593 at the start of the session, and 89.6 percent of that day's total WTI trading volume was already in June or later contracts.
• The negative print did not spread. Every later WTI contract month settled positive that day, and Brent crude, a cash-settled benchmark with no delivery obligation, closed its June contract at $19.33 the next day.
• The disruption was brief in the contract itself: prices crossed back above zero by roughly 8:04 p.m. ET that same evening, dipped negative again overnight, then the May contract expired the next day at a positive $10.01.

What Happened on April 20, 2020, Hour by Hour?

Trading on what the CFTC calls "April 20" actually began the evening before. NYMEX's WTI contract trades nearly 24 hours a day, Sunday through Friday, and the session covering April 20 opened at 6:00 p.m. ET on Sunday, April 19, as Asian markets came online. The May contract opened that session at $17.73 a barrel. It was already the penultimate day of its life: under NYMEX rules the front contract stops being "active" two business days before expiration, so June had taken over as the active, most-liquid month on Friday, April 17, leaving May to trade out its final two sessions as a shrinking, thinly-held position.

Prices drifted down for most of the session, and the pace of the decline accelerated around noon Eastern Time. The real break came late. At 2:08 p.m. ET, the May contract traded below zero for the first time in the history of the WTI contract, which had been listed since 1983. Over the following twenty minutes, from 2:08 to 2:28 p.m., prices fell from $0 to -$39.55. Because May was a non-active month, its settlement price was not simply the last trade; NYMEX rules set it as the volume-weighted average price of calendar-spread transactions between the May and June contracts during a two-minute settlement window, 2:28 to 2:30 p.m. ET. That window's trading pushed the contract to an all-time intraday low of -$40.32 at 2:29 p.m., and the settlement calculation over the full window landed at -$37.63.

The move was not confined to the outright price. The May-June calendar spread, the difference between the two contract months, widened from the previous close of -$6.76 on April 17 to -$58.06 at settlement on April 20, which is a cleaner way to see how much the May contract specifically had detached from everything trading around it. The June contract itself, one month further from delivery, settled positive that day and never came close to zero.

Times and prices are from the CFTC's interim staff report, published November 23, 2020, and the U.S. Energy Information Administration's April 27, 2020 explainer. All times are Eastern Time.

By the close on April 21, the May-June spread had narrowed back to -$1.56, roughly where it had been trading in mid-March, before the storage panic took hold. The extreme pricing was concentrated almost entirely in a single twenty-two-minute window on a single day, in a single contract that was one day from ceasing to exist.

Why Does an Oil Futures Contract Require Physical Delivery?

The NYMEX WTI Light Sweet Crude Oil contract is not a side bet on a number. Each contract represents 1,000 barrels of crude oil meeting NYMEX's quality specifications, deliverable at Cushing, Oklahoma. Anyone who still holds it at expiration is contractually obligated to either take delivery, if long, or supply it, if short. Almost nobody actually does this: under normal conditions, participants who do not want the physical commodity close out before expiration, by trading out or rolling forward. The EIA puts a number on how rare delivery actually is: only about 1 percent of futures contracts that trade ever go to delivery.

What made April 20 different is that the exit door narrows as expiration approaches, and this time it narrowed against a wall. NYMEX designates the nearest listed month as "active," but that status shifts to the next month two business days before expiration; May lost its active-month status at 6:00 p.m. ET on April 16, handing off to June on April 17. From that point, anyone still holding May who did not want delivery in Cushing had one trading day, April 20, then expiration on April 21, to get out. Most of the market had already left, which is precisely what makes the remaining trade informative: the traders still in May on the afternoon of April 20 were disproportionately the ones who had struggled to exit, and their forced selling set the price.

Cushing exists as the delivery point because of geography, not convenience: it sits at the intersection of dozens of pipelines connecting Gulf Coast, Midcontinent and Midwest production and refining regions, which is why NYMEX chose it in 1983 as the WTI reference point and why it remains the U.S. domestic crude benchmark. That same centrality is what turned a regional storage shortage into a globally visible price event: because WTI is priced off Cushing specifically, a full tank farm in one corner of Oklahoma was enough to move the number on every financial news ticker in the world.

How Did Global Oil Demand Collapse So Fast?

The starting point was already an oversupplied market. WTI had opened 2020 at $61.18 a barrel on January 2 and had already slid to $50.95 by February 6, a decline the CFTC's report attributes in part to falling demand in China and other Asian economies as the earliest phase of the coronavirus outbreak spread. The World Health Organization did not declare COVID-19 a global pandemic until March 11, 2020, but oil markets had been pricing in a demand shock for more than a month by then.

What followed in March and April was a step change, not a gradual slowdown. By March 2, WTI had fallen to $41.28; two weeks later, as lockdowns and travel restrictions spread across major economies, it hit $28.70 on March 16. The reason showed up directly in refinery data: for the week ending April 17, 2020, U.S. refinery runs, the rate at which refiners actually process crude into gasoline, diesel and jet fuel, fell to 12.8 million barrels a day, 4.1 million barrels a day (24 percent) below the same week a year earlier. When a quarter of that downstream demand disappears in a year, the crude refiners would have bought has nowhere to go but into a tank.

The CFTC's framing is useful: the industry's demand response to COVID-19 was faster than its supply response. Drivers stopped driving and planes stopped flying within days of lockdown orders; wells do not stop producing on the same timescale, both because shutting one in can permanently damage the reservoir and because producers with existing debt service often keep pumping even at a loss. That mismatch, demand falling in days while supply adjusts over months, is the same mechanism behind almost every commodity glut, in different clothing.

Why Did OPEC Plus Make the Glut Worse Before Making It Better?

The producers who might have offset the demand collapse instead added to it, briefly. In early March 2020, OPEC and its non-OPEC partners, together OPEC Plus, tried to agree a coordinated production cut in response to the emerging demand shock. The talks failed, the existing agreement dissolved, and the group's largest members responded not by cutting output but by defending market share, pushing WTI down further as unconstrained supply met collapsing demand, the opposite of the textbook response to a demand shock, and the reason March's collapse was sharper than the pandemic's fundamentals alone would have produced.

The reversal came more than a month later and arrived too late to prevent the storage crunch. OPEC Plus members reached a deal on Sunday, April 12, 2020, and, per the CFTC's own report, formally announced it before crude oil markets opened the next morning, April 13: what the report calls the largest single production cut in history, 9.7 million barrels a day, roughly 10 percent of pre-pandemic global supply. But the agreement's own terms meant it could not help the May contract: the cuts were not scheduled to begin until May 1, tapering to 7.7 million barrels a day from July through the end of 2020 and to 5.8 million from January 2021 through April 2022. Every barrel produced before May 1 still needed somewhere to go, which is why the market's initial reaction was muted: WTI kept declining through the following week, and by the close on April 17, three trading days before the negative settlement, the May contract had fallen below $20 to $18.27, while Brent closed at $28.08. The cut was real and eventually mattered, but it was a May Day fix applied to an April 20 problem.

Why Was Storage at Cushing the Binding Constraint?

Storage, not price, was the scarce resource by mid-April 2020, and the numbers on both a global and a local scale show why. Globally, floating storage, oil held in tankers at sea rather than on land because no onshore tank has room for it, stood at roughly 49.4 million barrels worldwide at the start of 2020. Weekly inflows into floating storage exceeded 10 million barrels during the week ending March 8, immediately after the OPEC Plus talks collapsed. During the first three weeks of April, with OPEC Plus producers still not subject to any quota, floating storage rose by almost 69 million barrels to about 127 million barrels, a 117 percent increase from the March 29 level. About 40 million of those barrels, the CFTC's report notes, were added in the single week ending April 19, the week the May contract went negative.

Onshore, U.S. commercial crude stockpiles told the same story with a sharper acceleration. During the week of March 27, stocks swelled by 13.8 million barrels, compared with an average build of about 3 million barrels a week in February and early March. Over the following four weeks, stocks rose by an average of 15.8 million barrels a week, and by April 17, U.S. crude inventories sat just 17 million barrels short of their all-time recorded peak.

Cushing specifically, as the WTI delivery point, is where that pressure became a price. The EIA puts its working storage capacity at about 76 million barrels, a figure the CFTC's report gives more precisely as 75.8 million, roughly 44 percent of all Midwest crude storage and about 11 percent of total U.S. commercial capacity. As of April 17, EIA data put Cushing's actual inventory at 60 million barrels, about 58 million in tank farms and roughly 2 million more in transit, working out to the tanks themselves being 76 percent full, up from roughly 50 percent in mid-March, a doubling of utilization in five weeks. Even the unfilled portion was not simply available: both EIA and CFTC note that a meaningful share of the remaining space had already been leased or committed to other parties, so a trader without a pre-existing arrangement faced less room, at a higher lease rate, than the raw vacancy figure suggested.

The May-June calendar spread is the clearest single number capturing this dynamic building over time rather than arriving all at once. It closed at a small premium of $0.55 on January 6, 2020, meaning the May contract traded slightly above June, a mildly backwardated market consistent with tight-but-manageable supply. It crossed below zero on January 31 and hovered between -$0.04 and -$0.24 through early March. From early March onward it widened steadily as storage concerns built, closing at -$6.76 on April 17, the last trading day before the negative print, before blowing out to -$58.06 on April 20 itself. A market moving from a $0.55 premium to a $58 discount over three and a half months is not a single-day event; April 20 was the day the accumulated pressure found its release, not the day the pressure began.

How Did Open Interest in the May Contract Set Up the Squeeze?

Open interest, the number of contracts still outstanding, is the CFTC's clearest lens on who was actually exposed when the price broke. Interest in the May contract peaked at 634,727 contracts on April 2, 2020, well above the trailing twelve-month average peak for an expiring WTI contract of roughly 430,000. That elevated starting point matters, because the ordinary pattern in any expiring futures contract is a steady decline in open interest as the expiration approaches and traders roll into the next month. On the May contract, that decline was historically large in absolute terms even though it followed the usual shape: interest fell from that April 2 peak to 108,593 contracts at the start of the April 20 session, which was still 69.4 percent higher than the trailing twelve-month average penultimate-day figure of 64,101 contracts.

By the numbers, most of the market had already left. Of the 2,390,935 contract-sides of total WTI open interest across all expiration months at the start of the April 20 session, the May contract accounted for only 4.5 percent, and 89.6 percent of that day's total trading volume across all WTI contract months was already occurring in June or later contracts. Trading in the May contract itself was thin by comparison: the outright May contract traded 58,693 contracts across the full day, against 914,148 in the June contract, more than fifteen times as much. A separate May outright TAS product, an order type that lets traders transact at a fixed differential to the day's settlement price rather than at a live market price, added another 51,867 contracts. Roughly 1,900 distinct accounts were active in the May contract on April 20, against more than 5,000 in June.

The CFTC's trader-classification data shows the exodus by category. At the start of April, 122 commercial traders and 217 non-commercial traders held reportable positions in the May contract, with non-commercial traders long 356,553 contracts and commercial traders short 298,944. By the opening of the April 20 session, those numbers had shrunk to 95 commercial and 78 non-commercial traders, with commercial short positions down 77.1 percent and non-commercial long positions down 88.1 percent from the start of the month; 64.1 percent of the non-commercial traders who had been in the contract on April 1 had exited by the morning of April 20. By the close of trading on April 20 itself, only 59 commercial and 16 non-commercial traders remained, and at the contract's expiration the following day only 2,427 contracts were left open across the entire market, below the trailing twelve-month average of 2,815 and less than 1 percent of the contract's own peak open interest three weeks earlier.

None of this compression is unusual in shape; every expiring futures contract sheds open interest as traders roll forward, a routine process this site covers under futures. What made April 2020 unusual was the combination: an elevated starting point, a delivery location running out of room, and a roll period coinciding with the sharpest demand shock crude oil markets had faced in decades. The traders still holding May contracts in the final hours were not a representative slice of the market; they were disproportionately the ones for whom exiting had proven hardest, exactly the population you would expect forced sellers among.

What Role Did the United States Oil Fund Play?

The United States Oil Fund, traded under the ticker USO, is a large retail-facing exchange-traded fund that normally tracks WTI crude by holding the front-month futures contract, and its own regulatory filings show it reacting to the same storage pressure well before April 20. On April 16, 2020, USO disclosed in a filing with the U.S. Securities and Exchange Commission that, because of market conditions and regulatory requirements, it would begin diversifying away from its usual practice of holding almost entirely front-month contracts. Effective April 17, three trading days before the negative settlement, USO said it would hold approximately 80 percent of its portfolio in the front-month contract and approximately 20 percent in the second-month contract, rather than concentrating in the front month alone.

That adjustment turned out to be far too small for what followed. In a further filing dated April 21, 2020, the day after the negative settlement, USO disclosed that "because of extraordinary market conditions in the crude oil markets, including super contango," it had, effective April 21, shifted to roughly 40 percent of its portfolio in the June contract, 55 percent in July and 5 percent in August, abandoning front-month concentration almost entirely. A subsequent filing dated April 24 disclosed two more moves in quick succession: as of April 22, USO had shifted to roughly 20 percent June, 50 percent July, 20 percent August and 10 percent September, and by April 24 itself that had changed again to roughly 20 percent June, 40 percent July, 20 percent August and 20 percent September. A filing on April 27 then disclosed a more permanent structural change: starting with the May 2020 monthly roll, USO's positions would roll over a ten-day period rather than concentrating the roll into one or two days, a change the fund attributed partly to risk-mitigation requirements imposed by its own futures commission merchant.

The CFTC's interim report is deliberately cautious on causation. It does not name USO or any other individual fund, states explicitly that it does not analyze the propriety of trading by any particular trader or group, and does not identify a root cause. It does note that "market participants, including those trading for commodity index funds or ETFs, rolled their positions from the May to June Contract during the month of April," the same category of participant USO's filings put a name to. The honest reading is that USO's disclosures document one large fund adjusting its own exposure in real time, not that any single fund caused the negative print; the CFTC's open-interest data shows the roll was already 95-plus percent complete across the whole market before the session began. For a reader who owned USO or a similar fund, the practical lesson sits closer to the mechanics covered under ETF investing than to any story about manipulation: a fund that tracks "the price of oil" holds a rolling futures position, not barrels of oil, and its own prospectus, not the spot price, determines how faithfully it tracks that price when the futures curve stops behaving normally.

Why Didn't Brent Crude Also Go Negative?

Brent crude, traded on ICE Futures Europe and priced off North Sea production, is the other benchmark most financial coverage quotes alongside WTI, and it never went negative, not on April 20 and not at any point during the episode. The EIA's own account of the day is explicit about why: Brent's June 2020 contract closed at $19.33 a barrel on April 21, the day after WTI's negative settlement, while WTI's own June, July, August and later contracts also stayed positive throughout. The negative print was confined to one contract, in its final hours, not a statement that crude oil in general had gone negative.

The structural difference is settlement, not geography. The Brent contract is cash-settled: at expiration no physical barrel changes hands, and the contract settles against a reference price rather than requiring anyone to arrange delivery. The CFTC's report draws the same distinction for a WTI look-alike product on ICE, noting its analysis excludes that contract because it is cash-settled and so was never exposed to the physical-delivery mechanics of the CME's NYMEX contract. A cash-settled contract can still reflect an oversupplied market through a lower price and a deeply backwardated curve, both of which Brent showed that spring, but it cannot be pushed below zero purely because a delivery point ran out of tank space, because there is no delivery point for it to run out at.

This is the single most transferable fact here for anyone who trades or invests in commodities: two benchmarks tracking the same underlying good can behave completely differently at the extremes because of how their contracts settle, not because of any difference in the physical market they describe. The price you see quoted describes a specific legal instrument, with its own delivery terms and settlement rules layered on top of the physical commodity, and those mechanics can dominate the headline number precisely when conditions turn extreme.

What Did CME Group Do to Prepare the Market for Negative Prices?

CME Group, which owns NYMEX, did not discover on April 20 that its own contract could theoretically go negative; it had spent the preceding three weeks quietly rebuilding its systems for exactly that possibility. On April 3, CME issued a notice on changes to price and strike eligibility flags for certain energy products, flagging WTI among contracts that might need to trade at negative prices. On April 8, it addressed a subtler technical problem: standard options-pricing models such as Black-Scholes rely on logarithms and mathematically cannot price an option on an underlying asset with a negative price. CME's notice that day announced its clearinghouse would give members one day's notice before allowing negative options pricing and strikes, so participants could switch to the Bachelier model, an older pricing framework that does not depend on the underlying staying positive.

Testing followed the policy notices. On April 13, CME told members that firms wishing to test negative and zero trade, settlement and strike prices could use its "New Release" testing environment, flagging the relevant contracts as eligible to trade negative on its market-data platform. A further notice on April 15 opened additional testing specifically for negative prices and strikes on NYMEX energy contracts. Then, on the morning of April 20 itself, before the May contract crossed zero, CME's Global Command Center notified NYMEX members directly that certain energy futures, including WTI, would have no low limits that day and could trade negative.

None of this preparation prevented the price move, and it was not meant to; a limit order book cannot manufacture storage space that does not exist. What it did do was keep the exchange's own systems, and clearing members' pricing models, from breaking when the number CME had been quietly preparing for actually printed. The alternative, an exchange whose software could not display or clear a negative trade, would have meant an outright market failure layered on top of an already disorderly one.

What Happened in the Final Twenty Minutes of Trading?

The CFTC's transaction-level data on the settlement window is unusually granular, since the Commission collects every trade and limit-order-book update from the exchange. It shows a market that did not merely decline but progressively lost the ability to absorb orders without moving. Between April 13 and April 20, resting depth at the best bid and ask in the May contract fell by about 50 percent, and by April 20 the order book had become visibly imbalanced, with far more size resting on the ask (sell) side than the bid. The bid-ask spread, already elevated on April 17, widened further starting around noon on April 20 and stayed elevated the rest of the session.

Trade-at-settlement, or TAS, contracts, which let a trader lock in a price at a fixed differential from that day's eventual settlement rather than trading at a live price, showed the clearest sign of stress. TAS has hard limits: plus or minus 10 ticks (10 cents) from settlement for outright TAS, plus or minus 20 ticks for spread TAS, and these limits are ordinarily hit rarely. On April 20, outright TAS traded at its maximum differential 1,101 times, for 11,568 contracts, more than seventy times the total across all of 2019. Spread TAS hit its limit 41 times, for 241 contracts; it had not hit that limit once in all of 2019 or January through mid-April 2020. A trade type built for transacting quietly near settlement was instead pinned at its outer boundary, one more way of saying everyone on one side of the market was trying to get out at once.

Underneath the price collapse itself, the exchange's automated safeguards, dynamic circuit breakers that halt trading for two minutes if a price moves more than 15 percent within a rolling 60-minute window, triggered repeatedly. Over 30 such halts hit the May contract on April 20, more than 10 hit the May-June spread contract, and over 15 more hit various related spread products. Because the May contract was not the exchange's active month that day, these halts paused trading in May-specific products without stopping trading in the active June contract or the broader WTI market, which is part of why the disorder stayed contained to the expiring contract rather than spreading across the whole complex.

Who Actually Lost Money on April 20?

The honest answer is narrower than the headlines suggested at the time. The CFTC's own open-interest data shows the population exposed to the negative settlement was small relative to the overall WTI market: 108,593 contracts of May open interest at the start of the session, against 2,390,935 contracts outstanding across all WTI expirations, a little under 5 percent. Within that group, the loss fell hardest on whoever was still net long the May contract in the final settlement window, unable or unwilling to accept physical delivery in Cushing. A long position that had cost, say, $20 a barrel to establish, and was still open at 2:30 p.m., did not merely lose its value; the holder owed roughly $37.63 a barrel more on top of that.

Retail investors holding oil-tracking exchange-traded products were disproportionately exposed to a version of this loss even without intending to hold a position through expiration, because those products roll their futures on a schedule, and a fund that has not fully exited the front month by the time it goes disorderly absorbs whatever price that month settles at. A saver who bought an oil ETF expecting exposure to "the price of oil" was, more precisely, exposed to a specific futures contract on a specific settlement date, with all the delivery mechanics that implies, a distinction worth remembering any time a fund's name promises simpler exposure than its structure delivers.

Producers and physical market participants were affected differently and, in most cases, less directly. The EIA's analysis at the time noted that positive spot prices persisted for most other U.S. crude oils even as the May WTI futures contract went negative, and that few physical sellers were actually paying counterparties to take barrels; the phenomenon was, in the EIA's phrase, "predominantly driven by the timing of the May 2020 contract expiration" rather than a collapse in the value of oil in the ground. A producer selling into the physical market that week faced a genuinely oversupplied market and falling prices, real economic damage, but not the specific loss suffered by a trader forced to settle a financial contract at -$37.63.

Did Anyone Profit From Negative Oil Prices?

A small number of participants profited directly, and the mechanism is straightforward once the settlement mechanics are clear: anyone who was short the May contract into the settlement window, or who bought it at a negative price with a genuine plan and the logistics to take delivery, was paid to do so. Being paid to receive a commodity sounds implausible until you separate the futures price from the cost of storing and transporting the physical barrel. If storage at Cushing effectively cost more, once leasing premiums and logistics were accounted for, than the negative price a seller was willing to accept, then buying oil at -$37.63 a barrel and having somewhere ready to put it could still be a rational, profitable trade. The population able to do this was small and specialized: firms with existing storage arrangements, pipeline access and the operational capacity to actually move physical crude, not the retail traders who made up a much larger share of the accounts active in the contract that day.

A second group benefited without touching a barrel: traders who had gone short the May contract earlier, whether as a directional bet on the storage crunch or as a hedge against other positions, and closed out during the collapse. The CFTC's data shows commercial traders' short positions in the contract fell 77.1 percent between April 1 and the opening of the April 20 session, which means most commercial short exposure had already been closed before the extreme move; the traders still short into the final settlement window, a much smaller group, captured the bulk of the price collapse as it happened.

What did not happen, despite some commentary at the time, was a broad windfall for "oil bears" as a class. Shorting an asset that falls from $18 to -$37 is profitable in principle, but realizing that profit required being positioned correctly, surviving the margin calls of the preceding weeks, and having a clearing arrangement able to settle a trade at a negative price at all, precisely the operational readiness CME Group spent the preceding three weeks building. The event rewarded specific operational capability far more than a generic bearish view on oil.

Did the Exchange's Circuit Breakers Do Anything?

They triggered, but not in the way "circuit breaker" suggests from equity markets. CME's dynamic circuit breakers halt a specific instrument for two minutes if its price moves more than 15 percent within a rolling 60-minute window; a related control, velocity logic, pauses trading on moves that are too fast even within a shorter window. Both fired repeatedly as May-contract prices approached and crossed zero: over 30 dynamic-circuit-breaker events in the May contract alone, more than 10 in the May-June spread.

What they did not do was halt the broader WTI market. A circuit-breaker event in the exchange's active contract, which that day was June, would pause the entire complex; May was no longer active, so its triggers paused only May-specific products while June and the rest of the market traded on normally. The result was a series of two-minute pauses that slowed, but did not stop, the slide toward and through zero, while participants who had already rolled out of May were entirely unaffected. That is the honest job of this kind of safeguard: it slows disorderly trading, it does not prevent a price from reaching a level the underlying constraint, a full storage tank at a specific delivery point, genuinely supported.

How Did Oil Prices Recover, and Over What Timeline?

The recovery in the contract itself was almost immediate, because the negative pricing had always been specific to the expiring May contract rather than to crude oil broadly. By roughly 8:04 p.m. ET that evening, about six hours after settlement, the May contract moved back above $0. It dipped negative again briefly overnight as European markets opened, then expired the next day, April 21, at a positive $10.01, with the May-June spread narrowing back to -$1.56, roughly its mid-March level.

The June contract, which became front-month once May expired, faced a version of the same storage arithmetic weeks later without repeating a negative settlement; EIA commentary at the time flagged continued Cushing constraints as a live risk into its own expiration. The broader price recovery for crude oil tracked the slow return of demand and the phase-in of the OPEC Plus cuts agreed April 12 but not effective until May 1, working back over the following months as economies reopened.

It is worth being precise about which recovery this section describes, because conflating them is the most common way this episode gets misremembered. The May 2020 contract's price "recovering" happened within hours, once its specific delivery deadline passed. The recovery of oil demand, refinery utilization and the broader price level took considerably longer, tracking the pandemic's own trajectory rather than one contract's expiration. Treating the fast mechanical bounce as evidence that "oil recovered quickly" misreads what actually snapped back.

Common Myths About Negative Oil Prices

"Oil was worthless on April 20, 2020." Only one contract, in its final hours, ever traded or settled negative. Every other WTI contract month settled positive that same day, Brent never went negative at all, and the EIA found positive spot prices persisting for most other U.S. crude oil grades throughout. The negative print described a specific, physically-delivered instrument at a specific moment, not the value of crude oil as a commodity.

"Someone caused this deliberately, or a single fund broke the market." The CFTC's interim report explicitly declines to analyze the conduct of any particular trader or identify a root cause. Its own data shows the roll out of the May contract was more than 95 percent complete before the session even began, and it describes the remaining pressure as coming from a broad category, "commodity index funds or ETFs" among others, not a single actor. USO's filings document a fund managing its own risk in real time, not evidence that any one fund caused the move.

"Negative prices meant producers were paying people to take their oil." Mostly not. The EIA found the phenomenon largely confined to the financial futures market; most crude oil in physical transactions that week still changed hands at positive prices. What went negative was one futures contract nearing its delivery deadline with almost no storage left at its delivery point, not the wellhead economics of producing crude.

"This could never happen again because the market has learned its lesson." The contract mechanics that allowed it, physical delivery at a single hub, and exchange systems able to clear a trade below zero, are still in place. CME Group's preparation was about handling a negative price, not redesigning the contract to prevent one. A similar combination of near-full storage and a sharp, fast shock could, in principle, produce something similar again.

"Brent crude and WTI are basically the same benchmark, so this should have happened to both." They track the same global commodity but are legally different instruments. WTI requires physical delivery at Cushing; Brent settles in cash with no delivery obligation. That structural difference, not anything about relative supply and demand, is why one could go negative and the other could not.

What a Reader Can Actually Carry Forward

It is tempting to file April 20, 2020 away as a piece of trivia, a single, freakish afternoon that will never recur in exactly that shape. That framing throws away the part most likely to be useful the next time a market moves in a way that looks, on the surface, impossible.

What generalizes

• A futures price describes a specific contract, not a direct quote on the underlying good. Delivery terms, settlement mechanics and expiration timing can dominate the number you see, especially near expiration in a physically-delivered contract. Before trusting a headline price, it is worth knowing whether the instrument behind it settles in cash or requires delivery, a distinction covered under futures.
• A fund's name is not its structure. An ETF promising exposure to "the price of oil" typically delivers that through a rolling futures position, not ownership of the physical commodity. What that rolling position does when the futures curve breaks down can diverge sharply from a simple spot-price chart, worth understanding under ETF investing before buying any commodity-tracking fund.
• Physical constraints can dominate financial pricing faster than models expect. Storage, transport and delivery logistics are binding constraints in a physically-settled market, not abstractions, and they can move a price further and faster than a purely financial supply-and-demand analysis would predict.
• Being early to identify a real risk is not the same as being able to trade it. Cushing's storage concerns were public for weeks before April 20. Identifying the risk did not, by itself, reveal the exact date or magnitude of the eventual break; most of the market had already rolled out well before the extreme move, which is itself a form of risk management.

What does not generalize

• The specific magnitude. A -$37.63 settlement required an unusually severe combination of a pandemic demand shock, a producer price war, and a storage hub genuinely running out of room in the same narrow window. Most storage-driven contango episodes are far less extreme.
• The speed of the bounce-back. The May contract's return to positive pricing within hours reflects that the negative print was a delivery-deadline artifact, not a reassessment of crude's underlying value. A move driven by a genuine change in fundamentals would not be expected to reverse nearly as fast.

The question worth asking now

Not whether oil prices could theoretically go negative again, which the mechanics above already answer, but a narrower one for anyone holding a commodity-tracking investment: does this product hold the physical asset, or does it hold a rolling futures position, and if the latter, what happens to it specifically when the front-month contract nears its own expiration under stress? For most long-term holders of diversified funds, the answer barely matters day to day. For anyone holding a single-commodity futures-based ETF through a period of genuine physical-market disruption, it is the single most important thing to know about what they actually own.

Could Negative Oil Prices Happen Again?

Mechanically, yes, though the exact conditions of April 2020 were unusual enough that an identical repeat is unlikely. CME Group's operational changes were about enabling its own systems to price and clear a negative trade, not about redesigning the WTI contract to prevent one; the physical-delivery requirement at Cushing remains the contract's defining feature, and it is precisely that requirement that made April 20 possible in the first place. Nothing about the contract's structure has changed to make a repeat structurally impossible.

What made 2020 unusual was several separately unlikely conditions arriving at once: a near-simultaneous global demand shock; a producer price war that briefly added supply when the market needed less; and a delivery hub already close to capacity before either happened. Reproducing a negative settlement would not require an identical pandemic, but it would require a similarly tight combination: a nearly full delivery point, a physically-settled contract approaching expiration, and a demand or supply shock moving faster than the market's ordinary capacity to roll or deliver could absorb.

The better question for an investor is not whether this exact scenario recurs in WTI specifically, but where else the same structural ingredients exist today: a physically-delivered or physically-constrained contract, a single dominant delivery or settlement point, and a market structure where forced exits can concentrate in a narrow window. Those ingredients are not unique to crude oil, and identifying them in advance in any market is a more useful exercise than waiting for a headline that says a price went negative.

Related Reading

• The 2020 COVID-19 market crash, the broader equity-market collapse running alongside the oil story, driven by the same pandemic demand shock.
• The 1973 oil shock, the mirror-image case: a supply shortage rather than a storage glut, and a reminder that "oil shock" can mean opposite mechanisms in different decades.
• Futures, for how contract specifications, delivery terms and settlement mechanics generally shape what a futures price actually represents.
• ETF investing, for how a fund's structure, not just its name, determines what exposure it actually delivers, including for commodity-tracking funds.
• All Swoopr market history case studies.

References

Every figure on this page was verified against the following sources, each retrieved on 26 August 2026:

• U.S. Commodity Futures Trading Commission: Interim Staff Report, Trading in NYMEX WTI Crude Oil Futures Contract Leading up to, on, and around April 20, 2020: the settlement price, intraday low, opening price, timeline of the settlement window, active-month transition date, storage percentages, floating-storage figures, U.S. stockpile figures, OPEC Plus cut terms, open-interest figures, trader-count and position tables, TAS trade counts, order-book depth and bid-ask spread findings, and dynamic-circuit-breaker counts.
• U.S. Energy Information Administration: Low Liquidity and Limited Available Storage Pushed WTI Crude Oil Futures Prices Below Zero: the Cushing capacity and inventory figures for April 17, 2020, the finding that negative pricing was largely confined to the financial market, the Brent June 2020 closing price on April 21, and the U.S. refinery-run figures for the week ending April 17.
• United States Oil Fund, LP: Form 8-K, Event Date April 16, 2020: USO's disclosed shift to an 80 percent front-month, 20 percent second-month allocation effective April 17, 2020.
• United States Oil Fund, LP: Form 8-K, Event Date April 21, 2020: USO's shift to roughly 40 percent June, 55 percent July and 5 percent August contracts, and its reference to "extraordinary market conditions... including super contango."
• United States Oil Fund, LP: Form 8-K, Event Date April 24, 2020: the interim April 22 reallocation to roughly 20 percent June, 50 percent July, 20 percent August and 10 percent September, and the further shift as of April 24 itself to roughly 20 percent June, 40 percent July, 20 percent August and 20 percent September.
• United States Oil Fund, LP: Form 8-K, Event Date April 27, 2020: USO's disclosed move to a ten-day rolling schedule beginning with the May 2020 roll.
• Federal Reserve Bank of St. Louis: Crude Oil Prices, West Texas Intermediate, Cushing Oklahoma, Series DCOILWTICO: an independent daily WTI spot-price series used to cross-check the direction and approximate scale of the April 20 to 21, 2020 price move.

Note on reconciling two price series. FRED republishes a separate EIA-sourced WTI spot-price series, distinct from the CME futures settlement price used throughout this page. It shows -$36.98 for April 20, 2020 and $8.91 for April 21, close to but not identical with the CFTC's reported CME settlement figures of -$37.63 and $10.01. The gap reflects different methodologies, an EIA spot-price calculation versus a CME futures settlement, not any disagreement about what happened. This page uses the CFTC's figures throughout, since they describe the specific NYMEX contract this article covers.

Figures deliberately not stated. No dollar total for aggregate trader losses or gains on April 20, since no verified source publishes one; no claim about USO's own realized profit or loss that week, since its filings disclose allocation rather than performance; and no single entity is named as the cause of the negative settlement, since the CFTC's own report declines to draw that conclusion.

Method note: all times are Eastern Time, matching the CFTC's report. May-contract trading-volume figures are drawn from the CFTC's per-product tables (outright, spread and TAS reported separately), since the report breaks the day down by product rather than publishing one combined May-contract total.

This is educational content about a historical trading event. It is not investment advice, it is not a forecast, and nothing here should be read as a claim about the current condition of any futures contract, storage facility or fund.

Frequently Asked Questions