Negative WTI (April 20, 2020)

Definition

On April 20, 2020 — the day before the May WTI contract’s expiry, with COVID shutdowns collapsing oil demand and Cushing storage nearly committed — the front-month future fell from ~$18 to settle at −$37.63: the first negative settlement in the contract’s history. Holders paid to be rid of oil. It is the KB’s case study in PHYSICAL-DELIVERY REALITY (the contract prices storage at the delivery point, not “oil” in the abstract) and in retail product design colliding with expiry mechanics.

How it works / structure

• The physics: WTI delivers physically at Cushing, Oklahoma (instrument-energy-futures); with storage capacity effectively spoken for, a long position at expiry meant taking delivery with nowhere to put it — negative prices are rational when disposal costs exceed the commodity’s value; CME had enabled negative-price trading systems weeks earlier (the possibility was announced, documented in the CFTC report).
• The expiry concentration (CFTC findings): the collapse concentrated in the final trading sessions of a contract with shrinking open interest — most institutional length had rolled (ms-futures-roll); remaining longs included retail-linked products and participants unable to take delivery, selling into a bidless expiry window.
• The product casualty: the largest retail oil ETF held concentrated front-month positions by mandate; it survived April 20 (it had rolled days earlier) but restructured its roll schedule and issued/reverse-split under stress — documenting how wrapper mandates interact with curve extremes (instrument-etf + ext-commodities roll-drag at its historical maximum: the super-contango).
• Engine-relevant parameters: days-to-expiry position flags on physical contracts, storage-utilization data as a curve-thesis input, and negative-price capability in every pricing and margin model touching commodities.

When it applies

Cited for physical-expiry discipline (the platform’s standing rule that positions in physical-delivery contracts exit or roll before the delivery window without exception); for model-domain honesty (lognormal assumptions forbid negative prices — model choice is a risk decision); for wrapper-mandate risk on curve extremes.

Risk profile & failure modes

• The central lesson: a futures contract is its delivery terms; trading it without the physical logistics literacy is trading a different, imaginary instrument.
• Model-domain failure: systems assuming prices ≥ 0 — margin engines, options models (opt-pricing-models lognormal), risk reports — produced silent nonsense; CME’s switch to Bachelier (normal) options pricing that week is the documented fix.
• Expiry-window liquidity: open interest drains before delivery; the remaining book is thin and one-sided — time-to-expiry is a liquidity variable, not a calendar fact.
• Misuse: “oil went negative” as generic tail-risk color — the event was contract-specific (Brent, cash-settled seaborne, settled ~$9 positive the same day); the mechanism is the lesson.

Evidence & limits

The CFTC interim report is the primary record (price path, open-interest anatomy, storage context); it explicitly declines a single-cause attribution. CME’s negative-price enablement and pricing-model switch are documented exchange actions. Product-level details are from public filings.

Falsifiable-thesis examples

Illustrations only, not signals:

• “No physical-delivery position in this book is ever held inside its final 3 trading days (expiry-discipline audit)” — falsified by the position-calendar scan.
• “Cushing storage utilization above 85% will coincide with front-spread super-contango (storage-pricing thesis)” — falsified by the paired series.

Cross-references

• The contract physics: instrument-energy-futures, ms-futures-roll, ext-commodities (storage theory vindicated)
• The model-domain lesson: opt-pricing-models, risk-scenario-analysis
• The macro backdrop: episode-covid-2020