Crude Oil Contango and Backwardation Explained

Contango and backwardation describe the slope of the crude oil futures curve. In contango, later-dated futures cost more than near-dated ones — the classic sign of an oversupplied market paying for storage. In backwardation, the front of the curve trades at a premium — the market is paying up for barrels now. That slope drives roll yield, storage trades, and hedge costs.

If you trade oil for any length of time, the curve will matter more to your P&L than the headline price. Plenty of traders have been dead right on direction and still lost money because they were long a contango market through a dozen rolls. This piece covers the mechanics, the math, and two case studies — the 2020 super-contango and the 2026 Hormuz whipsaw — that show both regimes at their extremes. It builds on the foundations in our complete guide to crude oil trading, which is the place to start if futures curves are new territory.

The futures curve: what you’re actually looking at

NYMEX WTI (CL) lists monthly contracts years into the future, each one for 1,000 barrels deliverable at Cushing, Oklahoma. Line up the settlement prices for each month — August, September, October, and so on — and you get the futures curve, sometimes called the strip or the term structure. ICE Brent works the same way, cash-settled against the Brent Index rather than physically delivered.

The curve is not a forecast. That’s the single most common misreading, and we’ll come back to it. What the curve actually encodes is the market-clearing price of time: what someone will pay you (or charge you) to shift a barrel from one delivery month to another. That price of time is anchored by two physical realities — the cost of storing oil, and the value of having oil on hand when you need it.

Traders rarely talk about the whole curve. They talk about timespreads: M1–M2 (front month minus second month), CL1–CL6, Dec-Red-Dec (December this year versus December next year). A quoted spread of +$0.80 means the front is 80 cents over the deferred — backwardation. A quote of −$0.50 means contango. Watch how those spreads move day to day and you’ll often learn more than the flat price tells you.

Contango: the market pays you to store oil

Contango is an upward-sloping curve: each successive month is priced higher than the one before it. It shows up when near-term supply exceeds near-term demand. Nobody urgently needs barrels today, so the spot price sags, while deferred prices hold up because storage, insurance, and financing costs get baked into later delivery.

Cost of carry sets the ceiling

Here’s the discipline contango lives under: it cannot exceed the full cost of carry for long, because arbitrage kicks in. The carry on a barrel at Cushing looks roughly like this:

• Storage: tank leases at Cushing have historically run in the range of 20–50 cents per barrel per month depending on how tight capacity is. In the March 2020 scramble, auction rates hit 50 cents, roughly double the pre-crisis norm.
• Financing: you have to fund the barrel. At a 5% interest rate on a $70 barrel, that’s about 29 cents a month.
• Insurance and losses: small but real, call it a few cents.

Add it up and full carry on land might be 55–80 cents per barrel per month. If the M1–M2 contango widens beyond that, a trader with tank space buys the front month, takes delivery, stores the oil, and sells a deferred future against it — locking in a riskless spread over their known costs. That buying at the front and selling at the back flattens the curve. This is why sustained “super-contango” only appears when storage itself is nearly full and the marginal tank (or chartered tanker) gets very expensive: the arb breaks down because carry costs explode.

The flip side: when the contango is steeper than your carry cost, storage owners mint money. In 2008–09 and again in 2020, trading houses and majors chartered supertankers purely as floating warehouses to harvest the spread. It’s the closest thing to a bond trade the oil market offers — and it’s why physical players with tankage consistently out-earn paper traders in glut years.

Backwardation: the premium for barrels today

Backwardation is the mirror image: the front trades over the deferred months, and the curve slopes down. There’s no arbitrage ceiling here, which is why backwardation can get far more extreme than contango. You can’t borrow barrels from the future. If refiners need crude this month and inventories are low, the front month can spike almost without limit while the back of the curve barely moves — the market is saying the shortage is temporary, but acute.

The textbook name for what drives backwardation is convenience yield: the implied benefit of physically holding inventory rather than a paper claim on future delivery. When stocks are scarce, having oil in your tank — feedstock security for a refiner, optionality for a trader — is worth more than the storage costs it incurs. High convenience yield pulls the spot price above deferred prices.

Backwardation is oil’s default state more often than people assume. CME Group research puts WTI in backwardation about 58% of the time since 1985, versus 42% in contango, measured front month against the sixth month out. Oil is not like gold, where contango is near-permanent because storage is trivial and the metal never gets “used up.” Crude gets burned, inventories draw, and OPEC has spent much of the past decade actively managing the market toward tightness — producers prefer a backwardated curve because it punishes anyone hoarding inventory against them.

Contango vs backwardation at a glance

One correction worth making explicit, because you still see it in print: high inventories at Cushing go with contango, not backwardation. When tanks are full, the front month is the weakest point on the curve — that’s the near-month discount that defines contango. If you internalize nothing else from this article, internalize the inventory-curve link. It’s covered from the physical side in our piece on oil supply and demand fundamentals.

Roll yield: where the curve quietly eats (or feeds) your P&L

Futures expire. Anyone holding a position for longer than a contract month has to roll: close the expiring contract, open the next one. The curve’s slope determines what that roll costs you, and the arithmetic compounds brutally over a year.

The contango drag, worked through

Say WTI trades at $70.00 front month and $70.40 second month — a modest 40-cent contango. You’re long one CL contract (1,000 barrels; each $0.01 is $10). At the roll you sell the front at $70.00 and buy the next month at $70.40. You now own the same barrels priced 40 cents higher, so your breakeven just rose $400 per contract. Do that twelve times and, if the curve doesn’t move, spot must rise about $4.80 — nearly 7% — for you simply to break even. That is negative roll yield, and it’s the main reason long-only oil ETFs like USO chronically underperform spot oil in contango regimes. USO learned this the hard way in April 2020, when it was forced to spread its holdings across multiple months mid-crisis and then executed a 1-for-8 reverse split after the front-month roll drag became unsurvivable.

Backwardation pays you to be long

Reverse the curve: front at $70.00, next month at $69.20. Rolling your long means selling at $70.00 and rebuying at $69.20 — you bank 80 cents ($800 per contract) of positive carry each roll, provided the curve holds. In the 2026 supply crunch, CME Group’s economists worked a live example: a long position in WTI futures rolled once during March 2026 returned about 53.4% between late February and early April, versus 48% for the spot price move — roughly 5.4 points of pure roll yield from the steep backwardation. Same direction, same market, materially different P&L. The mechanics of rolling — dates, liquidity, execution — are covered in our guide to how to trade crude oil futures.

The long-run numbers are stark. In the same CME study, a hypothetical strategy that was long front-month WTI only on days following a contango close lost roughly 95% since 1985, while being long only after backwardated closes gained on the order of 5,250% — before costs, and with all the usual hypothetical-backtest caveats. Nobody trades that literally. But the asymmetry tells you which regime you want at your back.

Case study: the 2020 super-contango

April 2020 is the reference event for contango, and it will be cited for decades. COVID lockdowns vaporized roughly a quarter of global oil demand inside two months, right as the March 2020 OPEC+ price war flooded the market with Saudi and Russian barrels. Oil went somewhere: into tanks. Cushing added a record 23 million barrels in April alone, reaching 65 million barrels by May 1 — about 83% of its 76-million-barrel working capacity, with most of the remainder already leased.

On April 20, 2020, the expiring May WTI contract (CLK20) settled at −$37.63. Holders who couldn’t take delivery — and with Cushing effectively full, almost nobody could — paid to escape. The June contract settled near +$20 the same day, an M1–M2 spread of almost $58. That is not a typo: the market charged fifty-eight dollars a barrel for one month of time. Traders with physical tankage or ships locked in carry returns that normally take years to earn; everyone long the front month via ETFs or brokers took catastrophic losses. The episode permanently changed how retail platforms, and USO itself, handle expiring energy contracts.

Case study: the 2026 whipsaw

The other extreme arrived six years later. WTI closed at $67.02 on February 27, 2026, with the curve in mild contango — a comfortable, well-supplied market. The strikes on Iran the next day and the subsequent disruption of Strait of Hormuz traffic took a double-digit share of seaborne supply offline within weeks. By early April, spot WTI was near $99 and Brent had spiked past $120; the curve flipped into some of the steepest backwardation since 1990, with nearby months commanding huge premiums while December 2026 barely moved — the deferred market never believed the disruption was permanent.

It was right. Ceasefire progress and recovering tanker traffic brought Brent back to the low $70s by late June, a $17 fall in four sessions at one point. Note the sequence: the curve steepened before the flat price peaked, and the front spreads collapsed before the headlines turned. Traders watching the M1–M2 spread had an earlier, cleaner read on the physical market than anyone watching the flat price — and the crisis premium lived almost entirely in the front six months of the curve, exactly where theory says a temporary shock belongs.

A worked storage trade, dollar by dollar

Since the storage arb is what anchors contango, it’s worth running the numbers once properly. Assume it’s a glut year. Spot WTI at Cushing is $62.00, the 6-month-out future trades at $66.20 (a $4.20 contango), and you can lease tank space at 35 cents per barrel per month. Financing runs 5% annually. You control 100,000 barrels of tankage.

1. Buy 100,000 barrels of physical crude at $62.00: outlay $6.2 million.
2. Sell 100 CL contracts six months out at $66.20. You are now flat price risk: whatever spot does, you deliver stored barrels against your short at a locked price.
3. Pay the carry. Storage: $0.35 × 6 months = $2.10/bbl. Financing: 5% on $62 for half a year ≈ $1.55/bbl. Insurance and sundries, call it $0.15. Total carry ≈ $3.80/bbl.
4. Collect the spread. $4.20 contango minus $3.80 carry = $0.40/bbl of nearly riskless margin — $40,000 on the position, plus optionality if the contango widens and you can re-lease or roll.

Forty cents doesn’t sound like much until you notice the trade barely has market risk — the residual risks are operational (tank availability, delivery logistics, margin calls on the short futures leg if prices rally before expiry). Scale it to a trading house with millions of barrels of tankage and you understand why Vitol, Glencore, and the majors’ trading arms report their best years in glut markets. In 2020 the spreads got so wide that six months of carry paid for itself several times over; firms that had leased Cushing space at pre-crisis rates were effectively holding lottery tickets that had already won.

The worked example also explains a rule of thumb: contango near “full carry” is a market running out of storage. The wider the spread relative to physical carry costs, the more distressed the front of the curve — and the closer you are to something breaking, as it did on April 20, 2020.

Where the words come from (and Keynes’s contribution)

“Contango” is Victorian London Stock Exchange slang — a fee a buyer paid to defer settlement to the next account day. “Backwardation” was the mirror charge paid by sellers. The terms migrated to commodities and stuck.

The theory came later. Keynes argued in the 1930s that futures should normally trade below the expected future spot price — “normal backwardation” — because hedging producers must pay speculators an insurance premium to absorb their price risk. Whether that risk premium actually exists in modern oil markets is still argued in the literature; what’s not arguable is that the physical drivers — storage cost on one side, convenience yield on the other — do most of the day-to-day work of setting the curve’s shape. Keynes gives you the vocabulary; Cushing tank levels give you the trade.

Common mistakes with the curve

• Reading the curve as a price forecast. A backwardated curve showing December $8 below spot is not the market predicting an $8 decline; it’s the market pricing today’s scarcity. Deferred prices historically say little about where spot actually lands.
• Ignoring roll costs on leveraged longs. The trader who bought a “cheap” oil dip in 2015 and rolled front-month longs for a year paid the contango twelve times. Many were right about the eventual recovery and still lost money.
• Confusing the inventory signal. Full tanks at Cushing mean contango pressure, not backwardation. Getting this backwards inverts every conclusion downstream of it.
• Assuming the regime persists. The 2026 flip from contango to violent backwardation took days. Position sizing that assumed a stable curve got carried out. Spreads are usually slower-moving than flat price — until they aren’t.
• Trading ETFs blind. If you can’t state your fund’s roll schedule and the current M1–M2 spread, you don’t know your own carrying cost.

Is contango bullish or bearish? (Mostly the wrong question)

Traders new to term structure want the curve to be a directional signal. It half is. Contango reflects a currently oversupplied market, which is bearish context; backwardation reflects tightness, which is bullish context. The CME data above shows most of oil’s bear-market damage has historically come during contango periods, while backwardated periods skew flat-to-higher. Persistent, deepening backwardation with drawing inventories is one of the more honest bullish signals this market offers.

But the curve is a description of today’s physical balance, not a prophecy. Two failure modes recur. First, the curve says nothing about shocks: the 2026 spike arrived out of a contango market, and no spread trader saw Hormuz coming. Second, prices routinely fall during backwardation — the back half of 2022 spent months backwardated while flat price ground down from the post-Ukraine highs, because tightness was easing at the margin the whole way. A downward-sloping curve is not a prediction that prices will fall, any more than an upward-sloping one predicts a rally. Treat curve shape as an inventory gauge with a delay, not a crystal ball.

There’s also a structural wrinkle: the shale era changed the curve’s behavior. From 2008 to about 2014, WTI spent far more time in contango than its long-run average, partly because surging U.S. production kept mid-continent storage brimming. Benchmarks matter here too — WTI’s curve answers to Cushing tank levels, while Brent’s answers to the seaborne market, so the two curves regularly disagree at the front. That divergence is half the reason the WTI–Brent spread exists as a trade at all.

How to actually use the curve

If you hold positions across rolls

Price your roll cost before you enter. A swing trade you expect to hold three months in a 50-cent-per-month contango starts $1.50 behind; the same trade in backwardation starts ahead. For longer horizons, consider deferred contracts outright — buying December instead of rolling the front twelve times sidesteps most of the drag, at the cost of lower liquidity and a muted response to spot moves.

If you trade ETFs

Check the fund’s roll methodology before assuming it tracks oil. A front-month-rolling fund in steep contango can lose double digits annually versus spot. In backwardation the same structure becomes a tailwind — the 2021–2023 period was unusually kind to futures-based oil funds for exactly this reason.

If you want to trade the curve itself

Calendar spreads — long one month, short another — let you trade tightness directly with less flat-price risk and far lower margin than outrights. Spread margins on WTI calendars run a fraction of outright margin because the legs hedge each other. That’s a topic that deserves its own treatment, but the core idea is simple: if you think inventories will draw faster than the market expects, buy the front spread; if you think a glut is building, sell it.

Signals worth tracking weekly

• M1–M2 and M1–M6 WTI spreads: direction and rate of change matter more than level.
• The EIA weekly report: Cushing stocks in particular — the front of the WTI curve is, functionally, a Cushing gauge. The data is free at eia.gov.
• Brent vs WTI structure: when one curve flips and the other doesn’t, something local (Cushing logistics, freight, a refinery outage) is driving it.
• Curve versus carry: contango near full carry says storage is the marginal buyer — a fragile state that ends violently when tanks fill, as 2020 proved.

CME Group publishes current settlement prices for every listed month on its WTI crude oil futures page, and its research desk’s 2026 analysis of the backwardation regime is worth reading in full for the roll-yield history.

If you hedge physical barrels

Producers and refiners live with the curve whether they like it or not. A producer hedging next year’s output in backwardation locks in prices below spot — painful to sign, but it’s the market’s honest price for deferred barrels, and it beats hedging nothing into a collapse. A refiner buying forward cover in contango pays over spot for the privilege of certainty. The curve is the hedger’s cost of insurance, and it moves: the difference between hedging a December strip in February 2026 versus April 2026 was measured in tens of dollars. Timing your hedge program against the curve regime matters nearly as much as the flat price level you lock.

The bottom line

Contango and backwardation are the oil market showing you its inventory position in price form. Contango: surplus, storage economics in charge, longs bleed on the roll. Backwardation: tightness, convenience yield in charge, longs get paid to wait. Neither is a forecast, both are context, and the transitions between them — 2020’s collapse into super-contango, 2026’s violent flip to backwardation and back — are where fortunes get made and blown up. Learn to read the spreads before you size up in flat price; the rest of our crude oil trading guide builds on exactly that foundation.