Contango vs Backwardation, Explained

Crude oil does not have one price. It has a calendar of prices: one for delivery next month, another for the month after, and so on for years into the future. Plot those prices in delivery order and you get the futures curve. When the curve slopes upward - later deliveries priced above sooner ones - the market is in contango. When it slopes downward, with prompt barrels priced above later ones, the market is in backwardation. Neither shape is a prediction. Each is a price signal about the present: contango is the market paying for storage and patience, backwardation is the market paying for barrels right now. This guide explains the machinery underneath both shapes, the arithmetic that keeps them honest, the week in April 2020 when the front of the curve broke, and how to read the curve with the data pages on this site.

1. One commodity, many prices

A futures contract is an agreement to deliver (or settle against) a set quantity of oil in a specific month. Because oil for delivery in December is a different economic good from oil for delivery in January - it arrives at a different time, into different storage and demand conditions - each month carries its own price. The difference between two months is called a calendar spread, and the pattern of spreads across the year is the curve.

Two vocabulary points before anything else. The front month (also called the prompt month) is the nearest contract still trading; it is the number quoted in headlines and the one PetroEyes tracks for WTI and Brent on the markets page. Contracts further out are deferred months. When you read that "oil fell," it almost always means the front month fell; the rest of the curve may have moved less, more, or the other way. Reading the curve means resisting the front month's monopoly on attention and asking what the whole calendar is saying.

The curve matters because it prices two things that flat prices hide: the cost of holding oil through time, and the urgency of having it immediately. A refiner deciding whether to buy cargoes now or later, a producer deciding whether to sell this month's output or store it, and a trading house deciding whether a tank is worth filling all read the same curve. Sections 4 and 5 build the two shapes from those incentives.

2. The contract underneath the curve

Curve mechanics only make sense once the contract plumbing is clear. The WTI futures contract on NYMEX (part of CME Group) covers 1,000 barrels per contract and is physically delivered at Cushing, Oklahoma, the pipeline and storage hub where the US inland market clears. A trader still holding a long position when trading in that contract ends is obligated to take delivery of physical barrels at Cushing. The Brent futures contract on ICE also covers 1,000 barrels per contract, but it is financially settled against the ICE Brent Index, an index derived from physical cargo assessments in the North Sea BFOE complex (Brent, Forties, Oseberg, Ekofisk, and Troll grades), with an exchange-for-physical route for parties who want actual cargoes.

That plumbing difference shapes how each curve behaves at its front end. A physically delivered contract converges, at expiry, to the value of barrels at one specific place with finite tank space: if those tanks are effectively full, the right to deliver into them is worth little and the obligation to receive barrels is worth less than nothing. A cash-settled contract converges to an index of waterborne cargoes, and waterborne barrels have an escape valve Cushing barrels do not - they can wait on a ship. Keep this in mind for section 7, where the two contracts faced the same demand collapse and only one of them printed a negative price.

One more mechanic: contracts expire in sequence. As the front month approaches its last trading day, open positions either close out or roll into the next month, and the second month becomes the new front. Around expiry, liquidity concentrates and the front spread can move for plumbing reasons - a fund rolling a large position, a physical player taking delivery - that say little about underlying supply and demand. Section 9 returns to this; for now, note that the curve is a market, not a thermometer, and markets have plumbing.

3. Two numbers that summarize the curve

You do not need to memorize twelve monthly prices to read the curve. Two spreads, both quoted in dollars per barrel, carry most of the information:

• The prompt spread - the second month minus the first month. Positive means the front of the curve is in contango; negative means backwardation. This is the tightest, most physical spread in the market: it compares barrels available almost immediately with barrels one month later, and it reacts first to storage, outages, and inventory surprises.
• The back slope - a deferred spread such as the twelfth month minus the first, or the average per-month change across the year. This describes the curve's overall tilt and moves more slowly; it reflects financing costs, expected balances, and the price producers accept to lock in future output.

Sign conventions cause endless confusion, so fix them now: throughout this guide, spreads are quoted as the later month minus the earlier month. Contango prints positive calendar spreads (you are paid, in price terms, to wait). Backwardation prints negative ones (you pay for immediacy). The Brent minus WTI spread covered in our WTI vs Brent spread explainer is a different animal - it compares two benchmarks in the same month, not one benchmark across months - but the two readings interact, and section 10 shows how to use them together.

4. Contango: the market charging rent

Contango exists because holding oil costs money. A barrel bought today and sold for delivery in six months has to sit somewhere for six months, and the holder pays tank rent or storage fees, interest on the money tied up in the barrel, and insurance against loss. Economists bundle these into the cost of carry. If storage is available and the market is orderly, the futures price for a later month settles near the prompt price plus the cost of carrying a barrel between the two delivery dates. The curve's upward slope is, in large part, a rent receipt.

This gives contango a natural ceiling, called full carry. Suppose the all-in cost of storing a barrel for one month - rent, financing, insurance, fees - comes to $0.90. If the prompt spread traded at $1.40, anyone with access to a tank could buy prompt barrels, store them, sell the next month forward, and pocket the $0.50 difference with the price risk hedged away. That trade, repeated at scale, buys prompt barrels (lifting the front) and sells deferred ones (pressing the back) until the spread compresses toward the cost of carry. Contango therefore tells you two things at once: prompt supply is comfortable enough that barrels are looking for a home, and the market is paying roughly warehouse rates to house them.

Two corollaries follow. First, when the spread sits below full carry, storing oil for the spread alone loses money; inventories then tend to drain unless someone needs the stock for operational reasons. Second, contango can only blow far past full carry when the arbitrage cannot be executed - when the tanks are already full, the pipelines are committed, and the "buy, store, sell forward" trade has nowhere to put the barrel. That exception is not a footnote. It is the story of April 2020 in section 7.

What contango is not: a forecast that prices will fall. A market in contango is saying prompt barrels are plentiful relative to prompt demand, full stop. It is entirely compatible with deferred prices rising over time - the whole curve can lift while its slope stays positive. Confusing the slope with the level is the single most common curve misread, and section 9 gives it the full treatment.

5. Backwardation: the market paying for barrels now

Backwardation is contango's mirror. When prompt barrels are scarce relative to prompt demand - inventories low, refineries running hard, a supply outage biting - buyers pay a premium for oil they can have immediately rather than a month later. The curve slopes down. The premium for immediacy has a name in commodity economics: the convenience yield, the implicit return to whoever physically holds the barrel. A refiner with crude in its tanks can keep running through a supply hiccup; a refiner without it bids for prompt cargoes. In backwardation, that insurance is expensive, and the curve says so.

Backwardation also reverses the storage incentive. Holding a barrel now costs the carry (rent, financing, insurance) and forgoes the higher prompt price: every month a barrel sits in a tank, its owner could have sold it sooner for more. Inventories therefore tend to draw down in backwardation unless operational needs force stock holding, which tightens the prompt market further until either supply responds or demand cools. Watch the weekly inventory data on our EIA inventory page confirm this in print: sustained backwardation and falling US crude stocks usually travel together.

For producers, backwardation has a hedging consequence worth understanding. Selling future production forward means accepting the deferred price, which in backwardation sits below today's prompt price. That is not the market predicting a fall; it is the market charging for immediacy. Producers who hedge into a backwardated curve lock in less than the headline price, and producers who skip hedging to chase the prompt price are taking the view that prompt scarcity will persist. Either way, the decision starts from reading the slope correctly - as a rental rate on time and storage, not as a prophecy.

6. Worked example: two hypothetical curves

The table below shows two invented curves for a hypothetical benchmark priced near $80. These are teaching numbers, not market data: they exist to make the arithmetic concrete. The storage cost assumption used underneath them - $0.90 per barrel per month all-in for rent, financing, and insurance - is likewise illustrative. Real storage and financing costs vary by location, contract, and credit, and Section 10 explains how to sanity-check the shape without knowing them exactly.

Now apply the carry test from section 4. In the contango example, the prompt spread pays $0.60 per barrel per month. Against the illustrative $0.90 full-carry cost, the "buy, store, sell next month" trade loses about $0.30 per barrel: the market is in contango, but it is not paying anyone to add storage, so this shape says prompt supply is comfortable without being desperate. If the same curve steepened to $1.40 in the prompt spread, the trade would earn roughly $0.50 per barrel per month hedged, tanks would fill, and arbitrage would lean against the spread until it fell back toward $0.90 - unless no empty tanks remained, which is the April 2020 case below.

In the backwardation example the test inverts. Buying the front month and selling month 2 pays -0.60: you would be selling later delivery for less, while still paying carry on the barrel in between. Storage is strictly punished, so the rational holder sells prompt and lets inventories fall - which is why falling stocks and backwardation reinforce each other until supply or demand breaks the loop. The $4.80 drop from month 1 to month 12 also frames the hedging question from section 5: a producer locking in year-ahead output at $75.20 against an $80.00 prompt price is not accepting a forecast of $75 oil; they are paying the market's immediacy premium for certainty.

The habit to build: never quote a curve shape without running it against carry. "In contango" is incomplete information; "in contango at $0.60 against roughly $0.90 of carry, so storage is not being paid to fill" is a reading. You will rarely know the true carry to the cent - financing depends on the borrower's credit, storage on the specific tank farm - but even a rough range separates "the market is relaxed" from "the market is paying a premium to park barrels," and that distinction is most of the signal.

7. April 2020: when the front of the curve broke

The cleanest stress test of everything above happened on April 20, 2020. COVID-19 lockdowns had cut global oil demand by an amount no living trader had seen, US production had not yet shut in at scale, and inventories at Cushing had climbed for weeks. Press reports at the time put Cushing storage at roughly three-quarters full, with much of the remaining space already leased or committed. The May 2020 WTI contract was one day from its last trading day, which meant anyone still long at expiry would have to take physical delivery of 1,000 barrels per contract at Cushing - into tanks they did not have.

The contract settled that day at negative $37.63 per barrel, down $55.90 on the day and the first negative settlement in the history of oil futures, with an intraday low of negative $40.32 (Reuters, April 20, 2020). Sellers were paying buyers to take the delivery obligation off their hands. The very next contract told the other half of the story: June 2020 WTI settled at $20.43 the same day. Subtract the settlements and the one-month calendar spread was negative $58.06 - the market valuing a barrel for June delivery at $58.06 more than the identical barrel for May delivery, a contango so extreme it has its own name in trading slang: super-contango. Brent, cash-settled against an index of waterborne cargoes that can wait on ships, settled in positive territory the same day (June Brent at $25.57): the world's oil was cheap, but only the contract that forced delivery into a full tank farm went below zero.

Read the episode with the tools from sections 4 and 5 and it stops looking like a glitch. The prompt spread was the price of one scarce asset - empty storage at Cushing, available immediately - and that price had blown through full carry because the arbitrage that normally enforces the ceiling requires the very thing that had run out: somewhere to put the oil. The curve was not forecasting that oil would be worth $20 in June and nothing in May; it was reporting, with brutal precision, that in late April 2020 a barrel you had to take now was a liability and a barrel you could take later was still a commodity. Prices along a futures curve are prices for delivery at a place and a time, and when the place and time bind, the price does whatever it has to.

8. Natural gas: the seasonal curve

Natural gas futures carry the same contango and backwardation vocabulary, but the curve's shape is dominated by the calendar rather than by storage scarcity alone. North American gas demand peaks in winter for heating and again, more moderately, in summer for power generation, while production is comparatively steady through the year. The market bridges the gap with underground storage: gas is injected during the traditional refill season, roughly April through October, and withdrawn from about November through March. The futures curve prices that rhythm in advance, so winter months normally carry a premium over the spring and autumn shoulder months even in a well-supplied year.

Two gas-specific habits follow. First, judge a gas curve against its seasonal norm before calling it contango or backwardation: a winter premium is the default state, and the informative signal is whether that premium is wider or thinner than the storage situation justifies. The EIA's weekly natural gas storage report, published Thursdays, is the confirming data - injections running behind schedule into autumn widen the winter premium, while a mild winter that leaves storage full flattens it. Our natural gas price page tracks the Henry Hub benchmark and the storage picture behind it. Second, respect the expiry mechanics at the seasonal seams: the spread between the last winter contract and the first spring contract (the March-April spread) is so sensitive to how much gas is left in storage at winter's end that traders long ago nicknamed it the widow maker. Curve reading in gas is storage reading, with a calendar attached.

9. Five common misreads

Misread 1: "Contango means the market expects prices to fall." No. Contango means later delivery costs more than prompt delivery, which is what you would expect when storage, financing, and insurance are priced in. A contango curve can sit at any price level and can rise or fall as a whole. The slope prices time and tanks; it does not prophesy the level.

Misread 2: "Backwardation is bullish, so prices must keep rising." Backwardation says prompt barrels are scarce right now. Scarcity can end abruptly - a refinery outage cuts prompt crude demand, an outage ends, imports arrive - and when it ends, the premium for immediacy deflates faster than flat prices move. Backwardation is a statement about today wearing a price tag, not a promise about tomorrow.

Misread 3: "The futures curve is the market's forecast." Prices along the curve are tradable delivery prices, not predictions of future spot prices. A December contract at $75 does not mean the market believes spot oil will be $75 in December; it means that is the price, today, for December delivery, given today's storage, financing, and scarcity. Forecasts live in analysis (our forecasts page collects them, labeled as the estimates they are); the curve is a price system, not a poll.

Misread 4: "Oil prices cannot go negative; April 2020 was a data error." It was a settlement, printed by the exchange, on a physically delivered contract one day before expiry, at a hub whose storage was effectively spoken for. Negative prices are an expiry-and-delivery phenomenon: they require a contract that forces someone to take physical product at a specific place. Cash-settled, waterborne Brent did not follow WTI below zero, which is the tell that the constraint was local plumbing, not the global value of oil.

Misread 5: "A steep curve move must mean supply or demand just changed." Check the calendar and the tanks first. Front spreads widen into expiry as positions roll, spike when storage at the delivery point fills or empties, and jump on single inventory reports that reverse the next week. A one-day slope change is a question - usually "what just happened at the delivery point?" - not a regime change. The checklist in the next section orders the questions correctly.

10. How to read the curve today: a checklist

When curve shape makes headlines, work through these steps before forming a view. Each step names the data that confirms or kills the story the slope is telling.

1. Start with the two spreads. Note the prompt spread and the year slope (twelfth month minus first), with signs as defined in section 3. Positive and rising: the market is paying more for patience. Negative and steepening: it is paying more for barrels now.
2. Run the carry test. Compare the prompt spread with a rough all-in cost of carry (storage quotes, short-term interest rates). A spread near or above full carry says tanks should be filling; well below it, storage economics are idle. April 2020 is the standing reminder that the ceiling only holds when empty tanks exist.
3. Check inventories at the delivery point. For WTI, that means Cushing stocks in the EIA's Weekly Petroleum Status Report, published Wednesdays. Our guide to reading the EIA weekly report walks through the tables in order; the analysis page tracks the current series. Rising Cushing stocks with a steepening contango confirm each other; rising stocks with backwardation mean the scarcity sits somewhere else.
4. Ask who wants prompt barrels. Refinery demand is the usual answer. The crack spreads covered in our refinery crack spreads explainer show what refiners earn for turning crude into products; strong cracks with backwardation describe refiners bidding for immediate crude, while weak cracks with backwardation point at a supply problem instead.
5. Compare benchmarks before generalizing. Read the WTI and Brent curves alongside the Brent minus WTI spread (the spread explainer covers that reading). Backwardation in both benchmarks signals global tightness; backwardation in one signals a regional story, and the regional story is usually the tradeable, fixable one.
6. Check the supply response. Sustained backwardation is the market asking for more oil soon. Rig counts on the rig counts page and the Midland differential in our Permian spread guide show whether US producers are answering, and how quickly the answer can physically arrive.
7. Rule out the plumbing. If the move is concentrated in the front month within days of expiry, or in a single seasonal seam in natural gas, assume mechanics before fundamentals. Check what the second and third months did; genuine scarcity reprices the whole front of the curve, not just the expiring contract.

11. Frequently asked questions

What is contango in simple terms?

Contango means oil for delivery later costs more than oil for delivery sooner. The difference is mostly the price of storing and financing the barrel in between, so a market in contango is roughly a market paying warehouse rent: prompt supply is comfortable, and time has a positive price.

What does backwardation signal?

Backwardation - prompt oil priced above later delivery - signals that buyers value immediate barrels more than future ones, usually because inventories are low, demand for prompt supply is strong, or supply has been disrupted. It rewards selling now and drawing down stocks, which is why backwardation and falling inventories tend to appear together in the weekly data.

Can futures prices go negative?

Yes, under narrow conditions: a physically delivered contract near expiry, at a delivery point where storage is effectively unavailable. On April 20, 2020, the May WTI contract settled at negative $37.63 per barrel because holding it meant taking delivery at a nearly full Cushing hub the next day. Cash-settled contracts and contracts with months left to run do not face the same forced-delivery math.

Is contango bearish for oil prices?

Not by itself. Contango describes the relationship between delivery months, not the direction of the price level. Curves in contango have risen, fallen, and gone sideways as whole structures. What contango does tell you is that prompt supply is not scarce - useful context for price views, but not a price view on its own.

Why does the natural gas curve look different from the oil curve?

Gas demand is strongly seasonal and production is comparatively steady, so the gas curve builds in a normal winter premium over shoulder months for storage withdrawals, with injection-season months priced to pay for refilling storage. Read gas curve shape against that seasonal template first; the signal is whether the winter premium is wider or narrower than current storage levels justify.

Where can I see the curve and the data behind it?

Exchange sites (CME Group for WTI, ICE for Brent) publish the full contract calendar. On this site, the markets page tracks the front-month benchmarks, the analysis page tracks the weekly inventory data that confirms what the curve is pricing, and the natural gas price page follows the Henry Hub market and storage. Read the curve for shape, then read the inventories for proof.

12. Related reading

• WTI vs Brent Spread Explained - the other spread every oil reader needs: two benchmarks, one month, and the freight arithmetic between them.
• How to Read the EIA Weekly Petroleum Status Report - the Wednesday data release that confirms or contradicts what the front of the WTI curve is pricing.
• Refinery Crack Spreads, Explained - what refiners earn per barrel, and why their appetite for prompt crude drives backwardation.
• Permian Rig Count vs. WTI Spread Analysis - how the US supply response shows up in regional differentials when the curve asks for more oil.
• Natural Gas (Henry Hub) Price Today - the seasonal curve's home market, with the storage data that shapes it.
• Crude Oil Prices: WTI, Brent, Henry Hub - front-month benchmark quotes, the starting point for any curve reading.