Contango and Backwardation: The Arithmetic Behind the Drift
4 steps · one page
In short
A futures curve is simply the set of prices for delivery at different future dates, and it has a shape.
When later months cost more than nearer ones, the curve is in contango. When later months cost less, it is in backwardation. That shape determines the return of anyone holding a rolling futures position, and it is entirely independent of whether the commodity price rises or falls. This article is the pillar's arithmetic centre: three earlier articles have shown a holder losing money on a favourable commodity move, and this one explains exactly why and quantifies it in both directions.
Why curves have shapes
Start with the mechanism that makes the shape possible. The energy article established the key relationship: storability is what links a futures price to a spot price. If you can buy the physical commodity now, store it, and deliver it later, then the future price cannot drift far from the spot price plus the cost of carrying it — arbitrage closes the gap. So contango is the normal state for a storable commodity, and the premium in later months reflects cost of carry: storage, insurance, and the financing cost of the capital tied up. That is not a market inefficiency; it is the price of not having to store the thing yourself. Backwardation is the more interesting state, because it means someone is paying more for the commodity now than for a promise of it later. That happens when physical availability is genuinely scarce — a user who needs the material this month to keep a plant running will pay a premium over a contract for next month, and that premium is called the convenience yield: the value of having the actual substance in hand. So a backwardated curve is a signal about physical tightness, and it is common in commodities that are hard to store, which is why energy curves flip between states more readily than metal curves. Three points that follow. The curve is not a forecast. A curve in contango does not mean the market expects prices to rise — it means carrying costs are being priced. Reading a curve as a prediction is one of the more common errors in this area. Shapes change. A curve can move from contango to backwardation and back within a year, particularly in energy and agriculture, so a holder's roll experience is not a fixed property of the commodity. And the front of the curve matters most for retail products, since most track near-dated contracts and therefore experience the steepest part of the shape.
The arithmetic, in both directions
Here is the mechanism in one sentence: a rolling position sells an expiring contract and buys a later one, and if the later one is more expensive, the same money buys less exposure. Repeat monthly and the effect compounds. Nothing is charged, nobody takes a fee, and the position simply owns fewer units of exposure each month. The reverse holds in backwardation: the later contract is cheaper, so the same money buys more exposure, and the position gains without the commodity moving. The reference table below is computed and is this pillar's single source for roll figures. It shows the annual effect of twelve monthly rolls at various curve spreads.
| Monthly spread | Annual effect — contango | Annual effect — backwardation |
|---|---|---|
| 0.3% | −3.5% | +3.7% |
| 0.5% | −5.8% | +6.2% |
| 0.9% | −10.2% | +11.5% |
| 1.2% | −13.3% | +15.6% |
| 2.0% | −21.2% | +27.4% |
| 3.0% | −29.9% | +44.1% |
Three observations from the table, each of which matters. The effect is large at spreads that sound trivial. A 0.5% monthly spread — half of one percent, easily dismissed — costs 5.8% a year. The effects are asymmetric. Backwardation adds slightly more than contango subtracts at the same spread, because the arithmetic is multiplicative rather than additive. And they compound across years, which is where the real damage sits: at the canonical 1.2% contango with a 0.75% charge, a holder facing a completely flat commodity price loses 14.0% in one year, 36.4% over three, 52.9% over five, and 77.8% over ten. Two consequences a reader should carry. The breakeven is not zero. At that same canonical contango, the commodity must rise 16.3% in a year for a holder to merely break even — so "I think this commodity will go up" is not sufficient; the question is whether it will go up by more than the structure removes. And this is not a fee, which is why it is dangerous. An expense ratio appears in a document and can be compared. Roll drag appears in no fee table, is not a charge, and is invisible until a holder compares their return to the spot price and finds a gap nobody mentioned. Two honest qualifications. Backwardation is real and has been persistent in some commodities for extended periods — a holder in those conditions earns a positive roll return, and presenting the roll purely as a cost would misrepresent it. And some funds use alternative roll strategies, spreading across multiple maturities or selecting contracts to reduce the effect. These can help and do not eliminate the underlying shape; a reader should check what strategy a product actually uses rather than assuming.
Worked example
Worked example (fictional; all figures computed). Omar holds $50,000 in a futures-tracking fund on a commodity at $75, ongoing charge 0.75%. Scenario one: contango at 1.2% monthly, commodity flat. Twelve rolls cost 13.3%; with the charge, he ends at approximately $43,005 — a 14.0% loss on an unchanged commodity. Scenario two: same curve, commodity rises 10%. The gain is met by the roll, leaving approximately −5.4%. He was right and still lost. Scenario three: same curve, commodity rises 20%. Now he is above the 16.3% breakeven, ending at roughly +3.2% — a 20% commodity move converted into a 3% return. Scenario four, at equal prominence: backwardation at 0.9% monthly, commodity flat. The roll adds 11.5%; after the charge he ends at approximately $55,320, a 10.6% gain on a commodity that did not move at all. And the decade. Held ten years through persistent 1.2% contango on a flat commodity, the $50,000 becomes roughly $11,100. The commodity never fell. The position lost more than three-quarters of its value to a structure that appears in no fee schedule. (All names fictional; figures computed from the reference table above, with the charge applied multiplicatively to each year's result.)
Frequently asked
10 questions
What are contango and backwardation?
Descriptions of a futures curve's shape. Contango is when later months cost more than nearer ones; backwardation is when they cost less. The shape determines the return of anyone holding a rolling futures position, independently of whether the commodity price rises or falls.
Why is contango normal?
Because for a storable commodity you could buy now and store, so the later price can't drift far from spot plus the cost of carrying — storage, insurance, and financing. The premium in later months is the price of not having to store the thing yourself, which is a cost rather than an inefficiency.
What does backwardation mean?
That someone is paying more for the commodity now than for a promise of it later — which happens when physical availability is genuinely scarce. A user who needs material this month to keep a plant running pays a premium over next month's contract, and that premium is the convenience yield. A backwardated curve is a signal about physical tightness.
Does contango mean the market expects prices to rise?
No, and this is one of the more common errors. The curve isn't a forecast — a contango curve is pricing carrying costs, not predicting appreciation.
How does the roll actually cost money?
A rolling position sells an expiring contract and buys a later one. If the later one is more expensive, the same money buys less exposure — repeat monthly and it compounds. Nothing is charged and nobody takes a fee; the position simply owns fewer units each month.
How big is the effect really?
Larger than the spreads suggest. A 0.5% monthly spread costs 5.8% a year. At 1.2% monthly with a 0.75% charge, a flat commodity produces −14.0% in one year, −36.4% over three, −52.9% over five, and −77.8% over ten.
So what return do I need just to break even?
At that canonical contango, the commodity must rise 16.3% in a year. Which means "I think this will go up" isn't sufficient — the question is whether it will go up by more than the structure removes.
Why is this more dangerous than a fee?
Because an expense ratio appears in a document and can be compared. Roll drag appears in no fee table, isn't a charge, and stays invisible until you compare your return to the spot price and find a gap nobody mentioned.
Is the roll always negative?
No. Backwardation is real and has persisted in some commodities for extended periods, and a holder in those conditions earns a positive roll return — on the illustration here, +10.6% on a completely flat commodity. Presenting the roll purely as a cost would misrepresent it.
Do some funds avoid this?
Some use alternative roll strategies — spreading across maturities or selecting contracts to reduce the effect. These can help and don't eliminate the underlying curve shape. Check what strategy a product actually uses rather than assuming.
References
- CFTC — Customer Advisory: Learn About Risks Before Investing in Commodity ETPs or Funds (roll mechanics; rising prices can still produce a drag) —
- CFTC — What is a Bitcoin Futures ETF? (contango and the roll premium, defined by the regulator) —
- FINRA — Exchange-Traded Funds and Products (commodity pools and other exchange-traded products) —
- EIA — U.S. Energy Information Administration (energy market data) —
Educational and informational only — not investment advice, a recommendation, or an offer to buy or sell any security. Investing involves risk, including the possible loss of principal. Worked examples use fictional companies and figures.