Contango and backwardation: why a long futures position melts away

A long futures position melts away not because the underlying asset falls, but because by the settlement date the futures price has to converge with the spot price. If the future trades above the underlying asset — this is contango — the premium does not belong to the buyer: it is extinguished as expiry approaches. While the underlying asset stands still, the future slides down towards it, and the holder of the long position pays for that slide. When they roll the position into the next contract, they buy a new premium and pay for it all over again. The erosion builds up with every roll, and it is this erosion, rather than a wrong call on direction, that most often explains why "the asset is up, but the position is down".

The basis: what exactly you are buying on top of the asset

The difference between the futures price and the price of the underlying asset is called the basis. It arises not from a forecast, but from the cost of owning the asset until the delivery date. Whoever sells you the future and hedges it by buying the asset on the spot market bears costs: the money for the purchase carries interest, and a commodity needs a warehouse, insurance and an allowance for shrinkage. The seller builds these costs into the price of future delivery. For financial assets — shares and indices — the main cost component is one in nature: the price of money. For commodities, the physics of storage is added on top.

The opposite situation is backwardation, when the future is cheaper than spot. It means that owning the asset now gives the holder something a paper contract does not: the commodity is needed in production immediately, the spot market is squeezed dry, or the share will pay out a distribution before expiry that the futures buyer will not receive. A detailed definition of the terminology is in the glossary: contango; the term card is available at the contango glossary entry.

Contango is not a forecast of a rise, backwardation is not a forecast of a fall

The most expensive misconception in this subject is to read the shape of the curve as the market's opinion on the future price. The curve is built by arbitrage, not by expectation. If the future deviates from spot by more than funding and storage cost, an arbitrageur will simultaneously sell the expensive side and buy the cheap side — and the deviation will close. That is why contango almost always speaks about the interest rate and the cost of carry, not about the asset getting more expensive.

An exception exists where arbitrage is physically impossible: in assets that cannot be bought and put into a warehouse — weather, volatility, network capacity. There the curve really does contain an expectation, because it cannot be closed with a trade. A similar logic is examined in the piece on the estimated nature of the hash rate.

The regulator's rate as the main engine of the basis

The basis on equity and index futures moves with the price of money. The more expensive the funding, the wider the contango, the more it costs to carry a long position and the more profitable the opposite side of the trade becomes. That is why the holder of a derivatives position needs to watch not only the rate decisions but also the wording around them — a separate piece covers this: how to read the regulator's signal. The same cause explains why the basis and yields on the debt market move in step: OFZ and the futures basis are fed by a common rate, and the yields can be viewed in the OFZ section.

Dividends: where backwardation in Russian shares comes from

The buyer of a future on a share does not receive the dividend — the owner of the stock does. So the payout that will take place before expiry is subtracted from the futures price. The closer the record date and the larger the expected payout, the deeper the future sinks below spot. For stocks with a notable dividend history — for example, {{instrument:SBER}} — the summer contracts traditionally live in backwardation, and after the record date it disappears, because there is nothing left to subtract.

Hence the practical conclusion: backwardation ahead of the record date is not a discount and not a gift. It is exactly the payout that you will not receive. Check the dates in advance: {{dividend_calendar|limit=5}}, the full schedule is in the dividend calendar, and corporate events are in the events calendar.

Why this hits commodity funds

A fund that holds not the commodity itself but a future on it is obliged to roll from the expiring contract into the next. In a prolonged contango every roll gives away part of the unit's value, even if the price of the commodity has not changed. That is why the performance of such a fund and the performance of the spot commodity diverge more and more the longer you hold. The composition and mechanics of the instruments can be found in the funds section.

How to look at this in practice

Step 1 — compare the price of the nearest contract and the next, and work out which way the curve slopes. Step 2 — ask what explains the slope: the price of money, storage or an upcoming payout. Step 3 — calculate what it will cost to carry the position out to your horizon, and compare that with your idea. If the idea does not survive the cost of rolling, it is not an idea but a bet that the basis will narrow.

Far-dated contracts are thinner than near-dated ones: spreads are wider, slippage is more noticeable, and getting out of a position may cost more than getting in — see why liquidity is noticed too late. The time of day matters too: the basis behaves differently in the morning and evening sessions, which is covered in the trading day and sessions. The financial result on derivatives trades goes into the tax base under the rule 13%.