In the spring of 2020, loaded tankers piled up off the coast of California and Singapore, and some remained at anchor there for weeks. Oil demand had just collapsed, the May WTI contract had closed at −$37.63 on April 20, and the maturity curve had been distorted to the point of absurdity: the barrel deliverable six months later was priced around ten dollars above the barrel available immediately, sometimes more. The traders who filled these ships, however, were not betting on oil. They bought for cash, sold the long term, tied up the cargo and collected the difference, reduced by a charter cost which soared as available ships became scarce.
It's cash-and-carry, and the name says exactly what the operation does: we buy the asset, we carry it, and at the same time we sell a contract which obliges it to be delivered later at a price fixed today. The strategy focused on gold, copper and wheat well before the invention of bitcoin, and it settled on the latter as soon as the first futures contracts appeared, several years before the CME opened its own at the end of 2017. The previous part of this series described its perpetual version, the one that lives on funding. The dated version differs on one point which changes everything. Where the gap comes from, how to reduce it to an annual return, what remains after margin and fees, and what can break before maturity: this is the subject of this second part.
This article is brought to you by DCY. To learn about their approach, visit dcy.fund.
Key Points
A dated contract converges mechanically towards the cash at maturity, so that the profit of the operation is known from the start: it is the initial difference between the two prices, what professionals call the basis.
A base of 4,000 dollars on a bitcoin of 80,000, for a maturity of 91 days, is worth 5% over the quarter, or 20.1% annualized.
The margin deposited on the selling leg brings this figure to 16%, even before counting the costs paid on two legs, on the outward and return journeys.
The base remunerates the immobilization of capital, then the demand for buyer leverage: it exceeded 20% annualized at the start of 2021, collapsed in 2022 until the fall of FTX and can become negative.
Oil tankers going nowhere
The episode of the oil tankers at anchor has an educational virtue that bitcoin does not have: we see the stock, we see the ship, and we immediately understand that the person setting up the operation has no opinion on the price of the barrel. Its result does not depend on the direction of the market but on one thing alone, the convergence of the two prices that it has placed face to face.
Let's transpose. Bitcoin prices 80,000 dollars in spot and the quarterly maturity shows 84,000, a deliberately generous gap, which we only come across in times of euphoria. Two orders, placed together: we buy 1 BTC in cash, we sell the equivalent of one bitcoin on expiry. No storage to pay, no ship to charter, but the same mechanics and the same indifference to the price.
Four thousand dollars, whatever happens
A dated contract has a property that the perpetual does not have, and it is on this that the whole reasoning is based: at maturity, its price joins that of the cash. He has no choice, since it concerns the same asset delivered at the same time. Let’s run through the two scenarios.
Bitcoin falls to 60,000. Spot leg loses $20,000; the selling leg, opened at 84,000 and bought back at 60,000, earns 24,000. Net result: 4,000 dollars.
Bitcoin rises to 100,000. Spot leg gains $20,000; the selling leg, opened at 84,000 and bought back at 100,000, lost 16,000. Net result: 4,000 dollars.
The arrival price therefore never enters into the equation. What we collect is the initial $4,000 difference, and this difference was displayed on the screen when entering. This is the basis, and its particularity is that it can be read before the operation rather than discovered afterwards.
A raw difference means nothing until it is related to a duration. Here, 4,000 dollars out of 80,000 is 5% over the quarter. To compare this figure to anything else, you have to annualize it: 5% multiplied by 365, divided by the number of days remaining until maturity. Over 91 days, this gives 20.1% per year, or 21.6% if we assume the operation is repeated four quarters in a row, gains reinvested, on an identical basis.
The formula fits in one line, and any reader can check it this evening on the public quotes of the expiry of their choice: divide the contract price by that of the spot, subtract 1, multiply by 365 divided by the number of days remaining. The result is the annualized return available at the time of calculation. It changes every minute, and this is precisely why a figure displayed in a commercial brochure always merits asking for the date of its reading.
Margin and fees: what’s left of the yield
There remain the two adjustments already applied to perpetual in the previous part, and they are not cosmetic. The first concerns the basis of calculation: the return is measured on the capital actually invested, never on the notional amount. To carry this position, you must take out $80,000 in cash and deposit margin for the selling leg. If this margin weighs a quarter of the cash value, the immobilized capital reaches $100,000 and the $4,000 gain is no longer worth 5% but 4% over the quarter, or 16% annualized instead of 20.1%. The second concerns execution costs, paid on two legs at entry and exit, which are subtracted from a gain whose gross amount will no longer change by a cent.
Known is not acquired
This is the difference with perpetual mechanics, and it is considerable. On a dated contract held to the end, the yield is neither estimated nor extrapolated from a past average: it is known. The funding of a perpetual is discovered payment after payment, so that any yield announced on this basis remains an estimate, never a commitment.
This certainty, however, has a counterpart that must be written in black and white: it only applies to those who hold on until maturity. Between today and convergence, the platform reassesses the position every day, and nothing prevents the base from moving further before closing. The final gain does not change; the path to get there, yes. Also, cash-and-carry is a strategy whose result is known in advance but the trajectory is uncomfortable, while investors often believe they are buying both at the same time: a fixed result and a smooth journey.
What can break before maturity
The risks of the opening component apply here, with the exception of funding becoming cost, which does not exist on a dated contract. This adds or aggravates others, and it is on these that we must stop.
Margin call comes first, and it often surprises those new to the strategy. If bitcoin goes from $80,000 to $100,000, the selling leg immediately shows a loss of at least $16,000, and the platform immediately demands coverage. The gain of $20,000 on the spot leg does exist, but it lies dormant in another account, often on another platform, and it can only be used as collateral once it has been moved. Add to that a broadening of the base, and the bill grows even bigger. Maintaining a cash-and-carry involves keeping in reserve enough to finance these calls without touching one or the other of the two legs, and accepting that this reserve further lowers the real yield.
If this reserve is missing, it is the liquidation of the selling leg which threatens, and with it the guarantee of convergence. The spot position finds itself alone, entirely directional, in a market which has just made a violent movement. The session of October 10, 2025 reminded us on a large scale: a blanket only protects if it survives the day.
A certainty to be renewed every quarter
Renewal is the risk specific to the dated contract, and the least visible. The yield is known for one maturity, not for the next. At the end of the quarter, it is necessary to reopen the position on the next contract, at the moment's basis, which may have halved or even disappeared. The 21.6% compound calculated above assumes four identical bases in a row, which has virtually never happened. The certainty of cash-and-carry is renewed quarter by quarter, and its price with it.
There remains the counterpart, already described in the opening section and which no basic calculation neutralizes. The funds that held their selling leg at FTX in November 2022 had a perfectly calculated base; it was of no use to them.
Who pays for this gap, and until when
The question is the same as in the opening section, and the answer too, with one nuance. In crypto, leverage is almost always taken in the same direction, upwards. An investor who wants to amplify a conviction can borrow to buy spot, or buy the futures contract and only lock in a margin. Many choose the second solution, and this preference mechanically pushes the contract price above spot. Part of the base simply remunerates the immobilization of capital, approximately at the monetary rate in dollars; it is the surplus which constitutes the price of leverage, and whoever collects it does nothing other than provide it. A cash-and-carry return is therefore compared to the risk-free rate, never to zero.
The basis, market thermometer
Its recent history therefore reads like a thermometer. It turned negative in March 2020, at the height of the panic. It exceeded 20% annualized at the start of 2021, when euphoria was general, then collapsed throughout 2022, until it became almost zero after the fall of FTX, when no one wanted leverage anymore. From the end of 2023, it recovered, driven by the wait and then the launch of spot bitcoin ETFs in the United States in January 2024, to the point that hedge funds accumulated a net short position of several billion dollars on CME contracts, backed by ETF shares: exactly the setup described here, on an industrial scale.
Cash-and-carry will never produce a spectacular return, and it dies when long leverage is withdrawn. It offers something else, which remains a rare commodity in this profession: a return whose amount we know before starting. You still have to agree to immobilize your capital until maturity, and to start all over again the next time, at the current price.
