Index Futures Basis (Cash and Carry)

Thesis (edge)

A futures contract on a stock index is a promise to buy the index at a set price on a future date. Its fair price is not a mystery. If you buy the index today with borrowed money and hold it until the futures expiry, you pay interest on the borrowing and you collect the dividends the stocks pay. So the future should trade at the index level, plus the financing cost, minus the dividends you would have received.

That is the whole model. It follows from the fact that if the future traded much above that level, anyone could buy the stocks, sell the future, hold both to expiry, and pocket the difference with no market risk. And if it traded much below, they could do the reverse.

Index arbitrage is exactly that trade. When the future is rich, you buy the basket and sell the future. When it is cheap, you sell the basket and buy the future. At expiry the two converge by construction, because the future settles into the index.

This is the closest thing to genuine arbitrage in equity markets, which is why it is also one of the most competitive.

Where it works (regimes)

• Works, for the right participant. The trade is real and it is done every day, mostly by banks and large market makers who can fund cheaply, replicate the index efficiently and hold the position without stress.
• Barely works for anyone else. The gap between the traded future and fair value is normally small. Your round-trip cost of buying hundreds of stocks and selling a future is usually larger than that gap.
• Gets interesting: when funding markets are stressed. In periods when bank balance sheets are constrained, especially around quarter and year ends, the futures basis can move to levels that imply an unusual financing rate. That is not a free lunch, it is the market pricing the scarcity of balance sheet, and you can only harvest it if you have the balance sheet others lack.
• Breaks: when dividends surprise, when funding costs move against you, or when a margin call forces you out of the futures leg before expiry.

Signals

• Fair value. Index level, plus financing to expiry, minus expected dividends. That is the number.
• The basis. How far the traded future sits from fair value. Most of the time this is small and inside the cost band.
• The implied financing rate. This is the more useful way to look at it. Rearrange the fair value relationship and ask what interest rate the futures price is implying. If the future implies you can effectively borrow at a rate well below your real cost of funds, that is interesting. If it implies you can lend at an attractive rate, that is interesting the other way. Expressing the basis as a rate makes it directly comparable to your actual funding.
• Dividend forecast quality. Half the fair value calculation is the dividend stream. If your dividend forecast is wrong, your fair value is wrong, and you will see a signal that does not exist.

Portfolio construction

• Two legs, held to convergence. The cash leg replicates the index, the futures leg is the hedge, and expiry does the rest.
• The cash leg is the hard part. Buying every constituent is expensive. Using a liquid ETF is far cheaper but introduces the ETF's own tracking error and its own basis. Using a smaller sampled basket is cheaper still but tracks worse. Each choice trades cost against precision, and the choice you make determines whether the trade is viable at all.
• Cash management matters. The trade is fundamentally a financing trade, so how you fund the cash leg and how you invest the margin is not a detail, it is the return.
• Size against your funding, not against your risk appetite. This trade uses balance sheet, and balance sheet is the scarce resource.

Risk model

• Financing risk. Your funding rate can move. Since the entire trade is a bet on financing, an adverse move in your cost of funds directly destroys the position's economics.
• Dividend risk. Companies cut or raise dividends. You have locked in a fair value based on a forecast, and the forecast can be wrong. Special dividends are a particular hazard.
• Margin risk. The futures leg is marked to market daily. If the market rallies hard and you are short futures, you post cash, even though your cash leg has an offsetting unrealised gain that you cannot access. Convergence at expiry does not help you if a margin call forces you out in month two. This is the classic way a supposedly riskless arbitrage kills someone.
• Tracking risk. If your cash leg does not exactly replicate the index, the two legs will not converge exactly, and the residual is a real risk you did not intend to take.
• Competition risk. You are trading against firms whose funding is cheaper than yours and whose execution is faster. If you see a gap and they do not act, ask why.

Costs & implementation

• The cash leg is expensive to trade. Hundreds of names, hundreds of spreads. This cost alone eliminates the trade for most participants.
• Financing spread is the real cost. The difference between the rate you pay to borrow and the rate the market's fair value assumes is, in practice, the whole game. Banks win this trade because they win on funding.
• Margin funding. You must hold cash against futures moves, and that cash has an opportunity cost that belongs in the model.
• The roll. Very few people actually hold to expiry. Most roll the futures position, and the roll has its own basis and its own cost. In practice the roll is where the economics live, and it deserves more attention than the theoretical convergence.
• Data: index composition, weights, a dividend forecast, live futures and cash prices, and your true funding curve.

Failure modes

• Using a textbook risk-free rate. Your funding is not the risk-free rate. Model your real cost of money or the basis you compute is fiction.
• Sloppy dividend forecasts. A large fraction of apparent basis mispricing is really dividend estimation error.
• Ignoring margin. Treating this as riskless because it converges at expiry, then being forced out by a margin call, is a well-worn path.
• Assuming the persistent basis is free money. When the basis has been rich for months, that is information about the price of balance sheet, not an error waiting to be corrected.
• Underestimating the cash leg cost. Backtests that assume you can replicate the index for free produce a strategy that does not exist.
• Forgetting the ETF's own basis. Substituting an ETF for the basket swaps one basis risk for another.

Our Notes & Suggestions

The most valuable insight from this trade is that a futures basis is really a financing rate in disguise. Once you learn to convert the basis into an implied rate and compare it against your own real cost of funds, you can immediately see whether a trade exists for you specifically. It very often does not, and knowing that quickly is worth more than a sophisticated model.

If you are not a bank, the realistic version of this is not classic index arbitrage. It is using the basis as information: knowing when futures are structurally rich or cheap tells you which instrument to use for exposure. If futures are cheap relative to fair value, get your long exposure through futures rather than an ETF. That is a modest, achievable use of the same analysis, and it does not require you to win a competition you cannot win.

Finally, treat "riskless arbitrage" as a warning label. The trades that get people into real trouble are rarely the ones they knew were risky. They are the ones that were supposed to be safe and were therefore sized as if nothing could go wrong.