Cost of Carry in Futures: The Indian Trader's Guide

Walk past a Reliance quote on a normal trading afternoon. The cash share prints at ₹2,840. The next-month future prints at ₹2,855. That ₹15 gap is not random, and it is not a forecast.

It is the cost of carry, priced into the curve by every arbitrage desk in Mumbai.

This article walks through what cost of carry actually means in the NSE F&O book, the formula a desk uses, the rupee math on one Reliance contract, why the next-month future usually trades above the cash, when it flips into a discount, and how the annualised carry number becomes a signal traders watch every morning.

Cost of carry is the price of time

A futures contract lets a trader take a view on a share without paying the full share price up front. Margin sits with the broker, the rest of the money keeps earning interest somewhere else.

That convenience is not free. The seller of the future is the natural counterpart of that bargain. They either hold the share or finance the share in some way, and they have to be paid for the cost of that capital until expiry.

The price they ask is the cost of carry. Multiply the spot price by the interest rate, scale it down to the number of days left until expiry, and that is what the next-month future should trade above the cash market by.

Net out anything the share is expected to pay back to the holder before expiry. The biggest item there is dividends. A share that goes ex-dividend before the contract expires hands a slice of cash to whoever owns the share, not to whoever owns the future, so the future shaves that expected payout out of its premium.

The headline definition that fits on one line is this. Cost of carry equals the interest cost of holding the underlying, minus any income the underlying is expected to throw off, scaled to the days left until expiry.

The formula the desk actually uses

The textbook version of the formula uses continuous compounding. Futures price equals Spot price times e to the power of (r minus q) times t, where r is the risk-free rate, q is the dividend yield, and t is the time to expiry in years.

The working version on a trading screen drops the continuous-compounding wrapper. The arithmetic on a one-month contract is small enough that the simpler form holds.

Futures minus Spot equals Spot times (r minus q) times Days to expiry divided by 365.

That single line is the entire model. Two inputs the trader supplies, two inputs the market supplies.

The trader supplies the risk-free rate, usually the 91-day T-bill yield or the MIBOR, and the dividend yield on the underlying for the period to expiry. The market supplies the spot price and the days remaining on the contract.

Plug in real numbers on a Reliance one-month future. Spot ₹2,840. Risk-free rate around 6.5 per cent, dividend yield close to zero for the month, days to expiry 30.

The interest cost of carrying a ₹2,840 share for 30 days at 6.5 per cent is roughly ₹15 per share. No dividend is expected in the month, so the dividend offset is zero. The theoretical fair futures price is ₹2,855, the lot-level carry is about ₹7,590.

If the next-month future actually prints at ₹2,855, the curve is fairly priced and the arbitrage desks have nothing to do. If it prints at ₹2,860, there is a ₹5 per share excess premium, and someone with the balance sheet to short the future and buy the cash will do exactly that until the gap closes. If it prints at ₹2,848, the future is cheap to the cash and the same desks reverse the trade.

That arbitrage is what keeps the actual basis pinned close to the theoretical carry on liquid Indian names. Reliance, HDFC Bank, ICICI Bank, Infosys, Nifty itself, the basis on these rarely drifts more than a few rupees from fair value for more than a few minutes.

The model bends on dividends. A share that is going ex-dividend before the future expires will trade with the dividend baked out of the futures price weeks before the ex-date. ITC announcing a ₹15 dividend with twenty days to expiry will see the future trade ₹15 closer to the spot, the carry effectively wiped out for that month.

Contango, backwardation and the shape of the curve

Three futures contracts on the same underlying are listed at any moment on the NSE. Near month, next month, far month. Lay their prices on a chart against the cash market and the line that connects them is called the futures curve.

In a normal market the curve slopes up to the right. The next month trades above the cash, the far month trades above the next month. The shape is called contango, and it just reflects the fact that each further-out contract carries the underlying for a longer period at the same interest rate.

Contango is the default state of an Indian equity contract. A typical Reliance or HDFC Bank curve in 2025 shows a ₹15 spread between cash and the near month, another ₹15 between the near month and the next, scaling roughly linearly with time.

The reverse case is backwardation. The next-month future trades at a discount to the cash market, the curve slopes down to the right.

Three things flip the curve into backwardation on an Indian equity contract.

One. A heavy dividend. ITC, Coal India, the public-sector banks, the IT majors at fiscal-year-end. The expected payout gets baked out of the next-month future weeks before the ex-date, and on a high-yield stock that can drag the future below the cash.

Two. A sharp bearish move. A market in freefall sees aggressive short-selling in the most liquid futures, pushing the basis negative even though dividends have not changed. The March 2020 covid window had Nifty future trading at a discount to the cash for several sessions in a row.

Three. Stock-specific selling pressure. A promoter unwinding, an FPI exiting a single name, or a fund forced to liquidate a position quickly often dumps it through the futures market first because the leverage is cleaner. The basis tells the story before the cash chart does.

None of these are mysterious. Each one has a clean reason behind it, visible on the corporate-actions calendar or in the F&O participant-wise data the NSE publishes after market hours.

How traders actually use the carry number

On a real trading desk the carry is not used to forecast where the future will close. It is used as a thermometer. A reading that drifts outside its normal band says something has changed under the surface.

The number a desk watches is not the rupee basis. It is the annualised cost of carry, the basis re-expressed as a yearly percentage. Annualised carry equals (Future minus Spot) divided by Spot, times 365 divided by Days to expiry, times 100.

The same ₹15 basis on a 30-day contract works out to about 6.4 per cent annualised. On a five-day-to-expiry contract the same ₹15 prints near 38 per cent. The annualisation is what makes the number comparable across days.

Three patterns repeat often enough that every Indian equity-derivatives desk has them mapped.

Carry rising on rising open interest. The future is gaining a premium even as fresh longs are being added. Aggressive buyers are happy to pay above fair value, and the curve is telling you long conviction is back in the name.

Carry collapsing on rising open interest. Fresh positions are being added on the short side, pushing the future toward or below the cash. The curve is telling you the fresh money has a bearish view, and the cash chart usually follows within a session or two.

Carry stable while open interest falls. Both sides of the existing book are getting closed out, the basis is unchanged because nobody is pressing either way. The signal is consolidation, not direction.

None of these calls work in isolation. They earn their weight when they line up with what the cash chart, the FII derivatives data, and the broader index basis are saying. A single carry print in a single name is noise. A pattern across the top ten F&O names, day after day, is signal.

The same logic loops back into the sizing decision covered in the guide on initial and exposure margin. A trader who is paying carry every month on a multi-month thesis has to size the position so the carry is a small fraction of the move the thesis expects to capture.

Three places where cost of carry quietly costs you

The carry is small. That is exactly why it gets ignored, and exactly why it adds up.

One. Buying the future when the share would do. An investor who plans to hold for six months and uses the next-month future instead of the cash share pays carry six times. On Reliance at a 6 per cent annualised carry, that is roughly 3 per cent of the position value over the period, just for the privilege of leverage that a long-only investor does not need.

The future is a leveraged instrument. If the leverage is not being used, the carry is a tax paid for nothing.

Two. Choosing the wrong contract on a stock with a known dividend. A trader bullish on ITC just before the ex-dividend date buys the next-month future and is surprised when the future barely moves on the day the dividend is announced. The market had already priced it in, and the position made the right call on the company while making the wrong call on the contract.

The fix is to read the corporate-actions calendar before sizing the trade, not after. The guide on stock futures covers the dividend mechanics in detail.

Three. Rolling a high-carry position three months in a row. A bullish thesis on a small-cap F&O name with a thin float and a high carry, rolled across three expiries, can quietly pay away most of the upside in carry alone. The chart says the position is up 8 per cent, the account says the position is up 2 per cent, and the gap is the carry.

None of these mistakes are about getting the direction wrong. They are about the small operational decisions on top of a correct directional call.