On Tuesday 16 Jun 2026 Nifty closed at 23,989.15. The June future settled at 24,001.00, 11.85 points above the index, with 14 days to run. Interest on the index for those days, at the bill yield of 5.17%, comes to 47.5 points. The future sat 35.7 points short of that. A premium smaller than interest is not a bearish vote. It is a sum with one line you cannot read off a screen: the dividends the shares pay before expiry. I read a futures price as that sum before I read anything into it. This lesson builds the sum, solves it backwards for the missing line, and shows why the answer wobbles. Every figure comes from NSE end of day files and is dated, not live.
3By the end you can
- Work out a futures fair value from the spot, the rate, the days and the dividend points.
- Solve the sum backwards to the implied dividend and say what else it holds.
- Explain why a gap to fair value is a reading of the model and not a free trade.
The idea, plainly
What is fair value?
Fair value is the price at which a future costs the same as owning the index until expiry. Owning the index ties up money, so it costs interest. It also pays dividends. A future avoids both. So the fair price of the future is the spot, grown at the interest rate, less the dividends the shareholder collects before expiry. That answers the question behind most premium posts: what is cost of carry, and why does Nifty futures sit at a premium? The premium is the carry, minus the dividends.
Fair value =
(spot less dividends
in today's money)
x e^(rT)
Read one more paragraph, 96 words
Here r is the interest rate compounded continuously and T is days to expiry over 365. Module 13.2 showed the same idea with simple interest. Dividends enter in index points. A stock's dividend, in points, is its weight in the index times its dividend as a share of its price, times the index level. Add up the stocks that go ex-dividend before expiry. NSE does not publish that total for each expiry. The Nifty 50 factsheet gives a dividend yield for the index as a whole, which is a starting point and not an answer.
Solve for the missing line
Dividends are the one unknown, so turn the sum around. Take the dividends that make fair value equal the traded future and call them the implied dividend, in points. It is what the model needs to explain the price, nothing more. If the market priced only interest and dividends, it would be the dividends. It also soaks up every other friction: the cost of arbitrage, demand for long or short exposure, and the gap between the bill yield and what traders pay for money. I use the bill yield as a proxy rate. A fund that borrows above it has a higher fair value.
Grow the index at the bill yield, then read the gap as dividends
Read 2 more paragraphs, 106 words
Timing of the dividends matters a little. A dividend that goes ex early in the window is brought back from almost the whole period, and a late one from a short stretch. Over two weeks the difference is tiny. Over many months it grows, which is why the engine behind the tool can take a single lump of points or a dated list.
On 16 Jun the June future implied 35.7 points of dividends and the July future 43.8. July holds more days, so more companies can go ex, and its number is larger. That reads sensibly. The next section checks whether it always does.
What stops the future drifting away?
Traders who can buy the whole basket and sell the future, or hold the shares already and do the reverse, keep the price near fair value. They act only when the gap is larger than what the round trip costs them: charges on both legs, funding and the price of trading fifty stocks at once. So fair value is not a line the future sits on. It is the middle of a band, and the width of the band is the cost of the trade. A reading inside the band proves nothing is mispriced. A reading outside it is the case the cash futures arbitrage lesson studies.
Worked example: four futures on two days
Here are the June and July futures on 10 Jun and 16 Jun 2026. Premium is the future less the index. Interest is spot times rate times days over 365, as in module 13.2. The last column turns the sum around, using the continuous formula above.
| Contract | Days left | Premium, points | Interest, points | Implied dividend, points |
|---|---|---|---|---|
| 10 Jun, June future | 20 | 25.15 | 65.0 | 39.9 |
| 10 Jun, July future | 48 | 129.95 | 157.2 | 27.7 |
| 16 Jun, June future | 14 | 11.85 | 47.5 | 35.7 |
| 16 Jun, July future | 42 | 98.75 | 142.2 | 43.8 |
The premium ranged from a quarter of interest to more than all of it
The next month premium wandered around the bill yield for a year
Read 5 more paragraphs, 337 words
Take the first row for 16 Jun, step by step. Grow the index close of 23,989.15 at 5.17% for 14 days and you get 24,036.72. The traded future settled at 24,001.00. The difference, 35.7 points once it is brought back by half the period, is the implied dividend. Nothing else in the sum was free to move, which is why every friction ends up in that one number.
On 10 Jun the order was odd. The longer July window implied fewer dividend points than June, 27.7 against 39.9. A longer window cannot hold fewer dividends, barring new announcements. So part of what the plug holds that day is not dividends. Six sessions later the order had flipped. The June plug fell by 4.2 points and the July plug rose by 16.1.
A cleaner way to see how much a premium differs from interest is to divide one by the other. A bar near one hundred percent means the premium was about equal to interest, and no dividends are left to explain. A bar below it leaves room for dividends. A bar above it asks for dividends below zero.
On 7 Apr 2025 the April future's premium was 102.2 points against 61.0 of interest with 17 days left. That is more than all of the interest, so the implied dividend was below zero, which no company pays. India VIX closed at 22.79 that day, up sharply, and an end of day file cannot say which friction did the work. Seen across a year, the pull between the premium and interest never settles on one number.
The fair value tool below opens on the June future of 10 Jun 2026. It starts with the dividend points that make the model equal the traded future, so you can move them and watch the gap open. It also draws a band after costs, which the cash futures arbitrage lesson builds. For the short answer to the cost of carry question, see the cost of carry page.
The common mistake
The belief: A future priced away from fair value is mispriced, so the gap must close for free.
Fair value is a model with one unobserved input and a proxy rate, so a gap is a reading of the model and not proof of a mistake. On 10 Jun the June plug exceeded the July plug by 12.2 points. Six sessions later the order had reversed. New announcements can change a dividend estimate, so I cannot rule them out, but a swing this size in six sessions looks more like the rest of the plug. Nothing about such a gap says it closes before expiry on its own. Closing it also costs money: delivery STT on the shares, charges on the future, funding. The cash futures arbitrage lesson puts those costs against the gap.
Key takeaways
- Fair value is spot grown at the rate, less dividends in today's money.
- Dividend points to expiry are not published, so they are estimated.
- The implied dividend is a plug that also holds frictions and demand.
- A longer window should not imply fewer dividends than a shorter one.
- A premium under interest needs no bearish view to explain it.
- A gap to fair value is not free: costs sit against it.
3 terms this section turns on
Before you trade. Derivatives trading carries a high risk of loss. SEBI's study "Profitability of Individual Traders in the Equity Derivatives Segment (FY25 to FY26)" (August 2026) found that 87.7% of individual traders in equity F&O made net losses in FY26, with aggregate net losses of Rs 91,685 crore. Read the SEBI study