Why selling an option needs so much more margin than buying one

Buy one lot of a ₹40 NIFTY option and it costs you ₹40 × 75 = ₹3,000. Sell the same contract and your broker blocks something in the region of ₹1,50,000.

Same contract. Same strike. Same expiry. Fifty times the capital requirement.

New traders usually read this as a broker being difficult, or as the market penalising sellers. It is neither. It follows directly from what the two positions actually are, and understanding it properly is the difference between sizing a short option sensibly and blowing up an account on a trade that "only" collected ₹3,000 of premium.

The two positions are not mirror images

When you buy an option you are buying a right. The most you can lose is what you paid. If the trade goes maximally against you, the option expires worthless and you are out ₹3,000. There is a floor, you have already funded it, and there is nothing left to collateralise. Hence: margin = premium.

When you sell an option you are taking on an obligation. You collect ₹3,000 today, and in exchange you have promised to settle a difference that is bounded only by how far the index moves. If NIFTY gaps 500 points against a short call with 75 units per lot, the settlement is on the order of ₹37,500 — twelve times the premium you collected, on one lot.

Nobody can know that number in advance, so the system does not try. It asks instead: how much would it take to cover a bad but plausible day? And the answer to that question has nothing whatsoever to do with the premium.

Margin is sized off the underlying's notional

This is the whole idea, and it is worth stating as plainly as possible:

Short-option margin is a percentage of the value of the underlying you have undertaken to deal in — not a percentage of the premium you received.

Concretely. NIFTY at 24,500, lot size 75. The notional value of one lot is:

24,500 × 75 = ₹18,37,500

The margin the exchange requires is computed as a percentage of that, roughly in the region of 10–15% once the standard SPAN and exposure components are added together. Call it 12% for the sake of arithmetic:

₹18,37,500 × 0.12 ≈ ₹2,20,500

The premium was ₹3,000. The margin is seventy-three times it. Not because anyone is being punitive, but because ₹3,000 is not remotely enough collateral against an obligation on ₹18 lakh of index.

The same logic explains a fact that confuses people even more: a far out-of-the-money short option requires almost as much margin as an at-the-money one. A ₹2 option and a ₹200 option on the same underlying and expiry have margins within shouting distance of each other, because the margin is driven by the notional they share, not the premium that differs by a factor of a hundred. Selling cheap options is not a cheap position. It is the same position with a smaller reward.

What this does to your return arithmetic

Once you see the margin as the real capital commitment, the economics of option-selling look very different from how they are usually pitched.

On the numbers above: collect ₹3,000, block ₹2,20,500. If the option expires worthless and you keep the whole premium, that is a 1.36% return on capital deployed, before charges. Charges on a round trip — at least ₹40 of brokerage plus STT on the sell leg plus exchange fees and GST — will take a visible bite out of it.

So the honest description of a naked short option is: a position that ties up a large amount of capital to earn a low single-digit percentage, most of the time, with an occasional loss many multiples of the gain. That can be a perfectly good business. It is a completely different business from the one implied by "I collected ₹3,000 on a ₹3,000 margin", which is what a badly-modelled simulator will show you.

And it is why the reward-to-risk framing that works for long options fails here. Your risk is not the premium. Your risk is the position size the margin is telling you about.

Spreads, and why the margin drops

Sell a call and simultaneously buy a further out-of-the-money call, and the margin falls dramatically — often to a small fraction of the naked requirement.

The reason is the same reason as before, read in reverse. With the long call in place, your obligation is no longer open-ended: above the higher strike, the two positions offset each other and your maximum loss is a known, finite number (the difference between strikes, minus the net premium, times the lot size). Once the worst case is bounded and provably bounded, the collateral only needs to cover that bound.

This is the mechanical, non-ideological argument for spreads. It is not that spreads are more sophisticated. It is that a defined worst case is cheaper to collateralise than an undefined one, and the capital you free up is real.

Three practical consequences

Size on margin, not on premium. Before you sell anything, ask what fraction of your account the margin represents. Two lots of a short NIFTY option at ~₹2.2 lakh each is ₹4.4 lakh of margin. On a ₹5 lakh account you are at 88% deployment on one directional view. The fact that you only collected ₹6,000 is irrelevant to that.

Intraday relief is not a smaller risk. Brokers offer reduced intraday margin on F&O. The reduced number is a funding concession, not a change in your exposure. Your obligation on a gap is identical; you simply have less collateral behind it. An intraday-margined short is more dangerous than a fully-margined one, not less.

No underlying price means no defensible margin number. This is a subtle one worth naming. To size a short-option margin you need the underlying's current price — that is the notional the whole calculation rests on. If you cannot get it, you cannot compute a margin. The correct behaviour then is to refuse the order, not to fall back on some multiple of the premium. A simulator that guesses in that situation will let you take a position you could never have carried in a real account, which is the one thing a simulator must never do: this app refuses the order instead, for exactly that reason.

The concentration problem margin does not catch

One last thing, because it is the failure mode margin rules were never designed to prevent.

Margin limits how much notional you can control. It says nothing at all about whether committing that much to a single view is sensible. A trader who sells the maximum number of lots their account permits, on one strike, in one expiry, has satisfied every margin rule in the system and taken a spectacularly concentrated bet.

If it works, the equity curve is a smooth line upward. No drawdown, no red day, an excellent return. Every risk statistic computed from the outcome will look superb, because none of them measure the risk that was taken — only the risk that happened to materialise.

That is why "how much of the account was committed to one position" deserves to be tracked as its own number, separately from drawdown and return. A margin requirement is a solvency constraint imposed by somebody else. It is not a position-sizing decision, and it is certainly not a risk assessment. Those are still yours to make.

Figures here are illustrative. Lot sizes, SPAN percentages and exposure margins change, and your broker's requirement will differ from any number above — check it against your own margin statement. The structure does not change: short-option margin scales with the underlying, and the premium you collect is not a measure of what you are risking.