Margin Requirements & Capital Rules

In the cash segment you pay for what you buy. Buy 100 shares of SBIN at ₹820 and ₹82,000 leaves your account; the shares arrive and the transaction is finished. A futures or short-option position is different. Nothing is bought outright. You have entered a contract that obliges you to settle a difference on some future date, and that difference has not happened yet. The exchange has no asset in hand and no way to know which way the obligation will run. So it holds a deposit instead.

The party the deposit protects is not primarily you. Between every buyer and seller on NSE and BSE sits a clearing corporation that legally becomes the counterparty to both. If a short seller cannot pay on Monday, the clearing corporation still owes the buyer. Its guarantee is only as good as the collateral it holds. Margin is that collateral, sized so that a single day of adverse movement across the whole market can be covered without anyone chasing a defaulter. That is why the number moves when volatility moves — the risk it is sizing has moved.

For the trader, the practical consequence is that margin is a block, not a cost. If ₹2,00,000 is blocked against a short strangle, that ₹2,00,000 is still yours; it is simply unavailable for anything else until you exit. It appears on your broker screen as a reduction in available margin, never as a debit on your ledger. What genuinely costs you money is the mark-to-market settlement of the position, the brokerage and statutory charges on the contract note, and the opportunity cost of capital that sat frozen.

SPAN stands for Standard Portfolio Analysis of Risk. It is a risk model, not a percentage. Rather than applying a fixed rate to each contract, it takes your entire portfolio in a segment and asks a single question: across a defined grid of adverse scenarios, what is the largest one-day loss this book could show? The grid moves the price of the underlying up and down in steps, and independently moves implied volatility up and down. Each combination produces a profit or loss for your whole portfolio. The worst of them becomes the SPAN requirement.

Two consequences follow, and both matter. First, SPAN is portfolio-level. A long call and a short call at a higher strike are not scored separately — the grid values them together, sees the long leg absorbing the damage in the scenario that hurts the short leg, and reports a far smaller worst case. Second, SPAN is volatility-sensitive. The same short position that blocked a comfortable amount in a quiet market blocks materially more when India VIX jumps, because the volatility-up scenarios in the grid now generate bigger losses. Nothing about your position changed. The measured risk did.

The grid parameters, the scenario count and the price-scan and volatility-scan ranges are published by the clearing corporation and revised periodically. Do not reason from a remembered percentage. The exchanges publish a margin calculator, and every broker exposes a basket-margin screen where you can enter the exact legs you intend to trade and read the requirement before placing a single order. Doing that before the trade, rather than discovering the block afterwards, is the entire discipline here — and it connects directly to the position sizing module, where lot count is derived from capital rather than guessed.

SPAN covers the risk the model can see. Exposure margin covers the part it cannot. It is a second layer, computed not as a scenario loss but as a flat proportion of the contract value, and it is added on top of the SPAN figure rather than replacing any part of it. Its purpose is a buffer against a move larger than the scan range in the grid — the gap opening, the circuit, the overnight event that lands outside the scenarios the model priced. Where SPAN is analytical, exposure margin is blunt on purpose.

Because it is computed on contract value, exposure margin scales directly with the size of the contract. That is the quiet link back to the SEBI 2025 framework covered in the F&O rules module: raising the minimum contract value for index derivatives raised the exposure component roughly in proportion. The percentage itself differs between index contracts and single-stock contracts, single-stock rates being higher because a single company can gap on news in a way an index of fifty companies structurally cannot.

The applicable percentages are set by the exchange and have been revised more than once. This page deliberately does not quote them. Naming a rate that has since changed would be worse than naming none, because a reader would size a position on it. Take the current exposure margin rate for the exact instrument from the NSE or BSE contract specification and margin circulars, or simply read the combined figure off the margin calculator, which already stacks SPAN and exposure for you.

The number your broker freezes is a stack, not a single figure. SPAN forms the base. Exposure margin is added on top. Then come additional margins that apply only in specific circumstances: an extra requirement on short option positions on the session they expire, a delivery margin that ramps up through the final sessions of a physically settled stock contract, and scrip-specific surveillance margins on names under exchange watch. Each layer is applied independently, and a position can carry all of them at once.

The sequence matters because the layers do not arrive together. SPAN and exposure appear the moment you enter. The expiry-session add-on appears on the morning of expiry, against a position you may have held unchanged for a fortnight. Delivery margin on a stock contract appears in the final week and climbs each session toward the full contract value, which is the mechanism explained in the settlement module. An account funded to exactly the entry-day requirement is an account that gets a margin call it did nothing to cause.

One further layer is worth naming because traders meet it without recognising it. Intraday, the clearing corporation samples open positions at randomly timed snapshots during the session, and the margin your broker must have collected is measured against the peak of those snapshots — not against your position at the close. Squaring off before 3:30 pm does not undo a shortfall recorded at 11:40 am. Reporting rules and the applicable percentage of that peak are regulatory and have been phased in over successive circulars; confirm the operative position with your broker.

An option buyer posts no margin at all. Their maximum loss is the premium they paid, so there is nothing further for the clearing corporation to collateralise. What changed under the SEBI framework rolled out through 2025 is when that premium has to be in the account. It must now be collected in full, upfront, before the order executes. Previously a broker could effectively fund an intraday long-option position, letting an account carry more premium than it held provided everything was squared off the same session.

Put the arithmetic on it. A Nifty call quotes ₹100. The lot size is 75, so one lot costs 100 × 75 = ₹7,500, and that ₹7,500 must be free in the account before the order goes through. It leaves your available margin immediately and does not come back until you exit. Reading the ₹100 on the screen as the size of the trade is the most common arithmetic mistake a new options trader makes, and the larger post-2025 lot sizes have made it an expensive one.

For a seller the premium works the other way. Premium received is credited and, in the exchange computation, reduces the net requirement against the position. This is why a short position often shows a block smaller than the raw SPAN-plus-exposure sum. It is also why the block moves against you as the position goes wrong: the option you sold is now worth more, the mark-to-market runs against you, and the requirement recomputes upward on the same position. The buyer-versus-seller asymmetry module works through why the two sides face such different capital demands.

You do not have to meet the whole requirement in cash. Approved securities already sitting in your demat — index ETFs, liquid funds, government securities, a defined list of shares — can be pledged to your broker and the value credited to you as collateral margin. The economic appeal is obvious: a long-term holding keeps earning its dividends and its price appreciation while simultaneously supporting an F&O position. The holding never leaves your demat account; a lien is marked on it.

Two frictions apply. The first is the haircut. Collateral is never credited at full market value, because the collateral itself can fall. The exchange assigns a haircut percentage that varies by instrument, deeper for volatile equity and shallower for liquid funds and government securities. Pledge ₹10,00,000 of a share carrying a haircut and you receive materially less than ₹10,00,000 of usable margin. The applicable haircut differs by scrip and is revised, so it must be read off the approved-securities list rather than assumed.

The second friction is the cash component. Regulation requires that a defined minimum proportion of the total margin be met with cash or cash-equivalents rather than pledged shares; the balance may be non-cash collateral. Fall short of that proportion and your broker levies a delayed-payment or interest charge on the gap, at a rate set out in its tariff. Both the required proportion and the penalty rate are live figures — confirm the current requirement with your broker and the exchange circular rather than working from a rule of thumb.

Because SPAN scores the portfolio, not the contract. Sell a naked Nifty call and, in the grid scenario where the index rises hardest and volatility expands, the position shows a loss with nothing offsetting it. That scenario becomes the SPAN requirement. Now buy a call at a higher strike in the same expiry. In that same violent up-scenario the long call is deep in the money and gaining, and it cancels most of what the short leg lost. The worst case across the grid collapses, so the requirement collapses with it.

This is not a discount the exchange grants out of goodwill. It is an honest measurement: the hedged position genuinely cannot lose more than the distance between the strikes, less the net credit received. The margin system is simply pricing a smaller risk. The corollary is that the benefit is only as real as the hedge. Close the long leg first and you are standing naked with a requirement that recomputes instantly upward. Carry a spread into the session where the short leg expires and, per the 2025 framework, the expiring leg can be margined as though it stood alone.

How much is saved is not something this page will quote, because it depends on strike width, expiry, the instrument and the prevailing volatility, and any single percentage would be wrong for most readers. The way to find out is to price both structures in the basket-margin screen — the naked short on its own, then the same short with the protective leg added — and read the two totals. The directional spreads and iron condor modules build the structures themselves; this is why a defined-risk trader can run a larger book on the same capital.

A shortfall arises in one of three ways, and only one of them involves you doing anything. You take a mark-to-market loss and your free balance falls below the requirement. Or volatility expands and the requirement itself rises against an unchanged position. Or a scheduled add-on lands — the expiry-session margin on a short option, the delivery margin in the final week of a stock contract. The second and third are the ones that surprise people, because the trigger was not a trade.

What follows is a margin call: a demand to fund the gap, usually with a stated deadline the same or the next session. If it is not met, the broker has the contractual right to square off enough of your position to close the shortfall, and it will exercise that right. Forced square-offs are executed at market, frequently in exactly the illiquid or fast-moving conditions that caused the shortfall in the first place. The exit price is whatever the book offers. This is also why unfunded shortfalls attract a broker penalty in addition to the loss.

The defence is structural rather than clever. Do not deploy capital to the full requirement. Size positions so that a meaningful multiple of the entry-day block remains free and unencumbered, in cash, specifically so that a volatility expansion or a scheduled add-on can be absorbed without a decision being forced on you. The risk-per-trade and position sizing modules give the arithmetic for that buffer. A structurally sound trade closed at the worst tick of the week because the account was fully deployed is not a market loss. It is a funding failure.