Capital allocation

Quick answer: Capital allocation is the division of a finite account across strategies, instruments and time horizons. The institutional standard is to allocate by risk contribution rather than by rupees deployed, because margin used and risk taken are not proportional, and to reserve headroom for intraday margin expansion.

In simple words

Capital allocation is how you split your account across different trades, instruments and times, rather than piling everything into one. Two ideas make it harder than it sounds. First, the margin a trade uses is not the same as the risk it carries — a small margin can hide a large possible loss, so dividing by margin misleads you. Second, margin is not fixed: on short options it can jump during the day if the market moves, and a position that looked fully funded at entry can be force-closed. So sensible allocation keeps a cash buffer unused, splits by what each trade can actually lose rather than by rupees, and checks that different-looking trades are not secretly the same bet.

Allocate by risk, not by rupees

The standard institutional practice is to allocate capital by the risk each position contributes, not by the rupees it ties up. The reason is that equal rupee allocations can carry wildly unequal risk: ten lakh in a defined-risk spread and ten lakh of margin behind naked short options are not the same bet, even though the capital figure matches. Allocating by risk means expressing every position in a common measure — worst-case loss at a defined move, or a value-at-risk figure — and setting the weights on that measure. This is the layer that sits above position sizing: sizing decides how many lots a single trade takes, allocation decides how the account's total risk is shared across many trades and strategies. A book that looks balanced by capital can be dangerously concentrated by risk, and only the risk view reveals it.

Margin used is not risk taken

The most common allocation error is to treat margin as a proxy for risk. They are not proportional. A defined-risk option spread can tie up modest margin while its maximum loss is strictly capped; a short strangle can tie up more margin while its worst case runs far beyond that margin. Two positions using identical margin can have completely different tails. Margin is the exchange's collateral demand under its stress assumptions; risk is what the position actually loses in the scenario that matters to you, and the two only coincide by accident. Allocating a fixed share of margin to each strategy therefore says nothing coherent about how the account's risk is distributed. The correct denominator is the position's own worst-case loss, and for undefined-risk positions that worst case is not the margin, it is a scenario the trader must impose.

Reserve margin headroom

Margin on short options is not a one-time charge at entry; it is recomputed intraday as SPAN plus exposure, and it expands when the market moves against the position or when volatility rises. A position that is fully margined at entry can face a margin call, and a shortfall the broker cannot ignore, after an adverse move — and brokers may square off positions to cure the shortfall, often at the worst possible moment. This is why an account that deploys close to its full margin capacity is fragile even if every individual position is sized correctly: it has no buffer to meet the expansion. Reserving headroom — deliberately leaving a slice of capital unused as free margin — is not idle cash, it is the reserve that keeps the account from being liquidated into a move. Allocation that ignores this reserve is allocation that has not accounted for the way margin actually behaves.

The correlation trap

Allocations that appear diversified can be a single concentrated bet wearing several names. Spreading capital across NIFTY, BANKNIFTY and FINNIFTY short-premium positions looks like three allocations, but they are three expressions of one view — Indian equity index volatility — and in a stress move they lose together. The danger is that the allocation framework counts them as independent and understates the combined risk, so the account carries far more of one factor than the capital split suggests. Genuine allocation requires looking through the instruments to the underlying drivers and asking how many truly distinct bets the account holds, not how many tickers. Correlations that look moderate in calm markets rise toward one in a crash, which means the correlation trap is worst precisely when the diversification was supposed to help. Counting positions is not counting risks.

Allocating across time

Capital is allocated not only across strategies and instruments but across time. Deploying the entire account at once concentrates exposure on the single entry point's conditions; staging entries, staggering expiries and keeping dry powder spreads the account across different market states and different points on the volatility surface. Time allocation also governs reserves: capital held back today is capital available to meet a margin expansion, to act when conditions change, or to survive a drawdown without forced selling. There is a genuine trade-off — capital held in reserve is capital not earning the strategy's edge — and the right split is a choice, not a formula. But an account fully deployed at all times has implicitly chosen zero reserve, and that choice is usually made by inattention rather than intention, which is the point at which allocation across time stops being managed at all.

Common mistakes

• Allocating capital by rupees deployed instead of by risk contributed lets equal-looking allocations hide unequal tails, so an account that appears balanced can be dominated by one undefined-risk strategy.
• Using margin as a proxy for risk misleads, because margin and worst-case loss are not proportional, and two positions with identical margin can have completely different maximum losses.
• Deploying close to full margin capacity leaves no headroom for the intraday expansion of SPAN plus exposure, so an adverse move triggers a margin call and the broker may square off positions at the worst moment.
• Spreading capital across NIFTY, BANKNIFTY and FINNIFTY short premium and calling it diversification counts one volatility bet three times, so the account carries far more of that single factor than the split suggests.
• Treating a fully margined position as safe until expiry ignores that margin is recomputed intraday, so a position funded at entry can still be liquidated after the market moves against it.
• Keeping the account fully deployed at all times implicitly chooses zero reserve, usually by inattention, leaving nothing to meet margin expansion or to act when conditions change.

Professional usage

Institutions run capital allocation as a formal risk-budgeting exercise. A total risk appetite is set at the book level, then divided across strategies and desks by their risk contribution — measured through value-at-risk, stress losses and scenario analysis — rather than by the capital they consume. Cross-margining and portfolio-margin systems let institutions net offsetting risks, so their margin efficiency and their risk view are computed together rather than confused. Reserves and liquidity buffers are mandated, not optional, precisely so that a margin expansion or a redemption never forces a fire sale. Correlation and factor models look through instruments to shared drivers, so three index positions are recognised as one volatility exposure. Retail traders lack portfolio margin, factor systems and mandated reserves, but the transferable disciplines — allocate by worst-case loss, keep free-margin headroom, and count distinct bets rather than tickers — require only arithmetic and attention.

Frequently asked questions

Voice search & related questions

Last reviewed 9 July 2026. Educational content only — not investment advice.