Introductory Context
"For professional options traders in India, VaR and stress testing are required by: (1) SEBI's ICAAP (Internal Capital Adequacy Assessment Process) for registered portfolio managers. (2) Institutional investor due diligence (investors in PMS and AIF funds require formal risk documentation). (3) Prop firm risk management (internal requirements to ensure desk-level risk limits are respected). Understanding these tools is essential for any options practitioner operating at professional scale. "
Value at Risk - The Three Methods
Historical simulation VaR: compute the portfolio's P&L for each day in the historical data, using actual historical price and volatility changes applied to the current portfolio. The 99th percentile loss of the simulated P&L distribution is the 99% VaR. Advantages: no distribution assumptions, automatically captures historical fat tails and correlation patterns. Disadvantages: backward-looking (past crises may not represent future risks), requires large historical datasets. Parametric VaR (Delta-Normal): approximate the portfolio's P&L distribution as normal using the portfolio's current delta and gamma, and the historical volatility of the underlying. VaR = z × σ_portfolio × portfolio value, where z = 2.33 for 99% confidence. Advantages: fast, analytically simple. Disadvantages: assumes normal distribution (ignores fat tails), inadequate for options portfolios with significant gamma (non-linear payoffs). Monte Carlo VaR: simulate thousands of hypothetical P&L scenarios using a stochastic model (e.g., Heston with jumps) and compute the 99th percentile. Advantages: handles non-linearities (gamma, vol surface sensitivity), can use sophisticated models. Disadvantages: computationally expensive, model-dependent.
Options-Specific VaR Limitations
Standard VaR methods significantly underestimate options portfolio risk because: (1) Fat tails from jumps: options are sensitive to rare large moves that VaR's normal distribution assumption misses. A 99% VaR based on normally distributed returns may miss the 1-in-100 event where Nifty crashes 15% in one day (which the jump-diffusion model correctly prices but the normal model assigns near-zero probability). (2) Volatility regime changes: VaR calibrated to normal volatility periods catastrophically underestimates risk when volatility doubles or triples (COVID 2020: VIX tripled in two weeks). Options' vega exposure means that a sudden VIX doubling can produce losses 2-3x larger than the historical-simulation VaR based on normal-period data would predict. (3) Non-linearity: the delta-gamma approximation in parametric VaR misses the higher-order non-linearities of large options books. Full revaluation (pricing the entire portfolio under each scenario) is required for accurate options VaR.
Stress Testing - Scenarios That VaR Misses
Stress testing complements VaR by evaluating the portfolio under specific named scenarios that VaR's statistical framework cannot capture: (1) Market crash scenario: Nifty -15% in one session, VIX triples from 15 to 45, correlations spike to 0.90+. (2) VIX spike only: VIX rises from 15 to 35 without a large underlying move (pre-event fear spike). (3) Sustained bear market: Nifty declines 30% over 3 months, VIX stays elevated at 25-30. (4) Rate shock: RBI emergency rate hike of 100 bps in one day. (5) Global liquidity crisis: global VIX spikes to 80+, all correlations go to 1.0, options bid-ask spreads widen 10x. Each scenario is applied to the portfolio and the P&L for each scenario is computed using full revaluation (repricing every options position under the scenario's market conditions). The stress test result is the portfolio's P&L under each scenario -- directly answering 'what is our worst case if X happens?'
Professional Risk Management Dashboard
Daily reporting requirements: 1-day 99% VaR (target: <2% of AUM). Current net delta (target: within ±20 Nifty units per Rs 10Cr AUM). Current net vega (target: within ±Rs 5L per VIX point per Rs 10Cr AUM). Stress test P&L for each scenario (flag if any scenario shows >10% AUM loss). Maximum drawdown (YTD and rolling 90-day). Portfolio Sharpe ratio (rolling 90-day). Strategy-level attribution (each strategy's contribution to VaR, Sharpe, drawdown). Limit breach protocol: if any metric exceeds limit → immediate notification to risk manager → mandatory position reduction to within limits within T+1 session.
Professional risk management is not about preventing losses -- it is about ensuring that losses, when they occur, are within the boundaries of what the strategy's capital structure and investors can survive. A well-run options fund can and will have losing months -- what it cannot have is a single losing month large enough to permanently impair the fund's capital. VaR and stress testing are the quantitative tools that maintain this constraint: they ensure that before every new position is entered, the fund's risk team has verified that the worst-case scenario, while potentially painful, does not threaten the fund's viability. This guarantee of survival through the worst case is the foundation of professional institutional confidence in an options strategy.