← Morgan Stanley Interview Insights
This is the kind of question where you know the answer exists and you've seen it before but reconstructing it under pressure is a different thing entirely.
Start by defining the setup: you hedge with implied vol σ_i but the stock realizes σ_r. Then decompose the PnL into the gamma-theta trade-off, showing that the daily PnL is approximately 0.5 * Γ * S^2 * (σ_r^2 - σ_i^2) * dt. Finally, integrate over the option's life to get the total PnL, which is path-dependent and proportional to the variance spread.
Pro tip: Emphasize that the PnL is not simply a function of the difference in vols but depends on the realized path of the stock price and the gamma profile. Mention that if realized vol exceeds implied vol, the hedger profits, and vice versa, but the exact PnL can vary due to path dependency.
Define that you sell an option and hedge it using Black-Scholes with implied volatility σ_i, but the stock actually realizes volatility σ_r. Assume continuous hedging for simplicity.
Using Itô's lemma, show that the change in the hedged portfolio value over a small time step dt is approximately 0.5 * Γ * S^2 * (σ_r^2 - σ_i^2) * dt, where Γ is the option's gamma.
Sum the daily PnL from t=0 to T to get the total PnL: ∫ 0.5 * Γ_t * S_t^2 * (σ_r^2 - σ_i^2) dt. This shows the PnL is the integral of the gamma-weighted variance spread.
Explain that if σ_r > σ_i, the hedger profits on average, and if σ_r < σ_i, the hedger loses. The magnitude depends on the path of S and the gamma profile.
Mention that in reality, hedging is discrete, leading to additional hedging error, and that the PnL can be decomposed into a variance swap-like payoff plus a path-dependent residual.
AI-generated suggestions, not part of the candidate's original notes. May be inaccurate — verify before relying on them.
Easier to articulate once you've done the derivation above.
Start by explaining that delta-hedging isolates exposure to realized variance versus implied variance. Then describe how gamma determines the profit or loss from rebalancing, and conclude with the relationship between the P&L and the spread between realized and implied variance.
Pro tip: Emphasize that the sign of the P&L depends on whether realized variance exceeds implied variance, and that gamma is the key driver. Mention that in practice, transaction costs and discrete hedging can erode profits, so the theoretical result assumes continuous hedging and no costs.
Explain that delta-hedging involves continuously adjusting the underlying position to keep the portfolio delta-neutral, aiming to eliminate directional risk.
Describe gamma as the rate of change of delta, which measures the curvature of the option price. Highlight that gamma determines the frequency and size of rebalancing trades.
State that the profit or loss from delta-hedging is approximately half the sum of gamma times the squared change in the underlying price, i.e., 0.5 * Γ * (ΔS)^2.
Connect the P&L to the difference between realized variance (from actual price moves) and implied variance (from option prices). If realized variance is higher than implied, a long gamma position profits; if lower, it loses.
Conclude that delta-hedging makes money when realized variance exceeds implied variance (for long gamma) and loses when realized variance is lower, assuming no transaction costs.
AI-generated suggestions, not part of the candidate's original notes. May be inaccurate — verify before relying on them.
I'd seen this formula before but never had to derive it from scratch on the spot.
Start by clarifying the assumptions (continuous hedging, no transaction costs, Black-Scholes framework) and then derive the PnL decomposition. Explain that the total PnL from delta-hedging a short option position equals the integral of (1/2) * S^2 * Gamma * (σ_realized^2 - σ_implied^2) dt, and then sketch how this approximates the given expression when the difference in squared vols is small. Use a simple diagram or equation to illustrate the relationship.
Pro tip: Emphasize that this is a fundamental result in quantitative finance: it shows that a delta-hedged option position is essentially a bet on realized volatility versus implied volatility. Mention that in practice, discrete hedging and transaction costs cause the actual PnL to deviate, but the integral approximation holds well under frequent rebalancing.
Assume a Black-Scholes world with continuous hedging, no transaction costs, and a short option position. Define the delta-hedged portfolio and the PnL over a small time interval.
Using Itô's lemma, show that the change in portfolio value over dt is approximately (1/2) * S^2 * Gamma * (σ_realized^2 - σ_implied^2) dt, where σ_realized is the actual volatility and σ_implied is the volatility used for pricing.
Sum the instantaneous PnL over the life of the option to get the total PnL as the integral of (1/2) * S^2 * Gamma * (σ_realized^2 - σ_implied^2) dt.
Note that the difference in squared vols is (σ_realized^2 - σ_implied^2). If we denote this as Δσ^2, the integral becomes ∫ (1/2) S^2 Γ Δσ^2 dt, which matches the question's description.
Draw a graph with time on the x-axis and the integrand on the y-axis, or simply write the integral expression and explain its meaning: the PnL is the weighted sum of the variance spread, with weights proportional to S^2 and Gamma.
AI-generated suggestions, not part of the candidate's original notes. May be inaccurate — verify before relying on them.
This tripped me up more than it should have.
Start by affirming that yes, a delta-hedged position can lose money even if realized volatility exceeds implied volatility on average, because the P&L depends on the path of volatility, not just its average. Explain that the hedging error is driven by the difference between realized variance and implied variance, weighted by gamma, and that path dependence (e.g., large moves when gamma is low, or small moves when gamma is high) can lead to losses. Conclude by emphasizing that the average realized vol is not sufficient; the timing and magnitude of moves matter.
Pro tip: Mention that the loss can be quantified by the integral of gamma times the difference between realized variance and implied variance, and that discrete hedging and transaction costs further exacerbate the issue. This shows you understand both the theoretical and practical aspects.
Explain that the profit or loss from delta hedging is approximately the integral of (0.5 * gamma * S^2 * (realized variance - implied variance)) over the hedging period. This sets the foundation for why average realized vol alone is insufficient.
Discuss how the hedging error depends on the path of the underlying and the gamma profile. For example, if the stock makes large moves when gamma is low (e.g., when the option is deep in or out of the money), the gains from realized vol may be small, while losses from implied vol can dominate.
Illustrate with a scenario: Suppose you are short an at-the-money option. If the stock stays near the strike (high gamma) with low realized vol, you lose due to time decay. Then, if it suddenly jumps away (low gamma) with high realized vol, the gains from that jump are insufficient to offset the earlier losses. Thus, even if average realized vol > implied vol, you lose.
Mention that in practice, hedging is discrete, and transaction costs and slippage can cause additional losses. Also, the presence of jumps or stochastic volatility can make the hedging error path-dependent.
Summarize that the profitability of a delta-hedged position is not determined by average realized volatility alone, but by the covariance between gamma and realized variance. If the stock moves when gamma is low, the position can lose money even if realized vol exceeds implied vol on average.
AI-generated suggestions, not part of the candidate's original notes. May be inaccurate — verify before relying on them.
More open-ended and I actually preferred it.
Start by clarifying that the choice of hedging volatility depends on the specific product and risk profile, then discuss how the analysis informs practical decisions such as selecting the appropriate vol surface, managing model risk, and optimizing hedging costs. Emphasize the trade-offs between accuracy and computational efficiency, and how these impact real-time trading systems.
Pro tip: Demonstrate awareness that in practice, traders often hedge with a blend of implied vols from different sources (e.g., listed options, OTC quotes) and that the analysis should guide the weighting scheme rather than prescribe a single vol. Also, mention that regulatory constraints and capital requirements can influence the choice.
Clarify whether the goal is to minimize variance, match a specific benchmark, or reduce tail risk. This determines which vol (e.g., ATM, skew-adjusted, or realized) is most appropriate.
Compare implied vols from different sources (exchange-traded, broker quotes, model-implied) and realized vols. Consider liquidity, bid-ask spreads, and data quality.
Discuss how the choice of vol interacts with the pricing model (e.g., Black-Scholes vs. stochastic vol). Highlight the sensitivity of hedges to vol misspecification and the need for robust calibration.
Factor in transaction costs, margin requirements, and system latency. The analysis should inform whether a more complex vol surface is worth the computational overhead.
Propose a dynamic hedging strategy that adapts to changing market conditions, with regular backtesting and performance attribution to validate the vol choice.
AI-generated suggestions, not part of the candidate's original notes. May be inaccurate — verify before relying on them.