Introductory Context
"In Indian markets, cross-asset correlation is relevant for several specific applications: Nifty options (which depend implicitly on the correlations between Nifty 50 components), FII flow impact analysis (where multiple asset class correlations -- equity, rupee, gold, crude -- determine the portfolio-level impact), dispersion trading (Topic 25.8), and any structured product linked to the performance of multiple assets. Understanding how correlation affects multi-asset options pricing and risk is the final analytical dimension that completes the advanced options practitioner's toolkit. "
How Correlation Affects Multi-Asset Options
The fundamental relationship: for a basket of n assets, the basket's variance = Σᵢ wᵢ² σᵢ² + 2Σᵢ<j wᵢ wⱼ σᵢ σⱼ ρᵢⱼ. Where wᵢ are weights, σᵢ are individual volatilities, and ρᵢⱼ are pairwise correlations. High correlation (ρ → 1): the basket's variance approaches the weighted average of individual variances -- limited diversification, basket behaves like a single asset. Zero correlation (ρ = 0): the basket's variance is the weighted sum of individual variances -- maximum diversification benefit, basket is much less volatile than individual components. Negative correlation (ρ → -1): the basket's variance is minimised -- a perfectly hedged portfolio.
Options pricing implications: a basket call option is more expensive when component correlations are high (the basket is more volatile) and cheaper when correlations are low (the basket is more stable). For the Nifty 50 index option: the index's implied volatility (India VIX) reflects the average pairwise correlation between Nifty 50 components. When FII selling drives all sectors down simultaneously (high correlation episode), India VIX spikes. When sector rotation occurs (IT rising while banking falls), the pairwise correlation falls and Nifty's movement is muted relative to individual stock movements -- the basis of dispersion trading (Topic 18.13, Topic 25.8).
The Correlation Risk Premium
Similar to the volatility risk premium (VRP) -- implied volatility systematically exceeds realised volatility -- there exists a correlation risk premium (CRP): the implied correlation (extracted from the relative pricing of index options vs single-stock options) systematically exceeds the realised correlation over the same period. This CRP exists because investors pay a premium for protection against correlation spikes during crises: when all stocks fall together (high correlation), the diversification benefit of holding a basket disappears and investors suffer large portfolio losses. Options on baskets (specifically Nifty index options) implicitly price this correlation risk premium, making index options systematically expensive relative to single-stock options -- creating the basis for dispersion trading.
Correlation Modelling Challenges
Correlation is harder to model than volatility for three reasons: (1) High dimensionality: the Nifty 50 has 50 × 49 / 2 = 1,225 pairwise correlations. Estimating and modelling 1,225 correlations that are mutually consistent (the correlation matrix must be positive semi-definite) is a significant statistical challenge. (2) Instability: correlations are less stable than volatilities. In normal markets: Nifty 50 stock pairwise correlations average approximately 0.25-0.35. During the 2020 COVID crash: average correlations spiked to 0.70-0.85. This factor-2 to factor-3 change in correlation during crises is much larger than the typical change in individual volatilities. (3) Observable only indirectly: while individual stock volatility can be observed directly from that stock's option prices, pairwise correlation is observable only indirectly from relative pricing of basket vs individual options (through the implied correlation extraction method). The model must make assumptions about the correlation structure that may not match reality.
Implied Correlation vs Realised Correlation (Nifty Context)
Implied correlation (from VIX vs average single-stock vol): approximately 0.40-0.55 in normal markets. Realised correlation: approximately 0.25-0.35 in normal markets. Correlation premium: implied exceeds realised by approximately 0.10-0.20 units. During crises: realised correlation spikes to 0.70-0.85, exceeding or matching the previously elevated implied correlation. Dispersion trading exploits this premium: short Nifty vol (via index options) + long individual stock vol (via single-stock options) = profit from implied correlation falling toward (or staying near) realised correlation.
Best-of/Worst-of Options and Correlation Sensitivity
The best-of option on n assets pays the maximum return of all n assets. The worst-of option pays the minimum return. These multi-asset options have strong correlation sensitivity: best-of options are most valuable when correlations are low (the best asset can diverge significantly from the worst, making the maximum return large). As correlation rises, the best-of's value decreases (all assets move together, so the maximum is not much better than any individual). Worst-of options behave oppositely: their value increases with correlation (when all assets move together, the worst performer is dragged down by the group's average, increasing the worst-of's expected loss). Best-of and worst-of options are common in equity-linked structured products -- the investor participates in the best-performing of several benchmark funds, or the redemption is linked to the worst-performing component. Understanding their correlation sensitivity is essential for evaluating these products.
Correlation is the financial market's most treacherous variable: benign in normal conditions, catastrophic in crises. The market's typical 0.30 pairwise correlation means diversification is working, portfolios are stable, and multi-asset options are reasonably priced. The crisis 0.80 pairwise correlation means diversification has failed, portfolios are moving in lockstep, and the correlation risk premium embedded in index options is being fully realised. The practitioner who understands correlation -- its normal level, its crisis level, its relationship to the dispersion trade, and its impact on basket options -- understands the final frontier of options market risk that the single-asset frameworks from earlier modules cannot address.