SABR and the Difference Between Static and Dynamic Volatility Cubes

Summary

The discussion distinguishes a linearly interpolated swaption volatility cube from a cube generated using SABR. Linear interpolation of market quotes can describe a static set of Black-model volatilities across strikes and maturities. SABR instead models volatility as stochastic and related to the underlying swap rate, so its implied volatility structure can change as rates move.

That dynamic behavior matters for valuation and hedging: the SABR framework produces different option sensitivities, or Greeks, from treating the quoted volatilities as fixed. The response therefore frames these approaches as answering different questions, rather than as competing interpolation methods. The exchange does not give calibration details, compare interpolation errors, or establish that SABR guarantees arbitrage-free surfaces; the stated distinction is about model dynamics and the resulting sensitivities.

Key ideas

• Linear interpolation of quoted Black volatilities creates a static volatility representation.
• SABR models volatility as stochastic and related to the underlying swap rate.
• A dynamic volatility model can produce different option Greeks from a static cube.
• The discussion distinguishes the models by their behavior as rates move, without comparing calibration quality.

Tags

# what is the point of SABR model as an interpolation tool if we can already observe the whole vol cube from the market


# what is the point of SABR model as an interpolation tool if we can already observe the whole vol cube from the market












on BBG and other data providers, it is common that you can find the whole vol surface/cubes. What is the point of the SABR model as an interpolation tool? why cannot people just linear interpolate the vol surface/cubes? instead of backing up the SABR params and interpolate on the param space? is it for the sake of non-arbitrage?

## Answer by user35980 (score 1)

https://quant.stackexchange.com/a/77145

SABR is a stochastic vol model with the vol parameter being a (stochastic) function of the underlying swap rate. Yes, you can use market skew data and build a linearly interpolated swaption vol cube. But this cube entails constant vols based on the standard Black model, so it's a static structure. The SABR swaption vol cube is a dynamic structure that evolves as rates move - that's kind of the point of the model. This also entails significantly modified greeks... etc.

So you're comparing apples and oranges.

Shown in full with attribution under the source's licence. Licence: CC BY-SA 4.0 (Stack Exchange)

This summary was written by Stratmill's research agent from the original; it is not a copy of the source.