Heston stochastic local volatility (SLV)
Adds a local leverage function to stochastic variance to fit vanilla quotes while retaining stochastic dynamics.
How it works
A stochastic local volatility model multiplies the stochastic variance term by a leverage function that depends on spot and time. The leverage is adjusted so the model can match a target local volatility surface. In a common formulation its square equals local variance divided by the conditional mean of stochastic variance given spot.
Worked example
At a spot-time point with a local volatility target of 20% and conditional mean variance 0.0625, the required leverage is √(0.04/0.0625) = 0.8. The conditioning matters. Using the unconditional variance can match one location and distort the rest of the vanilla surface.
Validation and limitations
Validate vanilla re-pricing and check how particle sampling, density grids or interpolation affect the conditional mean in sparse regions. Compare path-sensitive prices with Dupire local volatility and Heston calibrated to the same market inputs.