Local Stochastic Rough Volatility: Pathwise Filtering and the Conditional Density Equation

Battle Count: 96

Highly relevant to quantitative trading and derivatives desk operations. The paper provides a rigorous mathematical foundation for calibrating local stochastic rough volatility models (LSRV) to implied-volatility surfaces, which is central to exotic option pricing and hedging. The Rao-Blackwellized calibration ratio (equations 30-32) directly informs leverage-function updates used in production LSV calibration. The rough Heston explicit lognormal solution provides a benchmark for numerical implementations. The pathwise filtering framework enables conditional density estimation critical for risk management under rough volatility. However, the paper is theoretical and does not provide production-ready code or empirical backtests.

Very high implementation complexity. Requires: (1) joint simulation of correlated rough volatility paths (Z,V) using specialized Volterra discretization schemes (Euler-type, Markovian approximations, integrated-variance schemes); (2) evolution of the stochastic flow (Φ,J,H) along simulated paths; (3) solving deterministic Fokker-Planck PDEs with rough, path-dependent coefficients using finite differences or Galerkin methods; (4) coordinate inversion via inverse flow; (5) Rao-Blackwellized calibration as a fixed-point/Picard iteration. The reduced temporal regularity (H<1/2) necessitates very fine temporal grids and implicit/semi-implicit schemes. The rough Heston case is simpler due to the explicit geometric flow solution.

The paper is purely theoretical with complete mathematical proofs and explicit formulas. All assumptions are clearly stated (Assumptions 1-4). The rough Heston specialization provides explicit closed-form solutions (equations 33-39) that can be independently verified. However, no numerical experiments or code are provided. The theoretical framework is self-contained and reproducible for verification purposes, but practical implementation requires significant additional work (joint simulation schemes, PDE solvers).