We propose a finite-difference pricing method for FX Target Redemption Forwards (TARFs) within a local stochastic volatility (LSV) framework. The approach incorporates a multi-state regime-switching model for stochastic volatility, enabling fast and robust calibration. The local volatility surface is obtained in a single forward-sweep on the finite-difference grid. TARF valuation is achieved by solving coupled one-dimensional PDEs. Valuation results are compared with those from a Heston-LSV model. Numerical experiments demonstrate the stability and effectiveness of hedging TARFs with the proposed model.

Ferdinand Reuther is a manager at d-fine, where he focuses on developing, implementing, and validating valuation models. He has more than six years of experience working on projects in the valuation of financial instruments across multiple asset classes. He holds an M.Sc. in Quantitative Finance.