Jump Diffusion Model of Exchange Rate Dynamics – Estimation via Indirect Inference. Working Paper

Jump-diffusion processes have been widely used to model financial time se-ries to reflect discontinuity of asset returns. However, difficulty involved in the estimation of general jump-diffusion processes has prevented their im-plementation in empirical applications. This paper proposes the simulation-based indirect inference approach to the estimation of general parametric continuous-time jump-diffusion processes from discretely observed data. Applications to currency exchange rate models are undertaken to illustrate the estimation procedure and to present interesting empirical results. The es-timation results suggest that jumps are important components of the currency exchange rate dynamics even when conditional heteroscedasticity and mean-reversion are taken into account. However, models assuming conditional ho-moscedasticity tend to overestimate the jump frequency and underestimate the jump size