According to many recent studies, Lévy processes with stochastic volatility seem to be the best candidates for replacing geometric Brownian motion (GBM) as a price process model. This means that the GBM model has to be generalised by introducing the possibility of jumps and allowing the volatility to be a stochastic process. In this paper, we present a generalisation of the traditional Lévy–Merton jump-diffusion model, allowing discrete stochastic volatility. In order to estimate jump instants and jump amplitudes, we use, and improve on, a method based on quadratic variation. We apply this method to two time series provided by the “Banco de España” comprising daily observations of interest rate for operations of 1 day and 1 year (from 4 January 1988 to 31 December 1998).