Variance swaps on time-changed Lévy processes (Q1761447)

The paper studies variance swaps when the underlying is modelled as a time-changed Lévy process, where the time change does not need to be independent of the Lévy process. Such a model allows for volatility clusters, stochastic volatility, leverage effects and stochastic jump intensities. While classical theory based on the continuity assumption of the underlying implies that the value of a variance swap is given by two times the value of a log contract on the underlying, in the more general modelling framework proposed in this paper, the authors find that the value of the variance swap is also given by a multiple of the log contract price, but this multiple does not need to be equal to two. The authors derive the formula for this multiple and show that it exceeds two in the practically relevant cases.