Processes of Meixner type (Q1567713)

A Meixner(-type) process is a Lévy process with characteristic function (i.e. Fourier transform) \(\xi\mapsto ({\cos(\beta/2)\over \text{cosh}((\alpha\xi- i\beta)/2)})^{2\delta t} e^{i\mu t\xi}\), where \(\alpha,\delta> 0\), \(\beta\in (-\pi,\pi)\), \(\mu\in \mathbb{R}\) are parameters. Note that a Meixner process is not a (generalized) hyperbolic process. The author studies various aspects of Meixner processes such as their canonical form, self-decomposability, and Esscher transform. The latter yields a martingale measure for Meixner-type stock price processes and corresponding Black-Scholes formula. Mixed Meixner processes are introduced (mixing means integration over the parametrized family of Esscher-transformed processes) and their semimartingale nature is studied. Following \textit{O. E. Barndorff-Nielsen} [Finance Stoch. 2, No. 1, 41-68 (1998; Zbl 0894.90011)], the author constructs a driving Lévy process \(Z^{(\lambda)}_t\) such that the solution of the S.D.E. \(dX_t=\lambda X_t dt+ dZ^{(\lambda)}_t\) is stationary with prescribed one-dimensional Meixner marginals. The proofs are essentially those of the paper quoted above. Finally, self-similar Meixner processes are constructed and their Lévy characteristics are explicitly given.