Modelling Dividend Gaps with Jump Stochastic Processes

The aim of this paper is to propose a number of numerical procedures for calibration and path gen- eration of the heavy-tailed stochastic processes, in particular, to model the dynamics of the equity markets. I developed an APT-style two-factor model to separate market-wide effects from the id- iosyncratic shocks of the stocks under consideration. For the modelling of market-wide effects Vasicek and Heston models were used, while idiosyncratic returns were studied under the framework of the Merton Jump-diffusion process. All models were calibrated on historic data via Max-Likelihood esti- mation. I compared the accuracy of model calibration via Max-Likelihood and Method of Moments to the conclusion that the use of latter has led to a minor decrease in back-testing Log-Likelihood, while being highly computationally efficient.