We address several approaches to modelling high frequency financial data in continuous time.
At first we suggest a method of moment estimator for the parameters of the COGARCH(1,1) process, analyse the asymptotic properties of the estimator and fit the model to simulated and real high-frequency data. In the following chapter we develop the first new model, an exponential COGARCH(p,q) process. We investigate stationarity and moment properties and show that the model describes an instantaneous leverage effect. For the compound Poisson ECOGARCH(1,1) process we propose a quasi-maximum likelihood type estimator. To account for the strong persistence in volatility, which is sometimes observed in empirical data, we develop a fractionally integrated ECOGARCH(p,d,q) process. Finally we propose a new mixed effect model class for the absolute log returns. Explanatory variable information is used to model the fixed effects, whereas the error is decomposed in a non-negative CARMA process and a market microstructure noise component. In the end the model is applied to real high frequency financial data.
«
We address several approaches to modelling high frequency financial data in continuous time.
At first we suggest a method of moment estimator for the parameters of the COGARCH(1,1) process, analyse the asymptotic properties of the estimator and fit the model to simulated and real high-frequency data. In the following chapter we develop the first new model, an exponential COGARCH(p,q) process. We investigate stationarity and moment properties and show that the model describes an instantaneous l...
»