Nonstationary Generalised Autoregressive Conditional Heteroskedasticity Modelling for Fitting Higher Order Moments of Financial Series within Moving Time Windows

Figure 3

The stability phase diagram for the GARCH-normal (1,1) moments with an overlap of several company trajectories. (a) The trajectory for Lloyds Bank, where the shortest time window allows up to the tenth-order moment to exist. (b) The trajectory of Gold ETFs (Exchange Traded Funds), where the shortest time window allows to the sixth moment to exist. (c) The same for GSK, which allows up to the sixth moment to exist and lastly. (d) The same for Rio Tinto, which allows up to the twelfth-order moment to exist. We also show the error intervals of the trajectories, see Appendix A for details.

(a)

(b)

(c)

(d)