Stochastic volatility models with possible extremal clustering (Q2435218)

The paper focuses on a stochastic volatility model given by \[ Y_t =\sigma_t,\quad t\in \mathbb{Z}, \] in which the volatility sequence \((\sigma_t)\) and the i.i.d. noise sequence \((Z_t)\) are independent, \((\sigma_t)\) is regularly varying with index \(\alpha\) \((> 0)\), and the \(Z_t\)s have moments of order larger than \(\alpha\). After a presentation of basic definitions and results concerning a regularly varying sequence, the study investigates some important examples. In the first case \((\sigma_t)\) is an exponential AR(1) process with regularly varying marginals. Then a positive power of \((\sigma_t)\) satisfies a random coefficient autoregressive equation. Finally the case in which \((\sigma_t)\) is some power of the absolute values of a regularly varying moving average process is considered.