Indirect Estimation of Long Memory Volatility Models

An indirect estimator is proposed for two long memory volatility models; the fractionally integrated generalised autoregressive conditional heteroskedasticity (FIGARCH) model and the long memory stochastic volatility (LMSV) model. The small sample properties of the indirect estimator are compared to the small sample properties of conventional maximum likelihood estimators. It is found that the indirect estimator has the potential to perform favourably with respect to maximum likelihood for higher order parameterised FIGARCH and LMSV models