Title: The fractionally integrated log-GARCH model when the conditional density is unknown
Authors: Genaro Sucarrat - BI Norwegian Business School (Norway) [presenting]
Abstract: Financial time-series are frequently characterised by strong persistence. This can be due to ``long memory'' or structural breaks, or both. Recently, Fractionally Integrated log-GARCH (FI-log-GARCH) models have successfully been proposed as a model of such features. A drawback with these contributions, however, is that they rely on the conditional density being known. We propose estimation procedures that do not rely on the density being known. Monte Carlo simulations verify the usefulness of the procedures, and an empirical application provides an illustration.