CFE 2015: Start Registration
View Submission - CFE
A0776
Title: Whittle estimation of multivariate exponential volatility models with long memory Authors:  Malvina Marchese - Cass Business School (United Kingdom) [presenting]
Paolo Zaffaroni - Imperial College London (United Kingdom)
Abstract: The strong consistency and the asymptotic normality of the Whittle estimate of the parameters in a class of multivariate exponential volatility processes are established. The class includes the multivariate Stochastic Volatility model with Leverage and the Constant Conditional Correlation model with EGARCH individual volatilities. Under the general (MEV) model the logarithm of the squared returns is decomposed into the sum of a signal vector-linear process and a white noise. We allow for correlation between the signal and the noise since it arises in the CCC model and in the MSV model as a consequence of leverage. We allow for a wide degree of persistence of shocks to the conditional variance, including both short and long memory parametrization of the signal process. We assess the small sample properties of the estimator by means of a Monte Carlo exercise. We present an empirical application and we discuss diagnostics to evaluate the appropriateness of the multivariate exponential specification.