Article ID: | iaor20101192 |
Volume: | 26 |
Issue: | 1 |
Start Page Number: | 180 |
End Page Number: | 194 |
Publication Date: | Jan 2010 |
Journal: | International Journal of Forecasting |
Authors: | Yilmaz Fatih, Harris Richard D F |
This paper proposes a hybrid multivariate exponentially weighted moving average (EWMA) estimator of the variance–covariance matrix of returns. The proposed estimator employs a range-based EWMA specification to estimate the conditional variances of returns, and a standard return-based EWMA specification to estimate the correlation between each pair of returns. The hybrid EWMA estimator offers an improvement over the standard EWMA estimator, both statistically and economically. Moreover, the hybrid EWMA estimator is less sensitive to the choice of decay factor.