A Minimum Distance Estimator for Dynamic Conditional Correlations
The crucial problem in estimating dynamic conditional correlation models is the
need to guarantee a positive-definite covariance matrix. In order to avoid any violation
of this property, many estimators impose strong restrictions on the model. In addition,
many models do not parameterize the correlations directly but the covariance. This
paper avoids this problem in proposing a minimum distance estimator (MDE) to estimate
dynamic conditional correlations or multivariate GARCH models. The model
allows full flexibility in the estimation. Violations of the positive-definiteness of the
covariance matrices are part of the specification tests. A simulation study shows the
performance of the estimator and the empirical section compares estimates of the MDE
with the DCC estimator of Engle (2002). This paper does not only demonstrate an alternative
estimator but also outlines how this model can be used to analyze the influence
of the restrictions on the estimates in multivariate GARCH models.
BAUR Dirk;
2006-03-01
JRC31450
EUR 21889 EN,
https://publications.jrc.ec.europa.eu/repository/handle/JRC31450,
Additional supporting files
| File name | Description | File type | |
|---|
