MATH-605. STATISTICS SEMINAR. DEPARTMENT OF MATHEMATICAL SCIENCES.

DATE:	Thurday, November 29, 2007.
TIME:	1:15 pm. to 2:40 pm.
PLACE:	LN 2205.
SPEAKER:	Nadine McCloud, Department of Economics, Binghamton University.
TITLE:	"Specification tests for dynamic conditional correlations in multivariate GARCH models".

Abstract

The structure of conditional correlations has important implications for numerous types of eco- nomic and financial decisions including optimal portfolio diversification and hedging and risk manage ment. While many structures for dynamic conditional correlations have been put forward, specification tests for dynamic conditional correlations in multivariate GARCH models have not been introduced so far to the literature. This paper proposes an approach to testing the existence and specification of dynamic conditional correlations in multivariate GARCH models. The tests can identify linear and nonlinear misspecifications in dynamic conditional correlations. The tests are pure significance ests and are robust to the presence of conditional heteroscedasticity and higher order time-varying conditional moments of unknown form in the standardized error vectors of multivariate GARCH mod els. Our approach does not necessitate a particular parameter estimation method and distributional assumption on the error process. The asymptotic distribution of the tests is invariant to the uncertainty in parameter estimation. We assess the finite sample performance of our tests in the context of simulated data and two empirical applications and, where applicable, compare our results with some existing tests for constant conditional correlation.

This is a joint work with Yongmiao Hong .

Back to the Statistics Seminar webpage.