Dynamic hedging performance with the evaluation of multivariate GARCH models: evidence from KOSTAR index futures

摘要

This article examines the hedging performance of the conventional Ordinary Least Squares (OLS) model and a variety of dynamic hedging models for the in-sample and out-of-sample periods of Korean daily Korea Securities Dealers Automated Quotation (KOSDAQ) STAR (KOSTAR) index futures. We employ the rolling OLS and various popular multivariate Generalized Autoregressive Conditional Heteroscedasticity (GARCH) models to estimate and forecast the conditional covariances and variances of KOSTAR spot and futures returns. This article finds that dynamic hedging methods outperform the conventional method for the out-of-sample period. However, the simple rolling OLS is superior to all the other popular multivariate GARCH models.