台灣股票現貨對應股票指數期貨的避險比例與避險績效之研究-OLS, Rolling Regression, Bivariate CC GARCH and Bivariate DCC GARCH模型

Abstract

When investors hold stock future index to avoid investment risk, if they can choose the optimal estimated model to get the realized hedge ratio, they can get the better hedge performance. However, though the domestic relative researches about the estimation of the spot/stock future index have used OLS and GARCH models to estimate the realized hedge ratio, most of them just regard the reduction of variance as hedge performance. In this way, they may neglect the real demand of investors which is investors want to undertake every unit of risk to get the super profit. So we use the OLS, Rolling Regression, Bivariate CC GARCH and Bivariate DCC GARCH models to analyze three spot stock indexes and three future stock indexes. Therefore, we can get the realized hedge ratio and use the Sharpe index to compare the hedge performance of each model. The result shows that the realized hedge ratio of Bivariate CC GARCH model has better hedge performance than others. Besides, the realized hedge ratio of Bivariate DCC GARCH model is lower than Bivariate CC GARCH. It means when investors adjust the realized hedge ratio, they must avoid adjusting it frequently.