An Application of Nonparametric Regression Methods On Hedging with Taiwan Stock Index Futures

• Main Author: 蕭雅文
• Other Authors: 吳榮昌
• Format: Others
• Language: zh-TW
• Published: 2004
• Online Access: http://ndltd.ncl.edu.tw/handle/29309744233512043736

碩士 === 實踐大學 === 企業管理研究所 === 92 === Hedge ratio estimation of stock index futures is important for stock investors. Several recent studies have found that time-varying hedge ratios lead to more risk reduction than traditional constant hedge ratios. The research compares traditional model with nonparametric regression model. The focus is this article applies two nonparametric regression methods, Kernel Regression Methods and Local Linear Regression Methods, to calculate hedge ratios of Taiwan stock index futures. Naïve model, OLS model, OLS-ECM model, GARCH-ECM model, Kernel Regression model and Local Linear Regression model are involved. These ratios are used to derive hedging portfolios for three investment portfolios, which are Taiwan Stock Index, TSE Electronic Sector Index, and TSE Banking and Insurance Sector Index. Hedging performances of nonparametric regression methods are compared with some traditional hedge models such as Naïve, OLS, OLS-ECM, GARCH-ECM . The major empirical results are as following： （1）Using unit roots testing for price series of stock index futures, we find the significance of unit roots and thus the nonstationarity of the price series, so price series should be differenced to induce stationarity. （2）We also find evidence of cointegration between spot and futures prices. Consequently,a cointegration measure should be taken into account in the hedge models. （3）The volatility by the time between spot and futures prices are similar, so every kind of hedging model in this article has good effect under hedging. （4）GARCH-ECM model and the nonparametric regression methods, especially Local Linear Regression Methods GARCH-ECM, have good effect to reduce the variation of investment portfolios after hedging. （5）In conclusion, the variation of investment portfolios after hedging all can reduce over 90 % variation.