| Summary: | 碩士 === 國立高雄應用科技大學 === 商務經營研究所 === 93 === The purpose of this research is to discuss hedging effectiveness in GBF of R.O.C. under several hedging models. These models are Naïve model, OLS, and GARCH (1,1). All the data range from January 2, 2004 to March 31, 2005, 307 are collected. Moreover, the in-sample data is from January 2, 2004 to December 31, 2004, and then the out-of-sample data is from January 3, 2005 to March 31, 2005. The hedging period is divided into daily and weekly data. The hedging method is divided into dynamic hedging and static hedging. The hedging performance is measured by the decreasing percent of variance.
The empirical results find：Using unit roots testing for all data, we find the significance of unit roots and thus the nonstationarity of the return series, so return series should be differenced to induce stationarity. When hedging period is everyday, the performance of near future contracts is the best no matter in which kind of hedging models. Between Naïve model, OLS model, and GARCH model, there is not an absolute model can reach best performance in any different hedging periods and expiry contracts. There is no particular advantage of using the dynamic hedging than static hedging. Not a single approach performs better than any others ones. Hence a hedger should adopt proper approach by considering different hedging periods and expiry contracts.
|
|---|
