雜訊對臺灣股價指數期貨最適避險比率之影響

The Influence of Noise to the Hedge Effectiveness

期貨為避險之重要工具，過去文獻探討避險比率，均以成交價訊息作為避險比率之估計，然成交價之訊息中包含資訊與雜訊，而且期貨與現貨兩市場各自隱含之雜訊並不相同。因此，本文將避險比率的估計加入考量雜訊因素之影響，將原始價格訊息與去除雜訊之資訊，分別以OLS、ECM及GARCH模型估計相關避險比率樣本，並以台灣期貨交易所交易量最大的四種期貨商品加以實證。結果顯示，不論何種模型或商品，去除雜訊後，均能提高期貨與現貨資訊間之彼此相關性，因此所估計之避險比率較高，而在避險績效之表現上，在2008年市場波動較高而雜訊較大之情況下，以傳統成交價格估計之避險比率其避險績效較佳。然在市場較穩定下之2009及2010年，去除雜訊後估計之避險比率其避險績效則較佳。證明雜訊確實會干擾避險策略之運作，在市場盤整期間，則應以去除雜訊為考量，但在波動較大，雜訊大幅影響成交價格時，則應加入雜訊之全部資訊作為避險比率之估計，以獲得較佳之避險績效。

Future market provides an important tool for hedging. Previous studies on hedge ratio estimation have been focused on original transaction price. Nevertheless， transaction price includes both true information and noise. The markets for spot and future contain different kinds of noise. This paper proposes consideration of the effect of noise in original prices. Noise were both included and excluded in the computations of hedge ratio using OLS， ECM， and GARCH models. The approaches were verified in four largest futures products in Taiwan Stock Exchange. The results show that， after the removal of noise， the correlation between spot and future markets is increased. The estimated hedge ratio is thus higher. In regarding to the hedge effectiveness， using traditional transaction price to estimate the hedge ration has better performance in 2008 as there were high market fluctuation. Removing the noise has better hedge ratio performance in the years of 2009 and 2010 when more market were more stable. This shows that noise do have significant influence on the strategy of hedging. Noise should be removed when the market is relatively stable. When the fluctuation is considerable， noise should be included in order to achieve better hedge effectiveness.