- 學位論文
台灣指數期貨短線價差交易策略-以共整合方法
TAIWAN INDEX FUTURES SPREAD TRADING STRATEGIES IN A SHORT TERM- A COINTERGRATION APPROACH
摘要
台灣的期貨市場持續發展，隨著交易撮合方面的資訊技術提升，高頻交易的交 易量增加預期會較多，隨著法規管制的逐漸鬆綁，國內外機構法人的交易量增加預 期會較多，避險者成交量會跟隨股市成交量變化，投機者成交量則視行情的波動大 小變化，交易量增加較快的可能會是價差套利者。近期實務界與學術界更常利用的 方法是統計模型(Statistical Model)，這種模式的理論基礎，是基於兩組或兩組以上有 相關性的時間數列，如果可以找出一個長時間穩定的相關趨勢，可以利用來建立其 合理的價差區間，當實際成交價格超出此一價差區間，就可以進行買進相對低的， 賣出相對高的價差交易。 本研究以台灣期貨交易所台股期貨、電子期貨以及金融期貨的每五分鐘高頻資 料進行共整合分析，並擬定任兩期貨商品間的價差交易策略，藉由模擬結果驗證台 灣期貨市場的效率性，研究結果如下： 一、台指期、金融期與電子期皆為非定態時間數列，且三個變數中皆存在序列 相關現象。 二、台指期、金融期與電子指數期兩兩間皆有Granger 因果關係與共整合關係 存在。 三、台灣期貨交易所的三項指數期貨合約，任兩個商品的價差交易都能創造正 向報酬。
並列摘要
trading improved, the volume of high frequency trading will be expected much higher. As the regulation become loose, the trading volume of domestic and foreign institutional investors will increase more than anticipation. The volume of hedgers change by the stock market volume and the volume of speculators change by the variation of the whole market value. The volume of spread arbitragers may increase faster. Recently, the usual practice is Statistical Model. This kind of model is based on if we can find a long-term stable related tendency form paired time series, then we can use it to construct a reasonable spread interval. When the real price exceed over this interval, then we can long the weak and short the strong one. In this research, we use the five minute data of Taiwan stock index futures, electrical futures and financial futures to conduct cointergration test. Furthermore we work out a strategy and according to the strategy we simulate trading to test the futures market efficiency in Taiwan. The results as following: 1.Taiwan stock index futures, electrical futures and financial futures are nonstationarity and exist series correlated. 2.Taiwan stock index futures, electrical futures and financial futures are Granger causality and cointergation between one another. 3.The spread trading of three index contracts in TAIFEX can make positive profit between one another.