結合風險值與濾嘴法則交易策略之研究－ 以美國及亞洲重要股價指數期貨為研究標的

Abstract

本研究採用納入風險考量的濾嘴法則模型（動態與靜態），利用相對風險值中的歷史模擬法（ED）、等權重移動平均（EQMA）、指數權重移動平均（EWMA）以及GARCH（1，1）這四種模型求出濾嘴比率。以美國及亞洲重要股價指數期貨為研究標的，從事實證研究。而本研究結合風險值與濾嘴法則的觀念，更是文獻上的首度嘗試。從結果觀察到，濾嘴法則交易策略適用在股市行情明確，動態模型又比靜態模型有更佳的表現，各種濾嘴比率模型下的操作績效，皆不亞於傳統的買入持有法（buy and hold）的操作績效。其中濾嘴法則在美國的S&P500、日本的Nikkei225與香港的恆生指數這三個市場，創造了異常報酬（abnormal return）。

This paper uses the VaR approach to determine the filter ratio to examine the trading performance. Based on the relative risk management on VaR, we use the ED, EQMQ, EWMA and GARCH (1, 1) to derive the filter ratio. We use the major stock market index futures-S&P500, Nikkei225, TAIEX, KOSPI200, Hang-Seng, Straits as research samples. In the meanwhile, we also use the Static and Dynamic models with four approaches- ED, EQMA, EWMA and GARCH (1, 1) separately to generate the filter ratio and also examine the trading performance. The results show that the negative relationship between the trading frequency and performance exists for ED, EQMA, EWMA and GARCH (1, 1) respectively due to the trading cost. The results also show that the positive relationship between the volatility and filter ratio. In the meantime, we find there is a significant difference between the Static model and Dynamic model on the performance. The performance of Dynamic model is better, especially for ED, EQMA and GARCH (1, 1). Also the performance of filter-rule trading strategy is better than the one of B&H (Buy and Hold). What’s more, abnormal returns by filter rule exist respectively in S&P500, Nikkei225 and Hang-Seng.