以灰色預測模式改良台灣股票市場移動平均線投資績效之研究

本研究以台灣股票市場為研究對象，考量投資人實務交易之金融商品，採用台股期貨（TX）、台灣五十指數期貨（T5F）及台灣五十指數成分股，2009年1月1日至2017年12月31日研究期間之日、週及月股價資訊做為研究樣本，針對不同頻率之樣本計算移動平均線，建構短期、中期及長期波段操作策略，於考量交易成本的情況下進行模擬交易；日股價採用五日（MA5）、廿日（MA20）及六十日（MA60）區間之移動平均線，週股價採用四週（MA4）、廿四週（MA24）及四十八週（MA48）區間之移動平均線，月股價採用十二月（MA12）、廿四月（MA24）及六十月（MA60）區間之移動平均線。並藉由灰色系統理論改良移動平均線，將股價資料透過灰色預設模型GM（1，1）處裡，計算出白化後移動平均線，建構短期、中期及長期波段操作策略，進行模擬交易。最後分別針對白化前及白化後波段操作策略進行報酬差異分析，同時亦將白化前及白化後波段操作策略與買進持有策略進行報酬差異分析，最後本研究特別將白化前及白化後波段操作策略進行差分，比較差分後波段操作策略之差異。　　實證顯示，無論短期、中期或長期波段操作策略皆無法擊敗買進持有策略，而經過灰色預測模型或差分處裡之波段操作策略，亦然無法超越買進持有策略。而白化後波段操作策略無法超越白化前波段操作策略，差分白化後波段操作策略亦然無法超越差分白化前波段操作策略。亦即投資人於本研究樣本期間中，並無法透過移動平均線建構之波段操作策略獲得更高的超額報酬，且波段操作策略之績效並無法擊敗買進持有策略。推測因研究樣本取樣期間台灣股票市場為強勢多頭趨勢，由四千餘點一路上漲至上萬點，波段操作策略相較買進持有策略必須負擔更高的交易成本。台灣股票市場於2008年金融海嘯過後，市場效率性逐漸提升，投資人較難運用技術指標於股票市場中獲取較高的超額報酬，因此本研究不否認台灣股票市場似乎逐漸符合弱勢效率市場假說。

關鍵字

灰色預測模型； GM（1，1）；台灣股票市場；移動平均線；弱式效率市場假說；報酬差異分析

並列摘要

By using Taiwan index futures (TX), Taiwan 50 index futures (T5F), and Taiwan 50 index futures’ component stocks as examples in daily, weekly, and monthly from January 2009 to December 2017, this thesis discuss the performance of moving average considering the transaction costs, and try to establish a short-term, mid-term, and long-term exchanging strategies separately. 　　The various frequency samples are adopted in the following rules: using 5 days (MA5), 20 days (MA20), and 60 days (MA60) as a horizon to calculate daily moving average; using 4 weeks (MA4), 24 weeks (MA24), and 48 weeks (MA48) as a horizon to compute weekly moving average; and using 12 month (MA12), 24 months (MA24), and 60 months (MA60) as a horizon to calculate the monthly moving average. This thesis uses Grey System Theory to improve the investment performance of moving averages via Gray forecasting Model GM (1.1). This thesis focuses on the analysis of remuneration variances which are under the Band Trade Strategy (BTS) before and after the transparent treatment. At the same time, this study includes the analysis of remuneration variances on the Band Trade Strategy before-and-after transparent treatment and Buy and Hold Strategy. This study also includes the differentiations of before-and-after transparent treatment of Band Trade Strategy, comparing the differences after executing the difference method. 　　The study results show that, the short-term, mid-term, and long-term band trade strategies cannot beat the Buy and Hold strategy. Even though we improve the Band Trade Strategy through the Gray forecasting Model or differentiations, it is still not beat the Buy and Hold Strategy. Also, the Band Trade Strategy which through the processes of transparent treatment still cannot surpass the strategy which are not under the treatment or even under the differences methods. Hence, stock investors could not make more profits by the Band Trade Strategy structured by moving averages, and the performances of Band Trade Strategy cannot transcend the Buy and Hold Strategy in the Taiwan stock market. 　　The trend of Taiwan stock market was under the strong bull market after 2009. In this period, Taiwan’s stock markets upward from 4,000 points to 10,000 points, and the Band Trade Strategy bears more transaction costs than Buy and Hold Strategy. After the economic recession in 2008, the efficiency of Taiwan’s stock market increased. Investors had difficulties to earn excess return using technical analysis indicators. Therefore, the results show that the weak-form efficiency market hypothesis in Taiwan’s stock market cannot be rejected.

並列關鍵字

Grey Forecasting Model GM（1,1）； Taiwan Stock Market ； Moving Average ； Buy and Hold Strategy ； Band Trade Strategy ； Weak-Form Efficiency Market Hypothesis