| 摘要: | 國內債市在交易平台的有效率建置下，對於價格的透明度及交易的活絡度均有大幅度的提升，而一般交易商在判斷行情的走勢上除了基本分析外，最常用的就是技術分析法則，本研究主要就是在探討市場上常用的技術分析對於操作上是否有顯著之參考性。債市操作策略中，隨著借券制度之漸趨完備，價差交易策略可行性亦提升，因此本研究亦介紹了實務上一套利之機會及實證，結果為可行性相當高。另外經常發生在其他金融市場之季節效應，本研究亦一併予以實證。 技術分析法則中以移動平均線(MA)、指數平滑移動平均線(MACD)、相對強弱指標(RSI)及隨機指標(KD)作為本研究之實證對象。實證結果顯示利用MA訊號操作買進策略具顯著正報酬的頗多，賣出訊號參考性則較弱。利用MACD訊號亦顯示買進訊號部份出現顯著正報酬，賣出訊號參考性則較弱。利用KD訊號買賣都出現顯著之負報酬，顯示KD值在本研究中之操作式反而是相反方向有機會獲利。至於RSI訊號可分成兩部份，針對RSI值做買賣之策略出現顯著結果之報酬率幾乎為負數，而針對短期及長期RSI突破訊號做買賣之策略顯著結果中，買進策略為正報酬，賣出策略為負報酬，仍具交易上之參考性。 在季節效應之檢定上，發現有所謂之星期效應存在，即週三及週一之報酬顯著為正報酬，季節效應檢定結果為季節報酬無顯著差異存在，而檢定月效應上，發現九月份之報酬為顯著之負數，其他月份則不顯著，此皆可提供作為交易商進出之參考。 With the establishment of effective domestic bond trading platform, both the information disclosure and trading volume in bond markets have been substantially improved. Aside from fundamental analysis, proprietary traders frequently apply technical analysis to evaluate market trend. The purpose of this research is to investigate the effectiveness of technical analysis in bond trading. Along with the completeness of bond short-selling policy, bid-offer trading strategies have become more feasible. This study introduces an arbitrage trading strategy and its profitability. Furthermore, I also investigate the seasonality patterns in bond markets. These examines the validity of various technical tools, i.e. Moving Average (MA), Moving Average Converge/Diverge (MACD), Relative Strength Index (RSI), and Stochastic line (KD) in Taiwan’s bond markets. The results show that using MA signal to engage in a buying strategy often generates positive returns; however, it is less significant in a selling strategy. MACD signals also produce similar results, i.e. buying strategy generates positive returns and selling strategy does not show much relevancy. Moreover, using KD signal will result in negative returns, regardless of buying or selling. Using KD signal to engage in opposite direction trading has a greater opportunity to generate positive returns. As for RSI, the signals can be divided into two parts. Using RSI values in both buying and selling strategies will result in negative returns. On the other hand, by using short-term and long-term RSI signals, buying strategy will provide positive returns and selling strategy will provide negative returns. By examining the calendar effects, I find that there exist some weekday patterns; the returns on Wednesdays and Mondays are generally positive. When examining the monthly effect, the return for the month of September has notably been negative. There is, however, average returns are not significantly different across seasons. Overall, the evidence appears to suggest some potential profits for proprietary trading business. |
