最小化特定風險函數的動能策略

Abstract

動能策略的投資選擇方法曾被證實具有超額的報酬，其獲利原因有許多種說法。本論文首先探討動能策略在台灣股票市場是否同樣具有超額報酬，分析此策略在不同形成期與持有期是否能帶來超額利潤。因為動能策略為一個未考慮投資風險且採用均等權重的投資組合，所以我們進一步使用四種不同的資產配置模型對動能策略所選定的股票，在限制各別股票投資權重上限且限制股票種類數下，重新尋求風險最小的投資組合。這些模型考慮四個風險函數：平均絕對偏差、下方偏差動差、觀測期間的最大損失以及最小報酬，使得新建構的投資組合兼具動能效應且考慮投資風險。本論文以台灣股票市場2008至2012這五年的歷史資料做實證分析，結果顯示台灣股票市場具有動能效應，以及數學規劃模型能有效改良動能策略並降低VaR與CVaR風險值。

Momentum strategy has been proved existing excess return. However, there are many interpretations for the profitability of momentum strategies. This paper discusses whether momentum strategies in Taiwan stock market can earn excess returns and analyzes the strategy in different formation period and holding period. Since momentum strategies did not consider the investment risk and were equal weights portfolio, we use four different asset allocation models for the stocks has been selected by momentum strategy to construct a new portfolio that minimize the given risk function and limit the number of stocks. These models consider four risk functions: mean-absolute deviation, lower partial moments, and the maximum loss and minimum return in observation period, the resulting portfolio will have momentum effect as well as minimize the risk. The empirical study is performed by using the Taiwan market data from 2008 to 2012. The empirical results show that the Taiwan market has the momentum effect. The portfolio constructed by proposed models can improve the performance of momentum strategies and reduce the risk in terms of VaR and CVaR.