- 學位論文
可調式馬可維茲模型的績效驗證-以台灣50指數與產業研究報告為例
Validation of the Adjustable Markowitz Portfolio Model–Using Taiwan 50 Index and Industry Research Report
摘要
在傳統投資組合理論中，馬可維茲(Markowitz,1952)所提出的平均變異模型讓投資人可以權衡風險與報酬下進行資產配置，然而僅以過去的歷史報酬與變異來預測未來，許多學者認為仍有諸多限制與缺失。可調式馬可維茲模型應用了模糊理論來延展馬可維茲模型，讓專家與投資人的主觀判斷一起納入對未來證券報酬的評估中。因此本研究以台灣50指數的成分股與產業研究報告的建議來投入模型，探討可調式馬可維茲模型在台灣股市的實際績效變化。本研究以2009至2010年台灣股市產業研究報告所提出的個股建議，來做為可調式馬可維茲模型中專家對於個股期望報酬的調整參數，透過不同的形成期與持有期間下進行實驗，並分析可調式馬可維茲模型與傳統馬可維茲模型之間的差異。 實驗結果發現可調式模型透過研究報告建議調整下，在持有4個月後優於馬可維茲模型。以收集資訊成本考量，以近期3個月內發布的資訊，推估未來的台灣股票市場似乎最有參考價值，並持有4或10個月會有不錯的績效。
並列摘要
In the traditional portfolio theory, the Mean-Variance model proposed by Markowitz (1952) that investors can tradeoff the risks and return of asset allocation. However, only using historical return and variance to predict the future, many scholars believe that have many limitations and defects. The Adjustable Markowitz model adopt fuzzy theory to extend the Markowitz model, the subjective judgment of experts and investors together in the assessment of future securities return. In this research, we adopt Industry Research Reports as the expert advice adjustable parameters of the Adjustable Markowitz model. The experimental results showed that the Adjustable model through the adjustment of Industry Research Report is superior to the traditional Markowitz model after holding four months. If the collection of information cost considerations, the collection of information three months and held ten months is the preferred option. To collect information on cost considerations, the information published in the recent three months, it seems most valuable reference to estimate future stock market in Taiwan.