運用因子投資與懲罰範數建構投資組合 : 以美國股票市場為例

Abstract

根據 Singer and Beebower (1991)的研究，資產配置策略對於投資組合的報酬 績效貢獻高達 90%，因此藉由不同之建構投資組合的方法尋找資產的最適權重 分配一直是投資人所關心之重要課題。在過往的資產配置中，股票和債券等資 產一直是多元化投資組合的主要基石，然而自從 2008 年美國次貸危機爆發後， 許多金融資產遭到投資者嚴重的拋售壓力，導致各類資產普遍下跌的現象，金 融資產之間的相關係數上升，學界開始研究是否可以使用因子類比金融資產作 為投資組合的建構，降低投資組合內標的間的相關性。\n本文研究亦從此出發，從權威性金融期刊中挑選出因子，使用懲罰函數結合 平均數-變異數投資組合法，形成加權範數最小變異數投資組合;並運用十個績效指標來衡量加權範數最小變異數投資組合與其他三種標竿投資組合的表現

According to the research of Singer and Beebower (1991), asset allocation strategy contributes about 90% to the return performance of the investment portfolio. Therefore, finding the optimal weight distribution of assets through different methods of constructing investment portfolios has always been an important issue for investors. In the past asset allocation, stocks and bonds have always been the main points of diversified investment portfolios. However, since the outbreak of the U.S. subprime mortgage crisis in 2008, many financial assets have been under severe selling pressure from investors. With the phenomenon that assets are generally falling, and the correlation between financial assets has risen, researching whether factors can be used as financial assets as the construction of investment portfolios to reduce the correlation between investment portfolio internal assets.\nThe essay selected factors from authoritative financial journals, using norm penalty function combined with mean-variance portfolio method to form the Weighted-Norm Minimum Variance Portfolio (WNMVP) portfolio, then using ten performance indicators to measure the performance of Weighted-Norm Minimum Variance Portfolio and the other three benchmark portfolios.