規則導向式動態投資組合保險決策輔助模型之建構

Abstract

現今的投資環境複雜且多變，金融市場經常受許多因素的影響而瞬息萬變，不論對於一般的投資者或機構專業理財者而言都難以掌握其變化趨勢，本研究嘗試以時間不變性投資組合保險策略（Time Invariant Portfolio Protection，簡稱TIPP）為基礎，採用演化式演算法（Evolution Algorithms，簡稱EA）建構輔助決策模型，以協助投資者降低投資風險，避免鉅額損失。投資組合保險（Portfolio Insurance）是保本投資策略的一種，其概念是藉由付出少許的保險費用，以鎖定整個市場價格下跌時之風險，將損失控制於一定範圍之內，而市場上漲時卻又不失參與獲利機會。時間不變性投資組合保險策略是投資組合保險常用的策略之一，其中風險乘數（Multiple）及最低交易調整門檻值（Tolerance）是此策略的重要參數，全憑投資者經驗與喜好而設定，然而行為財務學者發現因為框架效應（Framing Effect）所致，投資者經常表現出過度反應或反應不足，因此無法根據市場變動設定適合的參數進行投資。本研究以演化式演算法動態調整風險乘數及最低交易調整門檻值並進行了兩階段式實證研究。實證結果顯示本研究所提之輔助決策模型確實可以尋得有效的參數，其模擬投資結果比傳統的TIPP策略及買入持有等策略表現更佳，顯示出本研究建議之模型是更保守且安全的投資組合保險模型。

Today‘s investment environment is complex and volatile, with numerous factors influencing the financial markets. It is difficult for individual and institutional investors to stay abreast rapid changes in this environment. This study attempts to investigate decision support models using evolution algorithms（EA）to explore the Time Invariant Portfolio Protection（TIPP） policy. Portfolio insurance is a principal-protected strategy that limits the investment portfolio losses to within a certain range when market prices decline, but also allows the portfolios to participate in profits when the market rises. Although Multiplier and Tolerance, concerned as the important parameters of TIPP, can be set up according to the individual's experiences and preferences, but scholars of behavioral finance have discovered that the investors tend to over- or under-react due to framing effect. Therefore, it is difficult for the investors to find out adequate parameters. The proposed models using evolution algorithms dynamically optimizes Multiplier and Tolerance parameters and two-stage experiments are conducted with the proposed models. The evaluation revealed that the evolutionary approaches dynamically identified satisfactory Multiplier and Tolerance parameters. The empirical results also suggested that the proposed models more conservative and safer than the buy-and-hold model and the conventional TIPP model.