台灣股市投資人之處分效果實證：結合學習效果、投資人情緒與波動度

Abstract

本研究旨在探討行為偏誤中的處分效果與學習效果之間的關聯性，同時分析投資人情緒與波動度對處分效果是否具有顯著影響。本研究之研究期間將劃分為非疫情與疫情期間，資料來源挑選在研究期間內曾被納入過元大台灣 50 之成分股共計 66 檔股票作為樣本，並將投資人類型區分為散戶、外資、投信與自營商，使用 Panel Data 迴歸模型中的固定效果模型與隨機效果模型對不同類型之投資人進行分析。實證結果發現，在非疫情期間下之散戶與自營商具有處分效果傾向，疫情期間下則是各類型投資人皆具有處分效果，且處分程度之排序由大至小皆為：散戶、自營商、外資、投信，而處分效果提升程度由大至小依序為：外資、投信、散戶、自營商。在投資人情緒變數中，三大法人買賣超金額與券資餘額比皆與處分效果呈現顯著正相關，但本研究較支持以專業機構投資人為數據來源的三大法人買賣超金額之實證結果，且其迴歸係數與顯著程度皆較券資餘額比來得有力。而波動幅度在非疫情期間下，對於散戶與自營商顯示為顯著負相關，但其影響力有限，而在疫情期間時則對各類型投資人則無一致的解釋。

The study aims to examine the relation between the disposition effect and learning effect in nonrational behavior, also analyze whether investor sentiment and volatility have a significant impact on the disposition effect. The research period of this study will be divided into non-epidemic and epidemic period, the data sources are select form Yuanta Taiwan Top 50 ETF, and divide the types of investors into retail investors, foreign investors, investment trust and dealers, use the fixed effect model and random effect model of the Panel Data regression model to analyze different types of investors. The empirical results show that during the non-epidemic period, retail investors and dealers tended to have disposition effects, while during the epidemic period, all types of investors had disposition effects, and the order of disposition effects from large to small is: retail investors, dealers, foreign investors, investment trust, and the degree of rising of the disposition effect is: foreign investors, investment trust, retail investors, and dealers. Among the variables of investor sentiment, there are significantly positive relationship between Institution Investment OverBuy / OverSell, RGZ Ratio and disposition effect. However, our study supports the empirical results of the Institution Investment OverBuy / OverSell with professional institutional investors as the data source, and its regression coefficient and significance degree is stronger than the RGZ Ratio. During the non-epidemic period, there is a significant negative relationship between the Average True Range and disposition effect for retail investors and dealers, but its influence is limited. However, during the epidemic period, there is no consistent explanation for all types of investors.