富櫃50指數之追蹤誤差成因探討

Abstract

本研究首先探討富櫃50指數的追蹤誤差大小，並分析富櫃50指數對於櫃買市場績效表現的代表性。接著，本研究進一步以Smart Beta策略概念，運用特徵選擇模型(CS)、Grinblatt and Titman模型(GT)、Carhart四因素模型及Jensen單因素模型等方式檢驗造成富櫃50指數產生追蹤誤差之成因。實證結果發現，富櫃50指數對於上櫃市場的代表性，並未有臺灣50指數對於上市市場的代表性來得好。本研究亦發現，富櫃50對於上櫃市場之追蹤誤差主要是來自於成分股的特徵屬性與上櫃市場全體公司的差異，而並非來自於ETF的交易成本。綜合四種績效評估模型結果來看，富櫃50指數成分公司的市值規模以及帳面市值比等兩項特徵因子，使富櫃50 ETF相較於大盤市場存在風險溢酬。

This study analyzes the tracking errors of TPEx 50 Index compared with Taiwan 50 Index to gauge how well TPEx 50 index is representing the performance of the over-the-counter market. Moreover, the study examines the determinant factors of tracking error with the concept of smart beta. We apply Characteristic Selectivity (CS) model, Grinblatt & Titman (GT) model, Carhart four-factor model, and Jensen one-factor model to evaluate the characteristics of the component stocks for Yuanta Taiwan TPEx 50 ETF.\nThe results suggest that TPEx 50 Index is less representative than Taiwan 50 Index for the respective markets tracked. The tracking errors for TPEx 50 are mainly stemmed from the excess return offered by the stocks characteristic rather than the expenses deducted from the ETF. Furthermore, the four fund performance evaluation models show that the market capitalization and book-to-market ratio carry risk premiums to Yuanta Taiwan TPEx 50 ETF.