本研究以2013年至2025年第三季臺灣國內指數一般股票型ETF為樣本，探討ETF成分股之上游曝險程度是否影響ETF追蹤誤差。本文依據產業價值鏈資訊建構ETF上游曝險變數，並以ETF報酬率與標的指數報酬率差異之標準差衡量追蹤誤差，同時控制費用率、成分股股數、成交量、基金年齡、複製方法及年度效果。研究結果顯示：（1）ETF上游曝險程度與追蹤誤差呈顯著負相關，表示上游曝險愈高，追蹤誤差反而愈低；（2）此一負向關係主要存在於市場型ETF與成長型ETF，產業型ETF與高股息型ETF則未達顯著水準；（3）改以TEA與TEC作為追蹤誤差替代衡量方式後，結果仍維持一致；（4）2019年前，上游曝險與追蹤誤差呈正相關，2019年後則顯著下降並可能轉為負向。本文補充過去ETF追蹤誤差文獻較少納入供應鏈位置觀點之不足，並提供投資人與ETF發行商評估追蹤風險之參考;This study examines whether the upstream exposure of ETF constituent stocks affects ETF tracking error， using Taiwan domestic index equity ETFs from 2013 to the third quarter of 2025 as the research sample. This study constructs the ETF upstream exposure variable based on industrial value chain information and measures tracking error as the standard deviation of the return difference between ETFs and their benchmark indices. The model further controls for expense ratio， number of constituent stocks， trading volume， fund age， replication method， and year fixed effects. The empirical results show that: (1) ETF upstream exposure is significantly negatively associated with tracking error， indicating that ETFs with higher upstream exposure tend to have lower tracking error; (2) this negative relationship is mainly observed in market ETFs and growth ETFs， while the results for sector ETFs and high-dividend ETFs are not statistically significant; (3) the results remain consistent when TEA and TEC are used as alternative measures of tracking error; and (4) before 2019， upstream exposure is positively associated with tracking error， whereas after 2019， this relationship significantly weakens and may turn negative. This study contributes to the ETF tracking error literature by incorporating the perspective of supply chain position and provides implications for investors and ETF issuers in evaluating ETF tracking risk.
