近年來，產業型ETF因聚焦特定產業及其相關產業生態系而快速成長，其追蹤品質逐漸受到市場重視。然而，既有ETF追蹤誤差文獻多聚焦於ETF特性與市場因素，較少探討產業結構與供應鏈網絡特徵對追蹤誤差之影響。因此，本研究以2013年至2024年間臺灣掛牌之產業型ETF為研究對象，分別以水平產業集中度（Horizontal Industry Concentration， HIC）與垂直雙向供應鏈瓶頸（Vertical Two-sided Bottleneck， VTB）衡量產業配置結構及供應鏈依存風險，並進一步分析中美貿易衝突、COVID-19疫情及俄烏戰爭等外部衝擊之調節效果。研究採固定效果追蹤資料模型進行實證分析。 研究結果顯示，HIC與ETF追蹤誤差呈顯著負向關係，而VTB則呈顯著正向關係，表示產業集中所帶來之流動性與交易效率優勢有助於降低追蹤誤差，而雙向供應鏈瓶頸則會提高ETF追蹤誤差。進一步分析發現，中美貿易衝突與COVID-19疫情會強化HIC與VTB對追蹤誤差之原有效果，而俄烏戰爭則削弱或抵銷其原有效果。此外，客戶集中度與供應商集中度於一般市場環境下未具顯著影響，但於外部衝擊期間則轉為顯著，顯示供應鏈依存關係對ETF追蹤誤差之影響具有情境依賴性。 本研究將產業結構與供應鏈網絡觀點導入ETF追蹤誤差研究，補充既有文獻對ETF追蹤效率決定因素之理解，並提供評估產業型ETF追蹤品質與結構性風險之參考。;In recent years， sector exchange-traded funds (ETFs) have experienced rapid growth due to their focus on specific industries and industry ecosystems. While prior studies on ETF tracking error have primarily emphasized ETF characteristics and market factors， limited attention has been paid to the roles of industry structure and supply chain networks. Therefore， this study examines Taiwanese sector ETFs from 2013 to 2024 and investigates the effects of Horizontal Industry Concentration (HIC) and Vertical Two-sided Bottleneck (VTB) on tracking error. The moderating effects of the U.S.–China trade conflict， the COVID-19 pandemic， and the Russia–Ukraine war are also examined. Fixed-effects panel regression models are employed for the empirical analysis. The results indicate that HIC is negatively associated with ETF tracking error， whereas VTB is positively associated with ETF tracking error. This suggests that the liquidity and trading efficiency associated with industry concentration help reduce tracking error， while vertical two-sided supply chain bottlenecks increase ETF tracking error. Further analyses show that the U.S.–China trade conflict and the COVID-19 pandemic amplify the original effects of HIC and VTB， whereas the Russia–Ukraine war attenuates or offsets these effects. In addition， customer concentration and supplier concentration are insignificant under normal market conditions but become significant during periods of external shocks， indicating that the effect of supply chain dependence on ETF tracking error is context-dependent. This study extends the ETF tracking error literature by incorporating the perspectives of industry structure and supply chain networks. It also provides practical implications for evaluating tracking quality and structural risks in sector ETFs.
