CEO雙重職位與ESG表現之關聯性研究

Abstract

本研究旨在探討董事長兼任總經理與企業ESG績效的關聯性，並預期在存有董事長兼任總經理的企業，會因為董事會喪失原先有效監督職能，使企業在ESG整體乃至各構面下的議題造成衝擊，令企業ESG績效因此下降。研究樣本取自於台灣經濟新報 (TEJ) 資料庫並使用「TESG永續發展指標」資料，以2018-2022年間的上市櫃企業作為研究對象，使用最小平方法進行迴歸測試，研究結果顯示，企業在存有董事長兼任總經理的狀態下，對企業整體的ESG績效會有負向顯著的影響，支持董事長兼任總經理對ESG績效會有負向關聯的假說。然而在各構面下則呈現不一致的情形，在環境構面大部分未達顯著關聯，顯示集權制度對於環境構面的關聯較無直接關聯；在社會構面下亦呈現相關情形，僅有對少部分議題呈現顯著，並且方向不一致，顯示企業可能因董事長兼任總經理帶來有正向以及負向的影響，而非單一負面的影響；在公司治理構面下則呈現明顯的負面顯著，顯示董事長兼任總經理對於公司治理的衝擊尤其明顯。在額外測試中則以staggered DID模型進行測試，結果與最小平方法測試相比，更為確認董事長兼任總經理對於整體ESG績效有負面影響，在環境構面包含各相關議題確認皆無達顯著水準，代表董事長兼任總經理與環境相關議題較無關聯；在社會構面則有較大的差異，結果顯示社會構面呈現顯著負向關聯，並在五個子議題中兩個子議題呈現負向顯著關聯，與先前結果方向不一致相差較大，然因此模型提供更為細緻的測試方法，可能因此捕捉到並彌補OLS模型的不足，才藉此確認董事長兼任總經理對於社會構面呈現顯著負向影響；在公司治理構面則與OLS模型結果一致，提供更穩健的支持。

This study investigates the relationship between CEO duality—the dual role of the chairperson concurrently serving as CEO—and corporate ESG (Environmental, Social, and Governance) performance. It is hypothesized that CEO duality undermines the board’s supervisory function, thereby negatively impacting ESG performance across its three dimensions. The empirical analysis draws on data from listed companies in Taiwan during the period 2018–2022, obtained from the Taiwan Economic Journal (TEJ) database and the TESG Sustainability Index. Using ordinary least squares (OLS) regression, the findings reveal that CEO duality is significantly and negatively associated with firms’ overall ESG performance, supporting the hypothesis. However, the results vary across individual ESG dimensions. In the environmental dimension, most results are statistically insignificant, suggesting a weak or negligible link between centralized leadership and environmental issues. In the social dimension, only a few sub-topics show significant associations, with mixed directions, implying both potential positive and negative influences. In contrast, the governance dimension consistently shows a significant negative relationship, indicating that CEO duality notably weakens corporate governance performance. Robustness checks using a staggered difference-in-differences (DID) model further confirm the adverse impact of CEO duality on overall ESG performance. In the environmental dimension, results remain insignificant across all sub-topics, reinforcing the earlier conclusion of minimal relevance. In the social dimension, the DID analysis reveals a more pronounced and consistent negative relationship, with two out of five sub-topics showing statistically significant adverse effects—findings not observed in the OLS model. This suggests that the staggered DID model may offer finer estimation and better capture the detrimental effects of CEO duality, especially on social issues. As for the governance dimension, the results remain consistently significant and negative across both models, offering robust support for the hypothesis.