以優勢約略集合理論與形式概念分析探勘標準普爾500公司的環境、社會責任及公司治理與獲利之關係

Abstract

近年來，全球氣候變遷加劇、貧富差距日增、各種社會問題嚴重，為追求永續發展，環境、社會和治理(Environment，Social，and Governance，ESG)相關議題備受關注，ESG的概念不僅為企業落實聯合國2015年所制定可持續發展目標 (Sustainable Development Goals，SDGs) 帶來了明確的指引，也為評估企業的重要指標。唯過去研究大多僅基於ESG之綜合分數，或是E、S、G個別分數進行分析，其中極端值所帶來的影響多被忽略。為探討ESG各子項目得分對企業財務指標之關係，本研究導入以優勢約略集合 (Dominance Based Rough Set Approach，DRSA)為基礎之多準則決策分析模式，以標準普爾500指數為研究對象，探討ESG各項目得點對於所有企業與一般公司獲利之推論關係 (inference relationship)。首先，本研究採用優勢約略集合和形式概念分析法 (Formal Concept Analysis，FCA)，分析2017年至2021年標準普爾500公司中，所有企業與科技公司之ESG得分與財務指標之推論關係。在第二階段，則邀集財經專家，提供ESG得分與財務指標間因果關係之意見，首先，採用基於決策實驗室分析法 (Decision Making Trial and Evaluation Laboratory，DEMATEL) 之網路流程(DEMATEL-based Analytic Network Process，DANP)，並針對兩階段方法所得之結果進行比較。研究結果指出，每股盈餘(Earnings Per Share，EPS) (f1)、生態環境照管(Environmental Stewardship) (e2)、訓練與發展(Training and Development) (s3)、資產報酬率(Return on Assets，ROA) (f3)，是優勢約略集合所擷取37條決策規則中最常見的概念。此外本研究也發現，優良的多樣性(Diversity) (s1)和法務會計(Forensic Accounting) (g4)表現，為低、中等及以上水平股東權益報酬率(Return on Equity，ROE) (f3)相對應的關鍵概念。而在資訊科技產業中，影響股東權益報酬率(f3)表現最多的ESG子項目為社會構面。最後，兩個階段呈現的結果並不完全相符，社會構面是影響程度最大的之構面，而各構面底下影響程度最大的準則分別為碳排放與空污控管(e1)、多元性(s1)、企業透明度(g3)、以及每股盈餘(f1)。本研究的結果可作為投資者的參考，幫助他們了解ESG的哪些因素能夠代表財務指標的增長，有助於識別關鍵的ESG要素。In recent years, global climate change has intensified, wealth disparity has grown, and various social issues have become more severe. In pursuit of sustainable development, the issues related to Environment, Social, and Governance (ESG) have garnered significant attention. The concept of ESG not only provides clear guidance for businesses to implement the United Nations' Sustainable Development Goals (SDGs) set in 2015 but also serves as important indicators for evaluating companies. However, most previous studies have primarily relied on the overall ESG scores or individual scores for Environmental (E), Social (S), and Governance (G) pillars, often overlooking the impact of extreme values. To explore the relationship between ESG sub-category scores and corporate financial indicators, this study introduces a Multi-Criteria Decision Analysis (MCDA) framework based on the Dominance Based Rough Set Approach (DRSA). The research focuses on the Standard& Poor 500 (S&P 500) index and investigates the inference relationship between ESG scores and the profitability of both firms and general companies. Firstly, this study employs the DRSA and Formal Concept Analysis (FCA) to analyze the inference relationship between ESG sub-category scores and financial indicators for all companies and information technology (IT) companies within the S&P 500 from 2017 to 2021. In the second stage, financial experts are consulted to provide insights into the causal relationships between ESG scores and financial indicators. Initially, the Decision Making Trial and Evaluation Laboratory-based (DEMATEL-based) Analytic Network Process (DANP) is utilized for network analysis, and the results from both stages are compared.The research findings highlight that Earnings Per Share (EPS) (f1), Environmental Stewardship (e2), Training and Development (s3), and Return on Assets (ROA) (f3) are the most common concepts extracted by the DRSA from the 37 decision rules. Furthermore, the study also identifies that high Diversity (s1) and high Forensic Accounting (g4) performance correspond to key concepts related to Return on Equity (ROE) (f3) at the level of low, middle, or above. In IT sector, the Social pillar emerges as the most influential ESG sub-category affecting ROE (f3) performance.Finally, the results found in the two stages do not align completely. The Social pillar exhibits the greatest impact among all the dimensions. The most influential criterion within each dimension are Emissions (e1), Diversity (s1), Transparency (g3), and EPS (f1). The outcomes of this study can serve as a reference for investors, helping them comprehend which elements in ESG can represent the growth of financial indicators. This understanding contributes to the discrimination of critical ESG factors.