運用資料探勘技術建構公司治理知識地圖與實證研究
No Thumbnail Available
Date
2010/08-2011/07
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
近年來，隨著美國安隆公司破產與國內力霸集團掏空等重大企業弊案，使得公司治 理議題廣受學術界、產業界與政府主管機關高度的重視，而其中問題的本源在於企業 所有權與控制權分離的公司制度設計，使得探討股權結構（包括內部人、外部人）、董 事會結構設計、董監事酬勞、經理人及員工的分紅制度設計、企業財務透明度、股東 行動主義活動等公司治理相關訊息與企業經營表現間的關係變得十分重要。然而，長 期以來，公司治理資訊與企業經營成效兩者間的關係呈現多元化實證見解，迄今仍有 許多爭議性的討論。舉例而言，管理股權與企業經營績效間的關係，究竟提升管理者 持股比例是會促使管理者的決策與股東利益趨於一致(alignment effect)，抑或增加管理 者圖利自我的機會而損及公司利益的情形(entrenchment effect)？總括而言，過去此議題 之相關研究因受到研究對象、研究期間長短、研究資料範疇等的差異，而呈現諸多不 同的觀察結論，因此，留下許多待進一步探討的空間。此外，過去相關研究的方法多 採用迴歸模型探討此議題，探討的範圍受限於研究者的假設範圍與相關的變數數目， 較無法作較多變數的觀察分析。因此，本研究計畫（二年期）之主要目的在於運用可 分析大量資料的資料探勘技術探討公司治理與企業經營表現間的關係，視覺化呈現多 個產業、多個期間的公司治理知識地圖，並從資料探勘的結果進一步利用結構方程模 式實證兩者間的複雜關係，以期能夠對於公司治理理論、實證研究及實務方案有所貢 獻。計畫第一年之重點在於蒐集與整理公司治理與企業經營績效資料，運用SOM 相關 演算法建構公司治理知識地圖系統，並以金融業為測試樣本。計畫第二年之主要重點 在延續運用第一年已發展之SOM 相關程式，再加上採行MDS 多變量統計方法，建構 其他產業之公司治理知識地圖系統，並採用科技接受模式評估系統成效，以及利用 LISREL 與PLS 實證公司治理與企業經營績效結構方程模式（係由前述資料探勘所發 現的關係知識發展而成）。
In recent years, researchers in academia, managers in businesses, and officials in government agencies have paid much attention to corporate governance since the explosion of several business scandals, such as Enron’s inside information trading in the United States and entrenchment of Rebar conglomerate in Taiwan. Corporate governance information has thus been one of the most important concerns in financial decision making for stock market investors, credit granting institutions, managers in businesses, and securities market regulators. Given this issue arising from the separation of ownership from control, it is important to explore the relationships among manifold ownership structures, compensation systems of managers and employees, board structures, financial transparency, shareholder activism, and performance indices of firms. However, considerable data exists regarding corporate governance. Therefore, an easy tool for visualizing and analyzing massive amounts of data and classifying companies into several categories according to corporate governance styles provides a useful decision support assistant for these stakeholders. Furthermore, the empirical results from previous research regarding the investigation of the relationships among the corporate governance information and firm performance show no inclusive findings. For example, alignment and entrenchment effects occur in different levels of managerial ownership, which generates several linear and nonlinear regression models (polynomial of degree n). Past researches heavily rely on regression methods to examine the relationships; however, they only examine the relationships between a few corporate governance variables and firm performance and seldom conduct a large scale study. Recently, data mining is an emerging and helpful approach for discovering some unknown patterns existed in mass data. Therefore, this research project will apply data mining techniques on the study of the relationships between corporate governance and firm performance indices for TSEC listed and GRETAI listed firms in Taiwan and use structural equation modeling for examination of the relationships discovered by data mining. Besides, technology acceptance model will be used to evaluate the usefulness and ease of use of the knowledge map system.
In recent years, researchers in academia, managers in businesses, and officials in government agencies have paid much attention to corporate governance since the explosion of several business scandals, such as Enron’s inside information trading in the United States and entrenchment of Rebar conglomerate in Taiwan. Corporate governance information has thus been one of the most important concerns in financial decision making for stock market investors, credit granting institutions, managers in businesses, and securities market regulators. Given this issue arising from the separation of ownership from control, it is important to explore the relationships among manifold ownership structures, compensation systems of managers and employees, board structures, financial transparency, shareholder activism, and performance indices of firms. However, considerable data exists regarding corporate governance. Therefore, an easy tool for visualizing and analyzing massive amounts of data and classifying companies into several categories according to corporate governance styles provides a useful decision support assistant for these stakeholders. Furthermore, the empirical results from previous research regarding the investigation of the relationships among the corporate governance information and firm performance show no inclusive findings. For example, alignment and entrenchment effects occur in different levels of managerial ownership, which generates several linear and nonlinear regression models (polynomial of degree n). Past researches heavily rely on regression methods to examine the relationships; however, they only examine the relationships between a few corporate governance variables and firm performance and seldom conduct a large scale study. Recently, data mining is an emerging and helpful approach for discovering some unknown patterns existed in mass data. Therefore, this research project will apply data mining techniques on the study of the relationships between corporate governance and firm performance indices for TSEC listed and GRETAI listed firms in Taiwan and use structural equation modeling for examination of the relationships discovered by data mining. Besides, technology acceptance model will be used to evaluate the usefulness and ease of use of the knowledge map system.