Please use this identifier to cite or link to this item:
|Title: ||Director network on firm performance and R&D expenditure decision|
|Other Titles: ||董事會網絡對公司績效與研發支出決策之影響|
|Authors: ||陳家妤;Chen, Jia-yu|
|Keywords: ||董事會;社會網絡;平方(非線性)關係;門檻效果;縱橫門檻自我迴歸模型;政策同質性;Boards of Directors;Social Network;Quadratic(Nonlinear) Relationship;Threshold effect;Panel Threshold Autoregression Model;Policy Homongeneity|
|Issue Date: ||2010-01-11 01:09:32 (UTC+8)|
此三部分的內容如下。第一部分研究董事會網絡、公司績效及所有權與控制權結構的相關性。有兩種競爭理論其結果是互相對立的。依據社會網絡假說，預期社會網絡與公司績效係呈正向關係(Kirchmaier and Stathopoulos, 2008)，然而代理假說則預期社會網絡與公司績效係呈負向關係(Kirchmaier and Stathopoulos, 2008)。實證結果顯示，董事會網絡一開始對公司績效有正向影響，但當董事會網絡連結太大時，則呈負向影響。本論文的另一個貢獻是控制股東的股權結構在董事會網絡與績效間扮演了一個重要的調節變數
第二部分本論文旨在探討台灣上市公司集團內與集團外董事會網絡運用及公司價值之門檻效果，運用縱橫門檻自我迴歸模型，檢測是否存在一最適集團內董事會網絡與集團外董事會網絡門檻值，使公司價值之走勢受到集團內董事會網絡與集團外董事會網絡之影響而呈現上下不對稱之兩個以上區間關係。研究以權益報酬率作為公司價值之衡量因子進行集團內董事會網絡與集團外董事會網絡門檻測試。實證結果發現集團內董事會網絡只存在單一門檻效果，估計之門檻值為32.15%，且其上下兩區間之影響係數 ( 及 ) 皆呈正向變動關係。但其中只有當集團內董事會網絡小於門檻值32.15%之區間係數 為顯著，高於門檻值之區間係數 則不顯著。另外，實證結果亦發現集團外董事會網絡只存在單一門檻效果，估計之門檻值為40.08%，且其上下兩區間之影響係數 ( 及 )皆呈負向變動關係。其中只有當集團外董事會網絡小於門檻值40.08%之區間係數 為顯著，高於門檻值之區間係數 則不顯著。建議公司可適度使用集團內及集團外董事會網絡，以達公司價值有效極大化。
第三部分本論文研究是否董事會網絡會影響公司的決策制定與政策的獨特性。研究依循Hochberg et al. (2007)和Kuhnen (2006)的社會網絡相似性分析方法並建構董事會網絡連結的衡量方法。實證結果顯示，連結越多的董事會在他們的研發決策上具有同質性。此外，產業集中度與公司的成熟度具有調節董事會網絡與研發支出決策獨特性的調節作用。研究結論對於公司的研發決策在公司財務及產業組織面向提供了一個新的看法，此外亦探究為何公司在創新決策具有群聚現象與台灣企業的研發浪潮。
This dissertation focuses on how director network impacts the firm performance and R&D expenditure decision. The first part entitled “The More Connections the Director has, The Better the Firm will Perform?”, the second part entitled “Threshold Effect between Intra-and Extra-Group Director Network on Firm Value of Listed Firms in Taiwan”, and the last part entitled “Director Networks and Their Similarity in R&D Expenditure Decision”.
These three parts are briefly introduced below: The first part of the dissertation is to test the relationships among director’s social network, firm performance, and ownership-control structure. There are two competing theories that predict contradictory outcomes. Following social network hypothesis, one would expect a positive relationship between social networks and firm performance (Kirchmaier and Stathopoulos, 2008), while agency hypothesis predicted a negative relationship between social networks and firm performance (Kirchmaier and Stathopoulos, 2008). This dissertation concludes that the director network initially have a positive influence on firm performance, however, when the magnitude of the director network is too high there is a negative effect. Our study also contributes to the literature by documenting that controlling shareholders’ ownership structure serves as an important moderate variable between director network and firm performance.
The second part of the dissertation aims at investigating whether application of intra-group director’s network and extra-group director’s network affect firm value of Taiwan listed firms. We apply panel threshold regression model to test if there exist optimal intra-group director’s network and extra-group director’s network, which may result in threshold effect and asymmetrical responses of the firm value to intra-group director’s network and extra-group director’s network. ROE is selected as the proxy variables to represent the firm value of the observed listed firms. Empirical result shows that intra-group director’s network only one single threshold effect exists, while all of the observations can be divided into two “regimes” depending on whether the intra-group director’s network is small or larger than the specific threshold value (32.15%). Two coefficients ( and ) are all positive with the evidence that the in the low intra-group director’s network is significant, while the in the high intra-group director’s network is not.
Another, our empirical result shows that extra-group director’s network only one single threshold effect exists, while all of the observations can be divided into two “regimes” depending on whether the extra-group director’s network is small or larger than the specific threshold value (40.08%). Two coefficients ( and ) are all negative with the evidence that the in the low extra-group director’s network is significant, while the in the high extra-group director’s network is not. This suggests that corporation should use intra-group director’s network and extra-group director’s network wisely in order to maximize the firm value and stockholders’ welfare.
The final part of the dissertation investigates whether members of a tight social network could impact directors'' decision-making and affect the policy uniqueness of a firm. In order to create a proxy for directors'' social connections, we employed methods from social network analysis similar to those used by Hochberg et al. (2007) and Kuhnen (2006) and constructed a measure that captures the various dimensions of connectivity, such as how well-connected the director is and the quality and predominance of each director''s connections in the network. Empirical results demonstrate that firms whose directors are well connected and whose connections are with better connected directors exhibit homogeneity in their R&D policy. Furthermore, empirical results suggest that both the concentration magnitude of industries and the maturity of the firms moderate the relationship between director network and R&D policy similarity. The results provided a new perspective of looking at firms’ R&D decisions by blending corporate finance and industrial organization together. We also provided some intuitions regarding why firms tend to cluster their innovation policy together and why we observe these policies occurring in waves.
|Appears in Collections:||[財務金融學系暨研究所] 學位論文|
All items in 機構典藏 are protected by copyright, with all rights reserved.