- 學位論文
員工獎酬及公司治理對於短視研發行為之影響
The Effect of Employee Compensation and Corporate Governance in Myopic R&D Behavior
摘要
本研究以2000年至2012年間台灣電子業上市(櫃)公司為研究對象，主要探討員工獎酬、公司治理對於短視研發行為之影響。依據Bushee(1998)作法，以今年減除研發支出及稅前淨利(EBTRDt)較去年之變動數，跟去年研發支出(RDt-1)的比較關係分類為盈餘增加樣本(IN)、盈餘小幅度減少樣本(SD)，盈餘大幅度減少樣本(LD)三組樣本，其中，員工獎酬以員工分紅做為變數衡量，公司治理則以董監事控制席次比例；董監事持股質押比例兩個變數衡量。 實證結果發現：(一)當公司員工獎酬發放越多時，管理階層越不會透過減少研發支出來提高盈餘，故員工獎酬增加不會造成研發支出的短視行為。(二)當公司控制股東擔任董監事比例越高時，管理階層減少研發支出藉以提高盈餘的可能性將提升，故董監事控制席次比例增加會造成短視行為。(三)董監事持股質押比例並不影響管理階層的短視行為。 在邏吉斯迴歸模型下，盈餘小幅度減少樣本(SD)及盈餘大幅度減少樣本(LD)，此兩組樣本相對於參照組盈餘增加樣本(IN)其短視行為較明顯；在多元迴歸模型下，僅盈餘大幅度減少樣本(LD)，相較於參照組盈餘增加樣本(IN)其短視行為較明顯。
並列摘要
This study is focused on the listing electronics firms in Taiwan from 2000 to 2012. The objective is to explore the impact of employee compensation and corporate governance on myopic R&D behavior. We follow Bushee’s method (1998) comparing each firm’s EBTRDt (earnings before taxes and R&D) with its previous year’s R&D expenditures (RDt-1), then investigate myopic R&D behavior for the collected samples of earnings increase (IN), small earnings decline (SD), and large earnings decrease (LD). In this research, the main explanary variables include Employee Compensation (employee bonus ratio) and Corporate Governance (shareholding ratio and pledged shares ratio of directors and supervisors, respectively). The empirical results are as follows: First, the higher employee bonus ratio offered by companies, the less R&D expenditures cut by managers. Therefore, managers are less likely to conduct myopic R&D behavior when employee bonus ratio is increased. Second, the higher shareholding ratio of directors and supervisors, the more likely that managers may raise earnings by cutting firm’s R&D expenditures. Thus, managers are likely to conduct myopic R&D behavior if shareholding ratio of directors and supervisors is increased. Third, we found that pledged shares ratio of directors and supervisors has no effects on myopic R&D behavior in management. Compared to IN sample, myopic R&D behavior is more evident in both SD and LD samples for the logistic regression model, but the myopic R&D behavior is more evident only in LD sample for the multiple regression model.