The Relationship among Institutional Ownership, Corporate Performance and Corporate Tax Avoidance

• Main Authors: LIANG, LI-HSIN, 梁立昕
• Other Authors: KUO, JENN-SHYONG
• Format: Others
• Language: zh-TW
• Published: 2019
• Online Access: http://ndltd.ncl.edu.tw/handle/9qeb8x

碩士 === 國立臺北大學 === 會計學系 === 107 === As there is no consistent conclusion of the relationship between institutional ownership and corporate tax avoidance, this study takes the empirical method of regression discontinuity design (RDD) to explore the relationship amoung institutional ownership, corporate performance and corporate tax avoidance. According to the empirical framework of Chen, Huang, Li and Shevlin (2018), the parameter method of the Sharp RDD was used. Taking the FTSE TWSE Taiwan 50 Index and Taiwan Mid-Cap 100 Index constituent stock as samples from 2008 to 2017, and taking the restructuring of exchange traded funds as an exogenous change to observe institutional ownership, corporate performance and corporate tax avoidance of the sample near the threshold. Indirect inference is the relationship between institutional ownership, corporate performance and corporate tax avoidance. The results suggest that the restructuring of exchange traded funds has a positive effect on institutional ownership, corporate performance, but there is no significant relationship between the extent of tax avoidance, and it can be indirectly inferred that institutional ownership has a positive relationship with the corporate performance. Further, the sample is stratified according to the level of corporate performance to explore the impact of exchange traded funds to the extent of tax avoidance, the results show better performance of tax avoidance companies are more aggressive. Therefore, this study suggests that the relationship between institutional ownership and tax avoidance is not a direct impact, but indirectly through influencing the strategy of tax avoidance, to achieve the purpose of improving corporate performance. In the sensitivity analysis, we also obtain the consistent result by using different estimation methods of sharp regression discontinuity design (nonparametric).