Study On Effect Of Institutional Investors’Heterogeneity On Earning Management

The theory of Ownership Structure Balance believes that the ownership structure can affect the allocation of the company’s internal control among different shareholders. A reasonable arrangement of shareholders’equity can stimulate the formation of mutual supervision and mutual containment so as to contributing to the improvement of corporate governance structure, as well as the quality of accounting information of listed companies, thereby affecting the company’s internal governance mechanisms and the company’s value. With the successful implementation of the split share structure reform, institutional investors in China develop rapidly. By the end of2012, the ratio of professional institutional investors holding A stock market value circulation up to17.4%.As an important factor affecting corporate governance, institutional investors are the large external shareholders’oversight on the companies’insiders in essence. Previous empirical studies have found that because of the difference of the scale of the holding, holding time, regulatory costs, etc. institutional investors will exhibit positive or negative in behavior. Aggressive institutional investors in corporate governance often adopt a "hand vote" approach, through the exercise of shareholders’ voting rights, submitting motion shareholder lawsuits and other ways to participate in corporate governance so that the earnings management of listed companies is constrained. Passive institutional investors often take in "vote with their feet" way when involved in corporate governance. It doesn’t inhibit earnings management of listed companies but exacerbate the occurrence of earnings management instead.The purpose of this study is to explore whether the different institutional investors would show different performance and play a positive or negative role of supervision supervision when participate in Corporation governance Therefore based on the different characteristics in holdings of institutional investors, this paper focuses on the effect of institutional investors" heterogeneity on earnings management in listed companies. Firstly, this paper reviews some studies before include the study of motivation of earnings management, the relationship between institutional investors and the earnings management, the relationship between institutional investors’heterogeneity and earnings management. Secondly, after reviewing the multiple principal-agent theory, the theory of Ownership Structure Balance and the theory of Corporate Governance, we analyze the mechanism of institutional investors’heterogeneity affecting earnings management. Third, based on the theoretical analysis, this study takes the historical transaction information in Jun.2007-Dec.2011semi-annual report of the securities investment fund and set up seven characteristic variables from three aspects-the ownership concentration, ownership turnover and earnings sensitivity. Then we use factor analysis to pick up the feature of institutional investors’holdings and institutional investors will be divided into three types-positive, negative and mixed. The classification results show that positive institutional investors exhibit high portfolio concentration, low portfolio turnover and low trading sensitivity while negative institutional investors exhibit low portfolio concentration, high portfolio turnover and high trading sensitivity. The mixed type is between the above two. Afterwards, we set up three holding variables-positive holdings, negative holdings and mixed holdings through a direct method of grouping. On this basis, we took the Chinese listed company from A-share main board stock markets as research samples and analysis the effect of institutional investors’heterogeneity on earnings management and verify that the positive behavior of institutional investors can constrain earnings management of listed companies, the negative behavior can stimulate earnings management of listed companies and mixed behavior stimulated the earnings management in a certain extent when use descriptive statistics, correlation analysis and multiple regression analysis and other methods. Finally some suggestions are put forward.