基于最小二乘与加权分位数回归法对比分析盈余管理在我国跨国公司中的应用

Application of Earnings Management in Transnational Companies of Our Country Based on Comparison Between Weighted Least Squares Regression and Quantile Regression

为了对比普通最小二乘法与加权分位数法的异同,并选择能够衡量我国跨国公司盈余管理程度的有效方法,按照跨国公司的定义,把2013年我国沪深两市上市公司分为跨国公司和非跨国公司两类,并在Jones模型的基础上,分别采用普通最小二乘法和加权分位数回归方法进行对比实证研究。结果显示,总资产是影响我国跨国公司盈余管理的重要因素。表明加权分位数回归的Jones模型能够作为衡量我国跨国公司盈余管理程度的有效方法。

In order to compare weighted least squares regression and quantile regression and find a good way to measure the degree of earning management, we dividend the listed companies of Shanghai and Shenzhen stock market in 2013 into two categories, including transnational company and non- transnational companies. Based on Jones model, we used weighted least squares regression and quantile regression method for empirical test. Comparing the results of two regression methods, this article draws the following conclusions： total assets are an important factor of transnational corporation earnings management, and the weighted quantile regression model can be an effective measure of transnational corporation earnings management.