Tax risk factors and consequence in China

The main financial data comes from the CSMAR and WIND database. The use of financial derivatives of listed companies is obtained from the annual report of listed companies through Python software. All the A-share listed companies of Shanghai Stock Exchange and Shenzhen Stock Exchange from 2008 to 2016 were selected as initial samples, which were then screened based on the following conditions. First, excluding companies whose denominator is negative for the formula for calculating the actual tax rate during the sample period. Second, excluding companies with actual tax rates greater than 1 or less than 0 to avoid the impact of outliers. Third, excluding special treatment (ST) companies and financial industry companies to avoid the impact of special companies and special industries. Fourth, excluding samples missing values for other data required for regression. Finally, 13,765 companies’ annual observations were obtained. In order to eliminate the influence of outliers, the continuous variables in the model were tailed at 1% and 99% levels.本数据集主要财务数据源自CSMAR与WIND数据库。上市公司金融衍生品的使用数据通过Python软件从上市公司年报中提取得到。初始样本选取2008至2016年沪深A股全部上市公司，随后基于以下条件开展筛选：其一，剔除样本期内实际税率计算公式分母为负值的公司；其二，剔除实际税率大于1或小于0的公司，以规避异常值干扰；其三，剔除特别处理（Special Treatment，ST）公司与金融行业上市公司，以避免特殊主体与特殊行业的影响；其四，剔除回归分析所需其他数据存在缺失值的样本。最终得到13765个公司年度观测值。为进一步消除异常值的影响，对模型中的连续变量进行1%与99%分位的缩尾处理。