Carbon emissions trading system and corporate tax avoidance

We selected companies listed on the Shanghai and Shenzhen stock exchanges from 2008 to 2020 as our initial sample. The list of companies participating in the ETS was obtained from the official websites of local Ecological Environment Bureaus and Development and Reform Commissions. Financial data were sourced from the China Stock Market & Accounting Research Database. The samples were processed as follows: (1) samples with missing data were excluded; (2) samples from the financial industry were excluded because of the special calculation and characteristics of financial statements; (3) samples from state-owned enterprises were excluded owing to their specific nature for tax avoidance; (4) companies identified as having special treatment were excluded to reduce the influence of outliers; and (5) winsorization was performed on continuous variables, truncating values at the 1st and 99th percentiles to eliminate the impact of outliers.

Between 2013 and 2014, China launched pilot ETS in Shenzhen, Beijing, Shanghai, Tianjin, Guangdong, Hubei, and Chongqing. In 2016, Fujian and Sichuan also joined the pilot program. To define the treatment group, we referred to the list of companies covered by ETS published by local regulatory authorities, which clearly identifies companies joining the ETS. Companies not included in these lists were categorized as the control group. To address potential endogeneity issues, this study employed the propensity score matching (PSM) method to select the sample. Referring to Chen et al. (2018) and Chen et al. (2022), this study considered companies who are in pilot ETS as the treatment group (Treat = 1) and non-pilot companies as the control group (Treat = 0). Specifically, this study used the following variables to conduct 1:1 nearest neighbor matching without replacement on the sample: static non-systematic risk (ALTZ), asset return volatility (EV), systematic risk (BETA), leverage ratio (Lev), fixed asset ratio (PPE), intangible asset ratio (Intang), return on investment (ROI), research and development expenses (RND), whether the chairman and CEO are the same person (Dual), the shareholding proportion of the largest shareholder (Top1), firm size (Size), return on assets (ROA), book-to-market ratio (Bm), audit opinion (Audit), whether the firm is in a loss in previous year (Loss), provincial-level EPU (EPU), and proportion of executives with political connections (PC).

Finally, 2,691 matched pairs of sample observations were obtained.

本研究以2008年至2020年沪深两市上市企业作为初始研究样本。参与碳排放权交易（Emissions Trading Scheme, ETS）的企业名单通过各地生态环境局与发展和改革委员会的官方网站获取。财务数据取自中国股票市场与会计研究数据库（China Stock Market & Accounting Research Database, CSMAR）。
样本处理流程如下：（1）剔除存在数据缺失的样本；（2）鉴于金融行业财务报表的特殊核算规则与特征，剔除金融行业样本；（3）由于国有企业在避税行为上具有特殊性，故剔除国有企业样本；（4）为降低异常值的影响，剔除被实施特别处理的企业样本；（5）对连续变量进行1%和99%分位的缩尾处理，以消除异常值的干扰。

2013至2014年，中国先后在深圳、北京、上海、天津、广东、湖北及重庆启动碳排放权交易试点；2016年，福建与四川也加入试点范围。为界定处理组，本研究参考各地监管部门公布的碳排放权交易覆盖企业名单，该名单明确标注了参与碳排放权交易的企业；未纳入上述名单的企业则被划分为对照组。为解决潜在的内生性问题，本研究采用倾向得分匹配（Propensity Score Matching, PSM）方法进行样本筛选。参考Chen等人（2018）与Chen等人（2022）的研究范式，本研究将参与碳排放权交易试点的企业划分为处理组（Treat=1），未参与试点的企业划分为对照组（Treat=0）。具体而言，本研究基于以下变量对样本进行1:1无放回近邻匹配：静态非系统性风险（ALTZ）、资产收益波动率（EV）、系统性风险（BETA）、杠杆率（Lev）、固定资产占比（PPE）、无形资产占比（Intang）、投资回报率（ROI）、研发支出（RND）、董事长与总经理是否两职合一（Dual）、第一大股东持股比例（Top1）、企业规模（Size）、资产收益率（ROA）、账面市值比（Bm）、审计意见（Audit）、企业上一年度是否亏损（Loss）、省级层面经济政策不确定性（Economic Policy Uncertainty, EPU）以及具有政治关联的高管占比（PC）。

最终，本研究共得到2691组匹配后的样本观测值。