business group contagion effect in china
Considering the limitation of data sources and referring to previous studies, this paper defines that when the actual controller of a listed company controls two or more listed companies at the same time, the listed company belongs to a business group. Based on the 2007-2019 Shanghai and Shenzhen A share listed companies’ samples, and according to the ownership structure chart disclosed in the company’s annual report, this paper determines whether a company belong to a business group and select the group-holding listed company as the research samples (Buchuk et al., 2014; Na et al., 2017). A-share listed companies in China generally have specific actual controllers. Consider the differences between state-owned firms and private firms, this paper defines the actual controllers of state-owned firms are the economic entity that closest to the governments in ownership structure charts. The actual controllers of private firms are the natural persons who control the listed company. This paper does not choose the government institutions as the controllers of SOEs are because by this definition, all the SOEs will belong to one group, which are going against our economic intuition. Then this paper excludes the group-holding samples of financial companies and with insufficient financial data to obtain control variables. Finally, this paper obtains 6693 firm-year observations from 2007 to 2019 with the necessary data to conduct our analysis. The data required in this paper comes from the CSMAR database and WIND database. To mitigate the effects of outliers, this paper winsorizes continuous variables at the 1% level in both tails. This paper performs Logit regression to study the contagion effect of overconfidence, performs PSM and controls firm level fixed effect to alleviate potential endogenous problems. The form of the data is widely used in the research of business group management and governance in both China and other countries (He et al., 2013; Buchuk et al., 2014), and this paper believes this data can continuously support the future research in this field.