Effects of CEO’s Geographic Proximity on Corporate Social Responsibility

In this study, we examine the impact mechanism of the geographic proximity between a CEO’s native place and their company’s headquarters on corporate social responsibility (CSR), from the perspectives of social identity theory and upper echelons theory, as well as the moderators of these relationships. We assume that the geographic proximity will increase CEO willingness to fulfill CSR, and we predict that these relationships will be moderated by external differences in the institutional environment index of a CEO’s native place and the difference in Gross Domestic Product between a CEO’s native place and the location of their company’s headquarters. Using data for publicly listed firms from 2012 to 2020, we test our hypotheses, and the results largely validate our theoretical framework. By introducing CEO geographic proximity and CSR as the main research focus, this study contributes to the literature on the heterogeneity of geographic location and CEOs’ strategic decision-making.
To examine our hypotheses, we utilize Chinese publicly listed companies that have published CSR reports as samples. More specifically, our research samples are data from companies listed on the Shanghai and Shenzhen Stock Exchanges A-share from 2012 - 2020. The sample data mainly include the latitude and longitude of the city where the CEO’s company headquarters and the CEO’s native place, as well as data on CSR. Our data are derived from Hexun.com, the China Stock Market & Accounting Research Database (CSMAR), corporate annual reports, the WIND Database, and the China Economic and Social Big Data research platform (http://data.cnki.net/). We obtain the longitude and latitude of firms’ headquarters from CSMAR and data on CSR report scores from Hexun.com and CSMAR. We also manually collected each CEO’s native place and birth-place. To ensure sample reliability, we exclude data with ST and PT class flags from our sample, because these companies received special treatment due to abnormal financial situations, as well as observations with missing data and financial firms. The final sample comprises 6717 firm-year observations. To control for the potential effect of extreme value, all continuous variables are winsorized at the 1% and 99% levels.