Permanency of CSR Activities and Firm Value

Among components of KEJI index scores: the soundness of corporate activities was estimated as 20 points, the fairness of corporate activities 10, contribution to social service 10, environmental protection 10, customer satisfaction 7, employee satisfaction 15, and contribution to economic development 28, respectively. Each component is calculated several subcomponents which consist of 64 items as of 2011. However, specific subcomponents have been changed over years, while seven major components are not changed in 2000s. The soundness of corporate activities is based on the composition of shareholders, financing methods, and soundness of spending activities, such as the relative magnitude of entertainment expenses to total salaries expenses, etc. The fairness of firm’s operation is based on the relative magnitude of internal transaction, use of market leading power for unfair transaction with clients, transparent disclosure of information, and the degree of cooperation with other firms. The subcomponents of contribution to social services consist of the protection of social minority groups and contribution to social welfare. Customer protection consists of the protection of consumer rights, product quality, and advertisement behavior. Environmental protection is based on the effort for environment improvement, the results of environmental improvement, compliance to environmental rules and regulations, environmental contamination, etc. Employee satisfaction is based on training, compensation, and welfare. Contribution to economic development is based on the efforts and results of research and developments, growth, profitability and exports, etc.

Teoh and Wong (1993) suggest that interaction variables with unexpected earnings and control variables be included in the ERC model. However, there exists multicollinearity issue when interaction variables are included in our regression model. Therefore, we follow Gelb and Zarowin (2002) who include only control variables in their regression. Also, analyst forecasts could be used when unexpected earnings are calculated. In this case, when there are no analyst forecasts available, smaller sample will be used in the analysis and reduce the statistical power of test. Due to this reason, there are many previous studies use last year earnings when they calculate unexpected earnings (Shivakumar 2000; Dhaliwal and Reynolds 1994; Ali and Zarowin 1992). When we use analyst forecasts in the analysis, we still obtain qualitatively similar results, even though slightly less statistically significant.

FnGuide is the database which provides stock returns and analyst forecasts for firms listed in the Korean Stock Markets.

TS2000 database provides financial statement data for firms listed in the Korean Stock Markets.

Even though all listed firms but firms in financial industries are evaluated, only top 200 ranked firms are announced. If the KEJI score of firms is not available, i.e., firm ranks outside top 200, we assign 0 for CSR activities (KDUM).