Business Groups and Corporate Social Responsibility

Standard data sources on CSR such as Thompson Reuters ASSET4 classify them as society, environmental, and governance CSR in addition to economic activities. A detailed list of CSR data from Thompson Reuter ASSET4 database is in Appendix 2.

Documented benefits from CSR for firms include higher analyst followings and recommendations, and forecasting accuracy (Hong and Kacperczyk 2009; Ioannou and Serafeim 2015; Dhaliwal et al. 2012); effective corporate governance and enhanced firm value (Blazovich and Smith 2011; Jo and Harjoto 2011, 2012); lower cost of equity and higher credit rating (Dhaliwal et al. 2011; El Ghoul et al. 2011; Attig et al. 2013).

We define the cash flow–control ownership disparity as the wedge between the voting rights and cash flow rights of the controlling shareholders. The wedge indicates a deviation from the one-share-one-vote rule as identified by Gompers et al. (2010). The disparity is quite large in Korea with voting rights averaging 48.7% while cash flow rights averaging 17.9% over the period of 2004–2010 (Korea Fair Trade Commission 2010).

The chaebol system is still highly susceptible to corruption scandals. For instance, Lee Jae-Yong, head of Samsung Group has been implicated in the scandal in regard to the impeachment of former President Park Keun-Hye, and is currently undergoing trial. This is an unfortunate example of unethical strategic positioning of a few business groups in Korea.

Zhang et al. (2016) examine business groups in three countries, China, Japan, and Sweden, by providing a view as to why and how business groups solve economic problems, and how the specific national context of the group affects its performance. Their first observation is that the business groups emerged during times of institutional instability and persisted afterwards thanks to cooperative capitalism and export-oriented economy. Second, business groups have used diversification strategy to share risk at the group level and to reduce costs at the firm level. Third, business groups utilize a group planning unit to reduce problems associated with high management and coordinate costs spread out in multiple industries and in multiple locations. With respect to internal markets, there are national differences in intra-firm behaviors and in the ways in which national governments mobilize business groups to reach certain economic and social goals. They claim that the emergence and persistence of successful business groups shows that the Anglo-Saxon model of the firm may not be the only viable business model globally.

A detailed description of the CSR index is available at http://​extranet.​datastream.​com/​data/​ASSET4%20​ESG/​documents/​Thomson_​Reuters_​DS_​ASSET4_​ESG_​Content_​Fact_​Sheet.​pdf.

The Z-Score (or standard score) is a relative measure which compares one company with the benchmark of ASSET4 company universe. The score indicates the relative CSR value in units of standard deviation of that value from the mean of all companies. (Thomson Reuter’s data collection and rating methodology, 2012, available at: http://​extranet.​datastream.​com/​logon.​aspx).

Random-effect specification can also allow more degrees of freedom than fixed-effect model, because rather than estimating an intercept for every cross-sectional unit, we can estimate the parameters that describe the distribution of the intercepts, and can estimate coefficients for explanatory variables that are invariant over time.

Notice also that the number of clusters in our sample is 82, which is greater than the threshold; small cluster problem generally arises with 20–50 clusters depending on the screening criteria (Cameron and Miller 2015).

Although the KLD data have been used extensively in research on accounting, economics, finance and management for many years, they have been criticized because of their binary and qualitative nature. In addition, the KLD data have an unbalanced panel structure, and may suffer from selection bias due to non-proportional coverage. For instance, large U.S. manufacturing firms are over-represented relative to their importance in the economy. Mishra and Modi (2013) and Chatterji et al. (2016) argue that this may dampen the generalizability of empirical findings based on the KLD database.