The Downside of Being Responsible: Corporate Social Responsibility and Tail Risk

There is a debate on whether businesses actually have responsibilities towards society, other than to increase profits. Friedman 1970) defends the opinion that businesses cannot have responsibilities, and that the best contribution businesses can make to society is to make profits. This debate, however, does not fall within the scope of this paper.

Another common measure of downside risk that has gained tractability over the past years is the expected shortfall (ES). The ES measures the expected loss on a portfolio given a violation of the VaR criteria. A common critique of the VaR measure is that it is not subadditive, and therefore, not a coherent risk measure. The ES resolves this concern by having the property of subaddivity. However, Danielsson et al. (2001) show that in the tail region the VaR also maintains a subadditive property.

Even when a stakeholder approach is chosen to measure CSR, several measurement issues remain. Turker (2009) categorises four approaches toward measuring CSR using the stakeholder method: the use of reputation indices, issue indicators, content analysis of corporate publications and scales measuring CSR perception at the individual managerial level. A well-known example of a reputation index is the Fortune index, which ranks the top 50 global companies according to their CSR performance. Issue indicators are one or more specific indicators, such as pollution or corporate crime, which are assumed to proxy for the entire CSR performance of a company. Content analysis of corporate publication is growing in popularity as companies pay more attention to social disclosure, but the information companies give in a corporate report can be different from their actual actions (see McGuire et al. 1988). The individual scale method measures the CSR values and attitudes of individual managers, but is not suitable for measuring organisational involvement with socially responsible activities (Turker 2009). In summary, there are various ways to measure CSR, but all methods have their advantages and disadvantages and it is presumably best to combine different methods.

DSR Bunnik was taken over by Sustainalytics Amsterdam in January 2010.

Established in Fribourg (CH), SiRi was a cooperation of 10 European social research companies that developed a common research questionnaire for analysing companies from different countries according to an identical structure on their corporate social responsibility. See www.siricompany.com for more information. The organisations ended their cooperation by the end of 2009.

A specific example of one aspect is the transparency of the employee policy. A score of 100 % is assigned if the company annually disclosed a report that contained a description of different programmes for employees, covered at least half of the employees of the company and is no older than two years. A score of 80 % is assigned if the company only published a code of conduct, although this code of conduct had to cover at least half of the employees. A score of 40 % is assigned if the report is older than two years, applies to less than half of the employees or had not been made public. A score of 30 % is assigned if the company is working on its first report. A score of 0 % is assigned if there is no report and the company did not plan to publish one.

The six control variables leverage, market-to-book ratio, price-earnings ratio (PE), size and the two proxies for liquidity, namely spread and turnover, are all retrieved from Datastream as well.

We measure the leverage as a ratio of debt-to-equity. The liquidity is measured in two ways: bid-ask spread and total share turnover, and the size of the company is captured by market capitalisation.

The regression model was extended with dummy variables for geography and industry.

The MSCI World Index reached its lowest point in February 2009.