Although there is $2 trillion in portfolios using socially responsible investing (SRI) criteria, it remains unclear how to measure “social responsibility.” We explore competing theoretical perspectives that explain the level of convergent and predictive validity across SRI ratings produced by competing social raters. While some prior literature predicts low convergent validity due to desire for differentiation, other work predicts high convergent validity driven by high true validity or by neo-institutionalist forces that reward imitation. We find that these ratings have low correlations and that firms with high and low social ratings are equally likely to be later embroiled in scandals.

Ratings of corporations’ environmental activities and capabilities influence billions of dollars of “socially responsible” investments as well as some consumers, activists, and potential employees. In one of the first studies to assess these ratings, we examine how well the most widely used ratings—those of Kinder, Lydenberg, Domini Research & Analytics (KLD)—provide transparency about past and likely future environmental performance. We find KLD “concern” ratings to be fairly good summaries of past environmental performance. In addition, firms with more KLD concerns have slightly, but statistically significantly, more pollution and regulatory compliance violations in later years. KLD environmental strengths, in contrast, do not accurately predict pollution levels or compliance violations. Moreover, we find evidence that KLD’s ratings are not optimally using publicly available data. We discuss the implications of our findings for advocates and opponents of corporate social responsibility as well as for studies that relate social responsibility ratings to financial performance.