Restatement of CSR Reports : Frequency, Magnitude, and Determinants

We provide the first direct analysis of the reliability of the quantitative information disclosed in Corporate Social Responsibility (CSR) reports. We examine the frequency, magnitude and determinants of CSR report restatements for the Global Fortune 250 (G250) from 2006 to 2013. During this period, 39% of G250 CSR reports were restated, with a monotonic increase across time from 29% in 2006 to 53% in 2013. The median magnitude of line item restatement is 10%, with a bias toward overstatement. In our exploration of determinants and economic incentives associated with CSR report restatements, we find that restatement frequency is positively associated with firm environmental and social complexity and that restatements occur more frequently in firms that have reported a high level of social performance and have environmental targets. We also find that frequency of restatement is positively associated with firms that reside in strong law countries and have their CSR reports audited. Our analysis of reporting bias indicates a negative association between adoption of GRI reporting guidelines and the likelihood of an overstatement. We also find a positive association between having the CSR report audited and the likelihood of overstatements. Together, our results indicate that CSR information could be unreliable and firms that face pressure to perform well have more restatements. However, use of measurement guidelines helps restrict managers' disclosure choices and thus prevents opportunistic choice of measurement methods. Auditors also appear to invest more time and effort to reduce over - rather than understatement of performance