Timing effects of corporate social responsibility disclosure: an experimental study with investment professionals

ABSTRACT

Companies disclose increasingly more corporate social responsibility (CSR) related information. However, CSR information is not always treated entirely rationally by capital market participants. In an experiment using experienced investment professionals, we investigate how the timing of CSR disclosure influences firm valuations by professional investors. The results suggest that CSR disclosure in a stand-alone report, temporally disconnected to firm’s financial disclosure, may lead to asymmetric anchoring, whereby simultaneous disclosure of CSR and financial information in an integrated report prevents anchoring in investors’ judgement. Investors’ asymmetric anchoring is induced by differences in cognitive effort invested in CSR information processing, which depends on whether CSR information signals future profits or losses. Our results contribute to the debate on disclosure standards for CSR information and the use of CSR information by professional investors.

KEYWORDS:

• Anchoring
• corporate social responsibility
• CSR information
• integrated reporting
• sustainability

JEL CLASSIFICATIONS:

• C90
• G11
• Q56

Disclosure statement

No potential conflict of interest was reported by the authors.

Notes

1. In general, recent research has shown that such self-generated anchors trigger psychological processes different from those activated by experimenter-provided anchors (Mussweiler and Strack 1999, 2000). While experimenter-provided anchors lead to an enhanced accessibility of anchor-consistent information (Mussweiler and Strack 1999), self-generated anchors lead to insufficient adjustments (Epley and Gilovich 2001).

2. Importantly, however, financial data did not signal financial distress in the case with positive CSR information. We avoided this because financial distress likely represents a special case that is hard to generalize.

3. We made this request even though we did not provide stock data about the firm. In the experiment, we emphasized that we were aware that such a recommendation would typically require some information on stock data and asked participants for an educated guess without having these data.

4. For all statistical tests reported, p-levels are one-tailed for directional predictions and two-tailed otherwise.

5. For example, forewarning of an anchoring bias and increased incentives are likely to activate effortful thinking, and prior research has shown that both reduce anchoring effects for self-generated anchors (Epley and Gilovich 2005).