Testing social and environmental disclosure-reputation relationship: a longitudinal two-way analysis

Abstract

Purpose: This paper reviews sustainability reporting understood as any type of social and environmental disclosures (SED) in its relationship with corporate reputation within the most reputed companies in Spain according to MERCO Business Monitor Ranking (2014-2016). Design/methodology/approach: To shed light on the relationship reputation-SED, two alternative models were tested, thought the use of Structural Equation Model (SEM) and Partial Least Squares (PLS), with longitudinal data. Findings: Both models supported the hypotheses although the model linking reputation to SED was slightly better, questioning the use of SED by reputation leader companies. Research limitations/implications: Our paper study the linkage, sign and causality, between reputation and SED by introducing two alternative models. SED and reputation are receiving considerable attention into the business scope, although their relationship is not agreed by previous literature. There are contradictory evidences that lead us to question the sense of this relation. Practical implications: Our contribution will be of interest to managers in terms of the value of this type of reporting from a strategic point of view. If reputation favours this type of disclosures, these will be issues to be considered in order to obtain a better competitive advantage through market differentiation. Social implications: The results will be of interest for future studies and actions aimed at regulating the improvement of this type of reporting in the hands of academics and practitioners but also investors and regulators. Originality/value: This study is an advance in the description of the SED-reputation relationship and contributes to this new line of research with new insights. Another contribution is the way to understand sustainability reporting. This paper analyses SED from the twofold point of view of the quantity of information and, the existing references about its quality, and adding the lag effect between both variables.