Corporate social responsibility (CSR) rating agencies have arisen with the aim of providing external and reliable information about business behaviour. The purpose of this paper is to present a multi-criteria methodology for integrating CSR valuations with the financial performance of companies in a unique measure of global sustainability performance.
The authors present a hybrid TOPSIS methodology on transformed scores of both the CSR valuations and the financial ratios. The “attribute-specific evaluation” approach into Multi-attribute Prospect Theory (PT) has been applied and the Design of Experiments (DoE) is used with the TOPSIS value of the firm as the response variable.
The proposal has been applied to 118 companies evaluated by Vigeo and Covalence CSR agencies. The authors also have considered five financial ratios of the companies in order to assess their financial performance. Consistent aggregation for firms has been achieved. Relationships between the different rankings, both those of Vigeo and Covalence and the ones constructed in this research, have been analysed. All top 10 Global sustainable firms rank among the top 10 positions in at least one of the remaining rankings. The results show that Vigeo and Covalence provide different information about the CSR behaviour of the companies.
Another interesting question is to study the discrepancies between Vigeo and Covalence, for example, in which areas there is the greatest divergence between the two agencies and what could be the reasons for this.
The results of this research could be of interest for both investors who want a global picture of companies in their selection process and stakeholders concerned with CSR issues who want to take advantage of different CSR ratings.
The application of PT softens the compensatory behaviour of the classical TOPSIS that may prove unsuitable for social evaluation. The DoE allows the aggregation of the weight sets from various decision markers. The combined methodology facilitates the scoring of new firms and the rank reversal problem can be mitigated with this methodology.
