Item response models to measure corporate social responsibility

Abstract

Corporate social responsibility (CSR) is a multidimensional concept that involves several aspects, ranging from environment to social and governance. Companies aiming to comply with CSR standards have to face challenges that vary from one aspect to the other and from one industry to the other. Latent variable models may be usefully employed to provide a unidimensional measure of the grade of compliance of a firm with CSR standards, which is both understandable and theoretically solid. A methodology based on item response theory has been implemented on the multidimensional sustainability rating as expressed by KLD data-set from 1991 to 2007. Results suggest that companies in the oil and gas industry together with firms in industrials, basic materials and telecommunications have a higher difficulty to meet the CSR standards. Criteria based on human rights, environment, community and product quality have a large capacity to select the best performing firms, as they are very discriminant, while governance does not exhibit similar behaviour. A stock selection based on the ranking of the firms according to the proposed CSR measure supports the hypothesis of a positive relationship between CSR and financial performance.

Keywords:

• item response theory
• latent variable models
• portfolio management
• ranking
• socially responsible investment

JEL Classification:

• C40
• G11
• G14

Acknowledgements

The authors gratefully acknowledge financial support from the Swedish Foundation for Strategic Environmental Research (MISTRA) through the Sustainable Investment Research Platform (SIRP). The authors also thank Leonardo Becchetti and Rocco Ciciretti for their useful comments.

Notes

¹ It is theoretically possible to estimate the model simultaneously for all years; however, it would lead to serious problems of nonidentifiability and instability as the ranked companies vary from one year to the other.

² KLD Research and Analytics was acquired by RiskMetrics at the end of 2009. We focus here on data released prior to the merging.

³ The Domini 400 Social Index is now called MSCI KLD 400 Social Index.

⁴ Financial data were downloaded from Datastream that uses ISIN codes to identify the companies. Matching the financial data with the KLD data, identified through the names or the CUSIP codes, produced a loss of about 5% in the total market capitalization of our investment universe as some companies of the KLD data-set were not identifiable.

⁵ Extension to unordered responses, not used here, exist, see De Boeck and Wilson (2004, Ch. 3).

⁶ As already explained in the KLD data-set, sparsity of joint distribution is a feature of the KLD data-set and it is not due to the particular aggregation scheme.

⁷ Statistical measures based on AIC criteria show that this additional feature is unnecessary for the data at hand. For example, in 2005, the AIC for the richer model is 8580 to be compared with the AIC of the proposed model, i.e. 7982. Results are similar also for the other years.

⁸ The time series of the risk factors were downloaded from the K. R. French’s website.