Unaware corporate social responsibility: impact of firm size, motivations and external pressures

ABSTRACT

We explore differences in firms’ attitudes towards corporate social responsibility (CSR). Using a unique dataset covering 8,857 French firms, collected by the National Institute of Statistics and Economic Studies (INSEE), we identify firms conducting conscious CSR and others with effective but unaware CSR activities. We then construct three CSR pillar scores for each firm, using Mokken scale analysis, a form of non-parametric item response analysis. The CSR scores, along with responses to specific questions, allow us to characterize firms that implement conscious or unaware CSR. We then estimate simple probit and count data models to show that a significant share of firms are in fact actually engaged in unaware CSR, with no monotonic size effect. Cooperation with external actors such as NGOs mitigates the effect of firm size on the likelihood of conducting unaware CSR, while the effect of NGO campaigns against large firms is mainly to increase the environmental score of small firms in the same industry.

KEYWORDS:

• Corporate social responsibility
• non-parametric item response theory
• scoring
• stakeholders
• SME

JEL CLASSIFICATION:

• M 14
• Q 59
• L 29
• C 38

Disclosure statement

No potential conflict of interest was reported by the authors.

Notes

¹ Magill, Quinzii, and Rochet (2015) argue that it would not be socially optimal for small competitive firms to have the same level of CSR as large firms with significant market power. This stems from the specific way in which they define CSR in their theoretical framework: a responsible firm cares about social welfare, by weighing stakeholders’ welfare instead of only shareholders’.

² MSCI KLD Research & Analytics https://www.msci.com/esg-ratings.

³ To date, this database has been little used, except in an approach quite different from ours by Roudaut (2015).

⁴ On the contrary of this greenwashing effect, Kim and Lyon (2014) also highlighted a potential under-declaration of certain firms to avoid the backslash effect from investors fearing that CSR may be sought at the expense of profitability.

⁵ Unfortunately, we only have this data for one year. Indeed, the survey was repeated in 2016, but with a modified questionnaire and a restricted sample, excluding very small firms.

⁶ In the remainder of the article we use the sampling weights provided by the INSEE ensuring that descriptive statistics and econometric results hold for the whole population of interest (around 150,000 firms).

⁷ The questionnaire is presented in full in Appendix A.

⁸ However, this law did not come into force until 2012. And companies with more than 500 employees are all concerned just since 2017.

⁹ The questions that make up each score are those indicated by a dagger in the full questionnaire provided in Appendix A.

¹⁰ Because our data were transmitted by the companies within the framework of a survey for statistical purposes only, with all the required guarantees of anonymity, we consider it unlikely that this reflects strategic under-reporting, or brownwashing, as described by Kim and Lyon (2014) as well as by Lyon and Maxwell (2016).

¹¹ Note that for each CSR pillar, the sample restricts to firms that either engage in conscious or unaware CSR, which implies that the sample sizes (number of observations) differ.

¹² Results of the LR-tests for each of the models estimated are available from the authors upon request.

¹³ Due to statistical confidentiality constraints, we are not allowed to explicitly link the Sigwatch database to the ENDD database. For this reason, we do not attempt to explore the direct effect of NGO pressure on target companies, but focus on the potential sectoral spillover effect.

Additional information

Funding

This research has been funded by the French National Research Agency (ANR-15-CE05-0008-01) and the program Investissements d’Avenir of the French government (ANR-17-EURE-001). The authors thank Thomas Lyon, Catherine Liston-Heyes, Bernard Sinclair-Desgagné, and two ano-nymous reviewers for helpful comments and suggestions. They are very grateful to conference participants at the FAERE (Nancy), CREE (London, Ontario) and GREENGO (Paris).