Signaling Corporate Social Responsibility: Third‐Party Certification versus Brands

• Published date: 01 July 2016
• Date: 01 July 2016
• DOI: http://doi.org/10.1111/sjoe.12150

Scand. J. of Economics 118(3), 397–432, 2016

DOI: 10.1111/sjoe.12150

Signaling Corporate Social Responsibility:

Third-Party Certification versus Brands∗

Fabrice Etil´

e†

Paris School of Economics, FR-74014, Paris, France

fabrice.etile@ivry.inra.fr

Sabrina Teyssier

French National Institute for Agricultural Research (INRA), ALISS, FR-94205

Ivry-sur-Seine, France

sabrina.teyssier@ivry.inra.fr

Abstract

Corporate social responsibility (CSR) is a credence attribute of products, which can be

signaled either through a label certified by a third party, or via unsubstantiated claims

used as part of a brand-building strategy. We use an experimental posted-offer market with

sellers and buyers to compare the impact of these signaling strategies on market efficiency.

Only third-party certification gives rise to a separating equilibrium and an increase in CSR

investments. Unsubstantiated claims can generate a halo effect on consumers, whereby the

latter are nudged into paying more for the same level of CSR investments by firms.

Keywords: Claim; halo effect; label; market experiment; social preferences

JEL classification:C92; D82; L15; M14

I. Introduction

Corporate social responsibility (CSR) implies that companies use social-,

environmental-, or health-friendly technologies in ways “that go above and

beyond what companies are legally required to do” (Vogel, 2005).1Thus,

the CSR commitments of firms differentiate consumer goods by their pro-

cess attributes. As these are generally unrelated to end-use attributes,

and consumers cannot check CSR claims without paying considerable

information costs, these products fall into the category of credence goods

∗We warmly thank the referees for their comments and advice, as well as Andrew Clark,

Patricia Crifo, Thomas Lyon, St´

ephane Robin, and participants at various workshops and

conferences. We are grateful to Maxim Frolov and Marie Font for research assistance. We

acknowledge funding from the research program ANR-07-PRNA-018 ALIMINFO.

†Also affiliated with French National Institute for Agricultural Research (INRA), ALISS.

1Policy makers view CSR as a means of promoting the production of public goods with-

out affecting private profits, in a win–win scenario. See, for instance, the summary of

the European Commission’s position “What is Corporate Social Responsibility (CSR)?” at

http://ec.europa.eu/social/main.jsp?catId=331&langId=en. CSR has also become one of the

major priorities of managers in the retail and consumer goods sector (Hartmann, 2011).

CThe editors of The Scandinavian Journal of Economics 2015.

398 Signaling CSR: third-party certification versus brands

(Darby and Karni, 1973; Nelson, 1974; Roe and Sheldon, 2007). Differ-

entiation by CSR attributes might increase consumer welfare, but it also

creates an asymmetric-information problem. Labeling can be used to signal

CSR to consumers and to address this market failure.

Most CSR labels are delivered to firms after a certif ication procedure

carried out by an independent agency (a private or public third party),

guaranteeing that the production process meets a certain quality threshold.

However, many companies choose to undertake some CSR investments

without being certified, and to make CSR claims in advertising cam-

paigns or on product packaging.2This practice is often seen as one element

of a more general brand-building strategy aimed at generating consumer

awareness and loyalty (Hoeffler and Lane Keller, 2002; Blumenthal and

Bergstrom, 2003; Kay, 2006; Anisimova, 2007; Guzman and Becker-Olsen,

2010). These CSR claims might be totally or partially unsubstantiated, and

might refer to superficial or even non-existent CSR engagements; they are

imperfect quality signals.3It has been argued that their proliferation has in-

creased consumer scepticism toward CSR initiatives, especially when CSR

efforts are perceived as not being driven by other-oriented values (Bronn

and Vrioni, 2001; Ellen et al., 2006; Jahdi and Acikdilli, 2009). How-

ever, economic theory suggests that the combined effect of entrepreneurs’

social preferences, media, non-governmental organization (NGO), or state

monitoring, and reputation might provide sufficient incentives for compa-

nies to self-regulate (i.e., not to mislead consumers regarding their CSR

commitments; McCluskey, 2000; Baron, 2001, 2007; Feddersen and Gilli-

gan, 2001; Baksi and Bose, 2007; Lyon and Maxwell, 2011).4Brands, then,

play the same signaling role as certified labels.5Our research here presents

experimental evidence on third-party certification and brands as potential

remedies for information asymmetry when CSR differentiates goods; we

also ask whether third-party certification is better at increasing market

efficiency than brands alone.

2In the food sector, Max Haavelaar is one leading example of a third-party certification

agency. Cause-related marketing (i.e., the funding of non-profit organizations or social causes

by firms) is an example of a CSR practice that does not require certification. See Bronn

and Vrioni (2001) for an overview.

3This situation covers the case of “greenwashing” (i.e., the disclosure of selected positive

information about firms’ activities). Lyon and Maxwell (2011) present a formal analysis of

greenwashing, and Kim and Lyon (2011) provide empirical evidence of greenwash in a US

government program of gas reduction with the voluntary disclosure of effort by firms.

4See also the references in Dulleck and Kerschbamer (2006) and Roe and Sheldon (2007).

5The Economist claimed in an answer to the “No-Logo” movement that brands “make firms

accountable to consumers”, and “brands of the future [...] will also have to signal something

wholesome about the company behind the brand [...] social responsibility.” See “The case for

brands” (http://www.economist.com/node/771049/print) and “Who’s wearing the trousers?”

(http://www.economist.com/node/770992/print), from the print edition (September 6, 2001).

CThe editors of The Scandinavian Journal of Economics 2015.

F. Etil´

e and S. Teyssier 399

We propose an experimental posted-offer market, where subjects are

randomly assigned to the role of buyers or sellers, and trade virtual goods.

Sellers have to choose a price and a level of CSR investment that can

differentiate the quality of their offer. The quality – the actual level of

CSR investment – remains private information, and we test the impact

of four informational environments. In the baseline, there is no signaling

device other than the market price, and buyers remain unaware of the

exact quality throughout the game (the no-signaling (NS) treatment). In the

research treatments, sellers can signal quality via a label. We compare the

case where labels guarantee a minimum quality standard (MQS) certified by

a third-party (the TP treatment)6to the case where labels are not regulated

by a third party.7These labels correspond to unsubstantiated claims that

might or might not correspond to CSR investments above the MQS. We

examine two mechanisms whereby these CSR claims can be self-regulated.

In the first treatment, there is a positive probability that fraudulent claims

(investment below the MQS) are detected and reported to buyers. This

simulates the pressure from activists, NGOs, and the media on companies

(the claim treatment).8The second treatment adds reputation effects to the

claim treatment, in order to account for the inclusion of CSR into brand-

building strategies (the brand treatment).9We compare these four treatments

in terms of market efficiency by considering the market shares, prices,

and qualities of labeled and unlabeled products. Our empirical measure of

greater efficiency is the emergence of a separating equilibrium, whereby

labels signal products of quality above the MQS (Spence, 1973; Rothschild

and Stiglitz, 1976; Conrad, 2005; Besley and Ghatak, 2007). We also look

at distributional effects, from sellers’ profits, buyers’ payoffs, and CSR

investments.

Our experimental design introduces four key innovations relative to the

existing body of literature (Cason and Gangadharan, 2002; Rode et al.,

2008). First, sellers not only choose prices and whether they want to be

6Third-party regulation corresponds to official certif ications such as ISO norms or labels

administered by independent certifying agencies with strict attribution criteria (e.g., the Max

Havelaar fairtrade label).

7The word “label” is used in the three research treatments of the experiments in order

to avoid uncontrolled wording effects. It is clear, however, that each research treatment

corresponds to a specific type of signal, and is a way of representing the signaling content

of certified labels, unsubstantiated claims, and brands. In addition, the three types of signal

often coexist in real consumer markets. For clarity reasons, we do not investigate such mixed

cases.

8In the case of credence attributes, fraud and mislabeling can be uncovered by inspections

carried out by external organizations, public authorities, or competitors (Emons, 1997; Vetter

and Karantininis, 2002).

9Ben & Jerry’s for ice cream, Altera for coffee, Body Shop for cosmetics, and American

Apparel for clothes are well-known examples of brand-building strategies.

CThe editors of The Scandinavian Journal of Economics 2015.