Title
Sustainability reporting in Brazil and Italy. Exploring distinct approaches.
Introduction
Institutional theory deals with the interaction between the organizations and the institutional dynamics, it points out the social expectations holds a main role in shaping firms’ behaviour and practices (Wright et al., 2005; Ostrom, 1991; Dillard et al., 2004). Institutionalized activities are the result of interrelated process between firms and institutional context and they are the results of organizational conforming mechanisms to existing social norms (Covaleski and Dirsmith, 1988). The institutional theory discuss about the adaptation of an institutional practice by an organization (coercive, mimetic and normative isomorphism mechanisms) and organizational practices being different from institutionalized ones (decoupling mechanisms) (DiMaggio and Powell, 1983; Dillard et al., 2004; Azlan and Roszaini, 2011).
In the field of Corporate Social Responsibility (CSR) institutional theory refers to the power social context exerts on stakeholder management and corporate social performance (Jones, 1999), to the relationship between social and environmental state-regulation and self-regulation (Neergaard and Pedersen, 2003), and it mainly discussed the national differences that affects firm approach to CSR (Maignan and Ralston, 2002; Matten and Moon, 2008). In the light of institutional theory, firms located in different countries may be recognized a different social role. A growing number of works found significant impacts of national differences on behaviours of firms referring to CSR (Habish et al., 2011; Papasolomou-Doukadis et al., 2005; Juholin, 2004; Fulop et al., 2000; Tencati et al., 2004; Abreu et al., 2005). Starting from institutional perspective, a “CSR divide” between developed and developing countries has emerged in literature (Barkemeyer, 2011) and it underlines the different impact CSR commitment determines on the surrounding conditions (Gugler and Shi, 2008; Tan and Wang, 2011). While organizations in the developed areas have been found to adopt an active approach to social or environmental oriented practices, companies in emerging nations predominantly apply reactive or defensive practices to respond to pressures (Jamali and Mirshak, 2006).
Among CSR practices, social responsibility disclosure (SRD) is particularly relevant because it expresses firms’ commitment and performance on CSR, thus making firms’ social and environmental policies more explicit (Demirag, 2005; Cooper and Owen, 2007). As stakeholder engagement holds a major role in CSR (Greenwood, 2007), firms’ commitment to disclose qualitative and quantitative information on CSR has acquired importance in the eye of stakeholders and become a more frequent and stable practice within corporate communication, since a growing number of firms has moved from the previously dominating shareholders’ point of view in recent years (Charreaux and Desbrières, 2001). The last extensive survey on SRD conducted in 2008 revealed that about 80% of companies integrate non-financial indicators in their SRD (KPMG, 2008).
Previous researches on SRD analyzed the phenomenon in terms of amount of social and environmental information (Deegan and Gordon, 1996; Gray et al., 1995; Jose and Lee, 2007); others again addressed the antecedents of publications of sustainability reports internal to the firms such as company size, company age, risk, corporate governance and values (Hackston and Milne, 1996; Deegan and Gordon, 1996; Roberts 1992; Haniffa and Cooke, 2005; Cambpbell, 2000; Walden Schwartz, 1997; Adams, 2002; Meek et al., 1995) and external, such as social, political and cultural context (Hogner, 1982; Burchell et al, 1985; Adams and Harte, 1998; Vormedal and Ruud, 2000; Haniffa and Cooke, 2005). In particular, this last one explores those factors (e.g. stakeholder pressure; industry-specific features) which are supposed to determine the level of SRD (Barako and Brown, 2008; Tagesson et al., 2009). Even if there is a large set of research on SRD, a lack of knowledge emerges regarding the differences between SRD in developed and developing counties from an institutional theory perspective.
Extant studies confirm that country context affect the disclosure of CSR related information by firms (Boesso and Kumar, 2007; Kolk, 2005; Kolk et al., 2001; Maignan and Ralston, 2002; Meek et al., 1995). For example, political, social and cultural factors determine the issues to be included in Codes of Ethics and the importance given to them (Zattoni and Cuomo, 2008; Langlois and Schlegelmilch, 1990; Schlegelmilch and Houston, 1989). Again, companies have resulted to respond to specific requests that arise in their surrounding context notwithstanding the growing relevance of international standards on CSR commitment and disclosure aimed at making firms’ behaviours converge in term of practices, models and contents and at fostering comparability and understandability (Maon et al., 2009).
The article is aimed at provide two main contributions to research. Firstly, it contributes to the extension of the literature around CSR in developing and emergent countries, which are under-studied compared to developed ones (Muller and Kolk, 2009). Most of the literature, indeed, takes into consideration western and northern countries which usually present a higher degree in social progress and development (Castelo et al., 2008; Chen and Bouvain, 2008; Perrini et al., 2007; Maignan and Ralston, 2002; Gray et al., 1995). Secondly, the paper explores the relation between institutional contexts and SRD in developed and developing countries. Giving the different social patterns and role of institutions (Puppim de Oliveira, 2002), the companies acknowledge a different role in social progress and environmental protection as the information provided in reports focus on different aspect of their CSR commitment Wanderley, 2008).
The article explores impact of institutional differences between one representative developed country (Italy) and one representative emerging economy (Brazil) through an original perspective which analyzes the social and environmental content of corporate reports – whether social or financial. We conducted a content-analysis of 100 reports published by the 50 major Italian and Brazilian listed companies in term of capitalization. The analysis is aimed at identifying differences in topics, type and amount of CSR related information disclosed by the companies.
The two countries have been chosen based on two main considerations: first, Brazil is the only Latin country among the emerging ones which are deemed to be at a similar advanced stage of development (Melé et al. 2006); second, Italy is the largest country representative of the approach devoted to through the involvement of a diverse group of social actors to the development of CSR (Albareda et al. 2007). Further, we focused our analysis on large firms for two reasons. First, they are more susceptible to scrutiny from stakeholder groups and they therefore are more sensitive to institutional pressures than smaller ones (Adams et al. 1998; Archel 2003; Neu et al. 1998; Patten 1991; Purushothaman et al. 2000). Second because, on average, they are more diversified across geographical and product markets, thus having a wider and more diverse stakeholder network (Brammer and Pavelin 2004a, p. 704).
The article is structured as follow. The first two paragraphs discuss the development of CSR in the Italian and Brazilian country context, so as to make the different approaches suggested by extant literature emerge. Next, the methodology of research is presented, with reference to sample description, measures and content-analysis. Finally, results are discussed and conclusions are drawn.
CSR and SRD in developed countries and Italy
Previous studies on CSR in European countries have shown that national governments promote CSR through a relational approach (fostering institutions, NGOs and businesses) to establish strict and collaborative relationships and promote sustainable development (Albareda et al., 2008). Hence, pressures on the context suggest that CSR polices and activities should be framed on a well-developed network of relationships among different subjects with different needs and expectations.
Literature on the development of CSR in Italy has mainly focused on the analysis of differences between SMEs and large firms, rather than on the comparison with other institutional settings (Perrini et al., 2007; Perrini, 2006a). This might be due to the fact that Italian context is characterized by prevalence of SMEs, while most of the international literature has been drawn on large and multinational companies which are expected to be more involved in social and environmental policies (Polonsky and Jevons, 2009; Russo and Tencati, 2009; Gugler and Shi, 2008;) due to a higher resource availability, a greater sensitiveness to media coverage and more need to communicate values and principles to external stakeholders (Graafland et al., 2003). Moreover, previous research have also suggested that while small and medium firms should be studied using the theory of social capital, large companies’ CSR behaviour is better explained through the adoption of the institutional theory (Perrini, 2006a; Russo and Tencati, 2009; Del Baldo, 2011).
Previous literature allow assuming that large companies in developing countries are at the core of a complex bundle of relationships and they are thus required to fairly respond to the social and environmental requests of various groups of stakeholders (Albareda et al., 2008). Therefore, the contribution to societal development should be given on a wide spectrum of issues and companies use sustainability reports to demonstrate their commitment to the growth and progress of all the subjects they enter into relationship with when realizing their operations and business activities (Habish et al. 2011).
Differently from what is going to be discussed in the following paragraph, companies in Italy do not receive institutional pressures to engage in the solution of wide-spread societal problems, which usually acknowledged to Government’s responsibility (Secchi, 2006). Instead, they receive pressures to commit to a broad and more general improvement of the impacts that might generate along their productive processes. In this sense, companies in developed countries are expected to voluntarily improve the living standards using their knowledge, capabilities and power. In the light of these considerations, we formulated the following hypothesis on the SRD in Italy:
HP1: SRD in Developed countries is aimed at creating a broad consensus among stakeholder without a focus on specific group of stakeholders or on specific social problems.
CSR and SRD in developing countries and Brazil
The privatization programs enforced in developing countries in the last decades have transformed the role of actors in the economic and social context (Damiano-Teixeira and Pompermayer, 2007). On one side governments had the prominent objective to favour the rapid economic development, therefore encouraging foreign companies and organizations to invest or move their business activity onto their territory and accepting looser regulatory regimes to allow favourable labour conditions. On the other side, private sector acknowledged the implications their activity could have on the national progress and development.
The rapid growth in the commitment to socially and environmentally responsible activities in the developing countries private sector is due to the need to support the institutional and governmental interventions in favour of societal advance (Boudon, 2002). Corporate engagement to CSR was driven by the recognition of a new role of companies in society and towards local community and the environment (Jamali and Mirshak, 2007). Therefore firms are expected to contribute to those main social problems which were of paramount importance for increasing the quality of life of the population.
Though researches which specifically explore CSR in Brazil are limited in number, at least to the extent of authors’ knowledge, this situation have been found to be common in developing countries (Arya and Zhang, 2009; Visser, 2008). Therefore a pattern for CSR development has been found which differs from the one discussed about Developed countries. Before such institutional weakness, companies tend to assume an explorative approach towards CSR, as issues to be addressed are determined by the specific condition in the local context rather than defined on the basis of the general accepted notion (Muller and Kolk, 2009; Logsdon et al., 2006; Kraisornsuthasinee and Swierczek, 2006). Hence, efforts are expected to focus on fewer precise areas or stakeholders. Although very limited studies have been conducted on the characteristics of SRD in different national and cultural settings, it is possible to assume that communication should tend to explain the positive effect determined on the quality of life and the environment, rather than demonstrating the compliance to institutional requirements and searching for a broad stakeholders’ recognition and consensus.
HP2: SRD in developing countries is aimed at describe the contribution of the firm to critical national problems rather than being oriented to establish a broad dialogue with stakeholders.
Methodology
Sample
To avoid sample bias, the large firms selection was determined by firms included in each country’s stock market index. In the research information included in the 100 selected reports from the major 50 Italian and 50 Brazilian firms per capitalization (on 12/31/2009) on the National Stock Exchange of each country was analyzed. The choice has been based on two main reasons: they are reliable representative samples of National context (Branco and Rodrigues, 2009; Habish et al., 2011); and they have a culture of higher transparency than those observed in smaller firms (Perrini, 2006b).
Reports to be content analyzed were collected on corporate websites in 2011. Since the communication of social and environmental performance is voluntary and no mandatory information are required by national laws or stock exchange rules in both countries, the annual report was included in content-analysis when sustainability report was not available. This decision was determined by the possibility that some social and environmental information could have been integrated. Only sustainability and annual reports have been considered as they represent the main communication tools on corporate performances towards stakeholders. At the end of documents collection, the final sample included 49 annual reports and 51 sustainability reports. The distribution of the documents by type and industry is reported in table I. 20 sustainability reports were gathered in Italy, while 31 in Brazil.
Breaking the data per industries, the largest part of the sample is made up by financial & real estate (30 firms) and energy, oil and gas (21) industries. Other industries are almost equally distributed.
[Table I about Here]
Measures and analysis
Type and amount of information provided in CSR reports were analyzed through the “content analysis” methodology. This methodology allows codifying content on written documents into categories, with the scope of transforming text into quantitative scales (Weber, 1988) and it has been largely used in the literature to examine SRD (Among others: Gray et al., 1995; Campbell, 2000; Deegan and Gordon, 1996; Deegan and Rankin, 1996; Holder-Webb, 2007; Milne and Adler, 1999).
A list of 36 different items grouped in 5 macro areas were defined starting from several empirical studies in the area which proved to be of great utility (see, for example, Adams et al., 1998; Gray et al., 1995; Hackston and Milne, 1996; Patten, 1991; Williams and Pei, 1999). The final list is based on a previous research by Branco and Rodrigues (2008), who developed and tested on report and website a SRD index considering items related to five categories: environmental disclosure; human resource disclosure; customer disclosure; community disclosure; supplier disclosure.
Previous content-analysis focused on the presence or absence of information (see, for example, Haniffa and Cooke, 2005; Patten, 2002; Puruoshothaman et al., 2000) or relied upon either word counts or page counts to quantify volume of SRD (Campbell, 2000; Deegan and Gordon, 1996; Deegan and Rankin, 1996; Gray et al., 1995; Milne and Adler, 1999). Recent studies suggested a new approach to content analysis which was adopted in this study. In line with Holder-Webb et al. (2009), a coding scale to represent a conceptual analysis with phrases or words as the level of analysis was adopted to avoid the difficulties due to the overestimation of the word repetitively offered. Each report in the sample was read to determine the level of coverage of a given variable. Each variable was rated using a 7-point Likert scale as suggested by previous researches (Holder-Webb et al., 2009). Each document in the sample was read to determine how intensively it covered a given variable. Each variable was coded for this intensity using a 7-point Likert scale (0 = No mention; 1 = mentioned but only in reference to another document or statement; 2 = Brief mention with little or no detail; 3 = Discussion with some detail but not extensive detail; 4 = Detailed discussion; 5 = Discussion of comprises over 50% of the document text; and 6 = Document completely dedicated to discussion). This coding scheme incorporates both coding for the existence (codes of zero versus non-zero) and the relative degree of dedication to the different information types (codes of 1–6).
Starting from the evaluation of each single item an index was calculated for each macro area. The score for each of the areas is computed as the mean of the scores of the items included in the areas, with a maximum value of 7 and a minimum value of 1. In accordance with Branco and Rodrigues (2008), items considered irrelevant for a given company or industry were excluded so as to avoid penalizing for not disclosing information. A total of five scores for the following areas were computed: Environmental disclosure, Human resource disclosure, Customer disclosure, Community disclosure, Supplier disclosure. The total disclosure score as the mean of the 36 items analyzed was eventually calculated.
Three researchers independently codified the documents. Coders were compared for their level of inter-rater reliability using correlation coefficients. All the variables showed coefficients exceeding 0.70, which is considered reliable (Frey et al., 2000; Neuendorf, 2002).
Differences in SRD between firms operating in Brazil and Italy were test through the significance of mean difference for each item considered and each score of macro areas. An analysis of the Kolmogorov-Smirnov (K-S Lilliefors) and the Shapiro-Wilk normality test suggests that variables are not distributed normally. Because the data broke the normal distribution assumption, a set of paired sample t-test and Wilcoxon-signed rank test were conducted to identify significant differences in SRD between Italy and Brazil. We tested also significance of adoption of Global d Initiative guidelines, the main format of SRD, as antecedent of ratings and the relations with disclosure indicators was not significance whit p-values less than 0.05.
Results and discussion
Table II shows the results of pair sample t-tests and Wilcoxon Test Z between Italy and Brazil. Results suggest that the area with the higher score is the environmental one both in Italy and Brazil, with a mean score of 2.08 (1.93 in Italy and 2.22 in Brazil). The issue with the lowest score is that on relationship with suppliers, with a mean score of 1.15 (1.06 in Italy and 1.23 in Brazil) and a highest value on disclosure related to codes of conducts (1.40).