Corps de l’article

Scarcities of resources, global warming, greenhouse effect, working conditions and treatment of employees are environmental and social issues that civil society recognizes. This has exacerbated the focus on corporate social responsibility (CSR, hereafter) in recent years among academics, consumers, governments and investors (Ditlev-Simonsen, 2011). In France, the AFEP-MEDEF code (2013) recommended the creation of a sustainable development committee to promote the integration of environmental criteria, social and governance to the company’s strategy. The law on New Economic Regulations[1] (NRE), requires companies to include the social and environmental activities in their annual reports without imposing any sanctions.

In early 2015, 47% of French companies have a performing CSR management system according to the study of Ecovadis[2] and the cross-business mediation. In comparison, the OECD average was at 40% and the BRIC Saverage at 15%.Moreover, in France, there is a growing integration of CSR in financial management. In recent years, CSR has grown substantially in France. According to Novethic, assets under management (all funds, mandates and direct portfolio management combined) raise from €3.9 billion in 2003 to €222.9 billion in 2014. In response to the ongoing growth in popularity of CSR,[3] research on this topic is gaining more attention. Several studies have investigated the nature of the relationship between CSR and financial performance (Roman et al., 1999; Margolis and Walsh, 2001; Jiao, 2010; Kim and Statman, 2012). Others have examined the relationship between corporate social responsibility and the cost of capital (El Ghoul et al, 2011; Dhaliwal et al, 2011; Goss and Roberts, 2011),Dividend Payout(Pijourlet, 2017), firm risk (Lee and faff, 2009), mergers and acquisitions (Deng et al., 2013) and corporate cash holdings (Arouri and Pijourlet, 2017; Cheung, 2016). However, despite the role of CSR, few studies make their contribution on the impact of CSR one arnings management (Chih et al. 2008; Kim et al, 2012; Choi et al., 2013; Pyo and Lee, 2013;Grougiou et al., 2014; Muttakin et al, 2015; Ester Gras-Gil et al, 2016). Earnings management is ‘’the process of taking deliberate steps within the constraints of generally accepted accounting principles to reach a desired level of reported earnings.’’ (Davidson et al. 1987: cited in Schipper 1989). Earnings management has received much more attention among investors, practitioners, regulators and scholars, especially after the collapse of several large firms in last few decades.

The study of the association between corporate earnings management practices and CSR is complex. Previous research draws on two main hypotheses. The first one is the legitimacy hypothesis (Schuman, 1995), which admits that socially responsible firms are encouraged to prove their commitment to ethical behavior based on trust and cooperation. In addition, “socially responsible” for the firm means to be ethic and transparent in its financial disclosure. The second hypothesis is the instrumentalist view which assumes the strategic use of CSR by managers (Friedman 1970, Mackey et al. 2007). This hypothesis assumes that managers can use CSR in pursuit of their own interests to hide their mismanagement (Jensen and Meckling, 1976; Williams et al. 2006).Therefore, the study of the relationship between CSR and earnings management is complex given the existence of two contradictory theoretical assumptions.

Our study contributes to the existing literature on CSR by providing further evidence on the effect of CSR on earnings quality in the French context. This study extends previous literature by examining the moderating effect of corporate governance and ownership structure on this relationship. To the best of our knowledge, this is the first paper to systematically test the relationship between CSR and earning management practices in France, which is a different institutional setting compared to the US.

Our findings show that CSR have a negative effect on earnings management practices. Our results also show that board independence and institutional investors reduce the extent of earnings management. We provide implications for academics, investors, analysts, business participants and regulators in the French context. Our results shed the light on CSR practices and help all market participants to better understand the importance of CSR in the company’s transparency processes.

This paper is organized as follows: the second section presents the theoretical framework of the relationship between corporate social responsibility and earnings management. The third section presents the sample and the methodology used, followed by the results and discussions in section 4. The last section concludes the paper.

Corporate Social Responsibility and Earnings Management: Literature and Hypotheses Development

Earnings Management

Schipper (1989) defines earnings management as a “Purposeful intervention in the external financial reporting process, with the intent of obtaining some private gains”. Earnings management occurs when managers manipulate accounting numbers. Managers are likely to engage in such practices when accounting information users are not able to detect earnings management effects. Earnings management reduce then the quality of financial statements and mislead investors about the real economic value of the firm. This leads investors to undertake non-optimal investment decisions. Kaplan (2001) consider earnings management such an unethical practice. Numerous empirical researches have focused on the factors influencing the quality of earnings during the last decades.

Corporate Social Responsibility (CSR)

According to Bowen (1953), CSR refers to “the obligations of businessmen to pursue those policies, to make those decisions, or to follow those lines of action which are desirable in terms of the objectives and values of our society”. The objectives of CSR are economic and social. Barnard (1938) argues that CSR “analyses economic, legal, moral, social and physical aspects of environment”. Friedman (1970) claims that social responsibility is likely to increase corporate profits. In other words, there could be important ethical reasons for firms to increase their economic performance.

Currently, how companies define and organize its CSR activities varies, but a common practice is to utilize the Triple Bottom Line (TBL) framework (Shnayder et al., 2015). TBL refers to three different entities, people, planet and profit, sometimes referred to as the 3P’s. The 3P’s have also been referred to social, environmental and financial dimensions of a company’s CSR performance (Slaper and Hall, 2011). These three dimensions should, if properly evaluated, cover corporate sustainability and capital growth and meet the needs of a company’s stakeholders (Dyllick and Hockerts, 2002).

CSR and Earnings Management: Theoretical Perspectives and Hypotheses

Chih et al. (2008) argue that the impact of CSR on the quality of financial reporting is at best inconclusive. Two competing views may explain the relationship between CSR and earnings management. The first one is based on the stakeholder theory (Freeman, 1984). The second is the agency theory (Friedman, 1970) and the managerial entrenchment hypothesis (Jensen and Ruback, 1983).The stakeholder theory (Carroll, 1979; Jones, 1995; Donaldson and Preston1995; and Phillips et al. 2003) supports the conflict resolution view of CSR i.e. CSR is intended to resolve conflicts of interest between stakeholders and mitigates then practices that generate a loss of confidence such as earnings management.

This perspective suggests that CSR may be an organizational device that leads to more effective use of resources (Orlitzky et al., 2003). Based on the stakeholder hypothesis, Jones (1995) documents that firms have an incentive to reveal their philanthropic and ethical behavior when they run their business with integrity. Such firms are more likely to engage in CSR initiatives and to provide reliable financial statements. In this regard, providing quality earnings is closely connected to CSR, especially to meet the needs of stakeholders (Choi et al., 2013).

Atkins (2006) claims that “being socially responsible” for the firm is “being transparent” in its financial statements by disclosing reliable and relevant information. The firm that makes an effort in implementing CSR and resources is able to constrain earnings management in order to comply with stakeholder’s ethical expectations. Kim et al.(2012) approve that socially responsible firms incur less earnings management manipulations by being more responsible in preparing financial statements. Indeed, ethics are motivating managers to be honest and trustworthy since this behavior is benefic for the firm.

Furthermore, according to Grow et al. (2005), CSR is a form of construction and retention of the reputation of the firm by ensuring confidence and support for different stakeholders. Hence, the firm’s reputation perspective shows that the desire to keep and protect firm reputation leads managers to avoid socially irresponsible activities and to constrain earnings management practices.

Numerous empirical studies support the perspective of stakeholder and corporate reputation. Chih et al. (2008) examine CSR and earnings management of 1.653 corporations in 46 countries. They study three kinds of earnings management: earnings smoothing, earnings aggressiveness, and earnings losses and decreases avoidance. They show that CSR firms are more aggressive in accruals management but are less likely to engage in earnings smoothing and earnings losses avoidance. Hong and Andersen (2011), for a sample of non-financial U.S. firms between 1995-2005, find evidence that socially responsible firms have higher quality accruals and less activity-based earnings management. Consistent with this argument, Cho et al. (2013) document empirical evidence of a negative association between CSR engagement and the level of earnings management using a sample of Korean firms from 2002 to 2008. However, the relationship is weaker for chaebol[4] firms and firms with highly concentrated ownership, which suggests that CSR can be abusively used by those firms to conceal their poor earnings quality. Kim et al. (2012) and Pyo and Lee (2013) find similar results in the American and South-Korean contexts, respectively. They find that socially responsible firms are less likely to manage earnings through discretionary accruals and to manipulate real operating activities.

However, the agency view supports a positive relationship between CSR and earnings management. The agency theory (Friedman, 1970; Jensen and Meckling, 1976) is based on the strategic use of CSR. Managers use CSR to serve their personal interests. They use, strategically, CSR to mislead stakeholders of the real value of the firm and its real performance. This confirms a positive relation between CSR and earnings management (Mackey et al. 2007).

Indeed, managers may launch societal responsible activities to obtain more media coverage, guarantee the legitimacy of the whole community and to limit detailed verification by investors and employees. According to Hemingway and Maclagan (2004), managers can take advantage of CSR to cover up their misconduct. Therefore, they seek to persuade stakeholders that the company is transparent. It is a form of reputation insurance, giving them a “license to operate” with respect to earnings management.

Mcwilliams et al., (2006) argue that managers can use CSR to advance their careers or personal agendas as CSR represents an indirect management advantage. In the same vein, Fritzsche (1991) assumes that ethical codes can be used by managers as a means of facilitating the pursuit of economic egoism and self-interest.

In addition, basing on the entrenchment theory (Shleifer and Vishny, 1989), CSR can be used by managers as an entrenchment tool. They seek to project a socially friendly image to benefit from the legitimacy of the entire financial community to draw attention to the reputation of the firm and thus divert stakeholders’ scrutiny to detect earnings manipulations.

Empirically, Prior et al. (2008) report a positive relation between earnings management and CSR by using archival data from a multinational panel sample of 593 firms from 26 countries between 2002 and 2004. They show that managers that have incentives to manage earnings will be very proactive in boosting their public exposure through CSR. Alternatively, firms with low levels of earnings management have fewer incentives to seek public exposure by promoting socially responsible activities.

According to this perspective, Cespa and Cestone (2007) consider that CSR can be used as an entrenchment mechanism by managers with the purpose of reducing the likelihood of being scrutinized by stakeholders in the context of earnings manipulation.

The preceding discussion shows that the relationship between CSR and earnings management is ambiguous and remains an empirical issue. We formulate then a non-directional hypothesis as follows:

H1. There is a relationship between CSR and earnings management.

The Moderating Role of Corporate Governance and Ownership

Corporate governance of French firms has undergone profound transformation (Law NRE 2001, Viennot Report, 1995, 1998, Bouton Report, 2002, AFEP-MEDEF). Good corporate governance has become an important value driver, enhancing ther eputation of a country or economic region among its financial and industrial partners. The term of Good Corporate Governance term was first introduced by the Cadbury Committee in 1992 in a report known as the Cadbury Report. The Cadbury Committee based on the stakeholder theory provides this definition of Good Corporate Governance, “A set of rules that define the relationship between shareholders, managers, creditors, government, employees, internal and external stakeholders to review their rights and responsibilities”.

In France, Afep-Medef Code (2016) plays a crucial role in the development of good governance practices. It provides a reference contributing to the improvement of the governance of listed companies and the widespread establishment of best practices. According to this code, the transparency of the information provided is one of the fundamentals of good corporate governance.

We examine here the impact of CSR on earnings management practices in presence of good corporate governance practices.

From the one hand, the over-investment hypothesis stipulates that managers try to overinvest in CSR to build their own reputations as good global citizens at the expense of other stakeholders. They can use CSR as an entrenchment tool (Cespa and Cestone, 2007).However, a good corporate governance system through different internal and external mechanisms should inhibit the manager’s ability to overinvest in CSR (Gompers et al. 2003). In the lack of effective control system, such an unethical behaviour takes place. More importantly, if the overinvestment hypothesis is confirmed, we suppose that corporate governance system is inefficient and managers overinvest in CSR and manipulate accounting numbers. We then expect a negative effect of CSR and earnings management in good corporate governance context.

From the other hand, according to the second hypothesis of conflict resolution, managers use CSR to resolve conflicts of interests between stakeholders in order to maximize shareholders wealth. A better quality of governance mechanisms increases the efficiency of the supervision of management, and consequently limits the discretionary behaviour regarding earnings management (Xue and Hong, 2016). Chi-Keung (2013) argues that corporate governance can reduce the extent of earnings management practices. Corporate governance would reduce agency problems between financial providers and managers and increase the efficiency of contracts (Gompers, Ishii, and Metrick, 2003). Then, we assume that CSR is negatively associated with earnings management in presence of good governance mechanisms.

Furthermore, until relatively recently, the ownership structure of French firms remains concentrated with a high presence of institutional investors (Nguyen, 2011). According to Madhani (‎2016), ownership concentration is a corporate governance device that allows large shareholders to control managers and protect their interests. In addition, institutional investors are main actors on governance structures which promote the good corporate governance (McNulty and Nordberg, 2016). They may encourage managers to take better accounting and financial decisions. They are able to reduce agency conflicts between stakeholders, re-establish trust, guarantee information transparency and contribute to good corporate governance, particularly, in a civil low country, where shareholders’ rights are poorly protected, as is the case in France.

If this framework is supported, we expect that good corporate governance system, together with CSR engagement, will contribute to reduce earnings management practices.

The preceding arguments lead to the following hypothesis:

H2: Under a good corporate governance structure, there is a negative relationship between CSR and earnings management.

Sample and Methodology

Data

Our purpose is to examine the relationship between corporate social responsibility and earnings management in the French context. In addition, we investigate the moderating effect of good corporate governance on the relationship between earnings management and CSR.

Our sample includes French firms listed on the CAC All-Tradable from 2010 to 2013.The CAC All-Tradable contains all the stocks of the Euronext Paris market that have an annual Free Float Velocity over 20%. We remove financial companies because of their atypical behaviour in financial reporting. Our final sample includes 101 companies over for years. Financial data were gathered from the Thomson One Banker database. Social responsibility information were collected from CSR Hub data base[5].

Variable Measurements

Discretionary Accurals

We measure earnings management by discretionary accruals estimated using the models of Dechow et al. (1995) and Kothari et al. (2005) to enhance the robustness of our results

  • Jones-modified model (1995)

With: TAi,t: Total accrual in year t; Ai,t- 1: Total assets in year t-1; _CAi,t: Change in sales; _CCRi,t: Change in receivables; PPEi,t: Gross proprety plant and equipement; εi,t: Residuals that represent the estimation of discretionary accruals.

  • Kothari et al. model (2005)

With: ROA is Return on assets

Corporate Social Responsibility Scores

We relied on ratings provided by CSRHub that provide ESG scores of 15,127 companies from 130 countries. CSRHubis the largest and comprehensive database of social data information. Data sources include research firms on socially responsible investment namely Asset4 / Thomson Reuters, Carbon Disclosure Project (CDP), EIRIS, Governance Metrics International /Corporate Library, IW Financial, MSCI (RiskMetrics IVA and Impact Monitor), RepRisk Trucost and Vigeo.

Regarding the scoring methodology, Hub divided CSR into four categories: community, employees, environment and governance. Eachc ategory includes four subcategories. Then, each collected information from different data sources is included in one or more sub category. Each data is converted into a score from 0 to 100. Corporate ratings for which there is not enough information is excluded.

Control Variables 

Firm Size: The size of the firm is vital in assessing the level of earnings management practices. Lobo and Zhou (2006) argue that large firms have incentives to increase the value of their earnings, because of the complexity of their business activities. Nevertheless, Jiang, Lee and Anandarajan (2008) find that large firms are associated with lower discretionary accruals. Firm size is measured as the natural logarithm of total assets. We do not expect the sign of the variable coefficient.

Firm performance: Firm performance is an important variable in the context earnings management. Haw et al. (2004) show a positive relationship between earnings management and firm performance. DeFond and Park (1997) show that managers save income for future periods (through negative discretionary accruals). Profitability is measured by the return on assets ratio. We expect the relationship between earnings management and firm performance to be positive.

Leverage: According to Jensen and Meckling (1976), debt plays a disciplinary role to address the discretionary behavior of managers. However, according to DeFond and Jiambalvo (1994), firms with high leverage are inclined to increase their accruals in order to avoid debt covenant violation. Leverage is the book value of non-equity liabilities divided by the book value of total assets. We do not anticipate the sign of the coefficient for this variable.

Board size: Ching et al. (2006) find that firms with larger boards are engaged in earnings management practices, consistent with Jensen’s (1993) view. Xie et al. (2003) report the opposite. For a sample of French firms, Jeanjean (2002) find no significant relationship between board size and earnings management. As the relationship between earnings management and board size is ambiguous, we do not expect the direction of the relationship.

Board independence: Agency theory argues that independent directors are required to provide effective monitoring of corporate boards. In a pioneer study, Klein (2002) finds a negative association between abnormal accruals and board independence in the US. However, some studies (Larcker et al., 2004; Bradbury et al. 2006) show that the presence of outside directors has no effect on earnings management.

CEO duality: Brickley et al. (1997) show that the combination of the CEO and the chairman positions has real benefits for shareholders. However, Peasnell et al. (2005) and Klein (2002) find a positive relationship between the violation of generally recognized accounting principles and CEO duality. Although the new conception of corporate governance adopted in 1999 in France leaves the choice to the board between the two forms monistic and dualistic.

Ownership concentration: Where shareholders have a low stake in a firm, they have little or no incentive to monitor managers, because the monitoring cost will exceed the benefits of monitoring managers (Ramsay and Blair, 1993; Hart 1995). In the American context, Chtourou (2000) show that there is a negative relationship between earnings management and the cumulative percentage of blocks of shares held by investors holding more than 5%. Lopez Iturriaga and Hoffmann (2005) argue that the level of earnings management decreases when ownership is concentrated. Therefore, we expect that ownership concentration reduces the extent of earnings management.

Institutional ownership: Koh (2005) examined the relationship between income smoothing and institutional ownership. Results depicted positive relationship of institutional ownership with firms earnings smoothing by the companies. Emamgholipoura et al. (2013) show that the percentage of shares held by institutional investors is likely to increase earnings management practices. However, Jiambalvo et al. (1999) show that managers do not manipulate their earnings, given the pressure from institutional investors who are more interested in the long-term profitability.

Methodology

The multiple regression methodology with panel data is used. Panel data analyses include two special dimensions: an individual dimension, as indicated by the iindex, standing for the company, and a t index standing for the period dimension (Gujarati, 2004). Our data have a double dimension. In fact, we collect data every year for every company. The Hausman test is used to choose between fixed effect and random effect models. The Hausman test compares the variance-covariance matrix of the two estimators:

With: βf, fixed effects estimator and βa, random effects estimator.

The results of the Hausman test, not reported here, show that the fixed effect model is preferable to the random effect. This decision is predicated on the fact that the asymptotic significance is below the 5% level. We estimate the following models:

With: DA: discretionary accruals estimated using two models: Jones-modified model (1995) and Kothari et al. (2005).

CSR: the total points from the index published for each company for every year

To test the moderating effect of corporate governance attributes on the relationship between CSR and earnings management, we introduce interaction variables in our models as follows: (DA)it = α0 + α1(CSR)it + α3(Bsize)it + α2(Duality)it + α3(Duality*CSR)it + α4(Indep)it + α5(Indep*CSR)it + α6(Conc)it + α7(Conc*CSR)it + α8(IINST)it + α9(IINST*CSR)it + α10(Lev)it + α11(ROA)it + α12(Fsize)it + εit

With:

Duality: a dummy variable that equals to 1 if the chair is also the CEO of the firm and to 0 otherwise

Bsize: total number of board members

Indep: the proportion of independent directors

Conc: the Square of shares held by each shareholder

IINST: the percentage of shares held by institutional investors

Lev: the ratio of total debt to total assets

ROA: the ratio of net income to total assets

Fsize: the natural logarithm of total assets

Results and Discussion 

Descriptive Analysis

Table 2 provides the descriptive statistics for the variables used in this study. The dependent variable, measured by the absolute value of discretionary accruals displays a level of 0.908 using the Jones-modified model (1995). This is comparable to the level of 0.503 obtained using Kothari et al. (2005) model.

The average CSR score is 58.7. This score ranges in a scale between 0 and 100. By industry, the highest rating is 75 in the tourism sector. The lowest score (28) is recorded for the food-processing industry, the pharmacy, the environment and the medical biology.

The percentage of independent directors is on average 43.4%. We notice, a disparity between companies with a percentage of independent members ranging from zero to 78%. Sampled companies have a board of directors with around 12 members on average. The CEO is also the chair of the board of directors in 69.2% of cases.

The percentage of shares held by majority shareholders is 38.2%. This shows that the French listed companies have a relatively concentrated ownership structure. Finally, the percentage of the institutional investors in the capital of the French firms is on average 39.8%. This type of investors is indeed in very strong expansion in France.

Table 3 presents the correlation matrix between our independent and control variables in order to assess any potential multicollinearity problem betwee nany independent variables, which might cause estimated coefficient instability and increase standard deviation. According to Gujarati (2004), such a problem might occur when the variable correlation exceeds 0.80.Results have shown that correlations between explanatory variables are quiet low. A second multicollinearity measure has therefore beenused (the VIF: Variance Inflation Factor). The VIF values range from 1.62 to 2.91, far below 10 which is the critical value as defined by Neter, Wasserman and Kunter (1989).

Table 1

Variables definitions

Variables definitions

-> Voir la liste des tableaux

Table 2

Descriptive statistics

Descriptive statistics

-> Voir la liste des tableaux

Multivariate Analysis

Table 4 presents the results of multivariate regression analyses of discretionary accruals. In the first model the discretionary accruals are measured by the modified-Jones model (1995). We find that the estimated coefficient of CSR score is negative and significant at the 10% threshold. For accruals estimated by the model of Kothari and al (2005), we find similar results as those found by the first model. This confirms that the degree of CSR constrains earnings management practices.

The discretionary accruals are substantially lower for more socially responsible firms. These results provide strong evidence for the effect of CSR in reducing earnings management by French firms. Our findings provide support for hypothesis according to which there is a negative relationship between the level of CSR and discretionary accruals. Firms engaging in CSRare less inclined to manipulate their earnings and are more transparent in their reporting practices. These results support the stakeholder theory,the legitimacy theory and the corporate reputational view. The negative relationship between CSR and earnings management is consistent with the findings of Chih et al. (2008) and Gras-Gil et al. (2016), who argue that companies with greater CSR incur less earnings management.

This finding suggests that French firms that have provided the effort to implement CSR are able to constrain earnings management to comply with the ethical requirements and to satisfy stakeholders’ interests. This result is in line with the reputation hypothesis suggesting that firms have incentives to act socially responsible. In France, these firms are under high shareholder and press scrutiny and, hence, are more concerned about their image and their sound financial reporting.

As for control variables, we find that large boards are likely to limit the managerial opportunistic behaviour. Indeed, small boards seem more prone to failure to detect earnings management. Thus, a larger size of board assumes a better supervision of the management team and a higher quality of financial reporting. Frias-Aceituno et al. (2012) show that the complexity of managerial monitoring requires the presence of a large number of directors with different skills to effectively control managers. The results in Table 3 also show that companies with an independent board are less inclined to engage in earnings management practices. This relationship is negative and significant at the 5% level suggesting that independent directors enhance the monitoring of the board and reduce managerial discretion. Regarding ownership characteristics, the percentage of shares rights held by the first shareholder has a significant negative impact on earnings management. This result corroborates the efficient monitoring hypothesis which suggests that large shareholders reduce the scope of managerial opportunism. This finding is consistent with Wang (2006).

We find that there is a negative and statistically significant relationship between the percentage of shares held by institutional investors and earnings management, which suggests that these investors constrain incentives of managers to engage in earnings management. Institutional investors are considered as better controllers and a major corporate governance device. These results are similar to those found by Cornett et al. (2008).Among firm characteristics, firm size positively influences discretionary accruals because large firms face more pressure from investors and financial analysts to show positive earning or increase in earnings. The relationship between firm performance and earnings management is negative not as expected and not significant using two models to estimate discretionary accruals. This relationship is not consistent with the results of Haw et al. (2004). We also find a negative relationship but not significant between leverage and earnings management suggesting that highly indebted firms may be less able to practice earnings management because they are under close scrutiny of lenders.

Table 3

Correlation Matrix

Correlation Matrix

-> Voir la liste des tableaux

Table 4

Panel regression

Panel regression

-> Voir la liste des tableaux

Table 5 shows the results of the moderating effect of corporate governance attributes on the relationship between CSR and earnings management. We note in Table 4 that the results obtained in the previous analysis remain unchanged

The effect of the interaction between CSR and board size (CSR * Bsize) on earnings management is positive and significant at the 1% level suggesting that large boards do not control the efficient use of CSR by managers leading to increase the earnings management propensity. This result suggests that a small board seems desirable to control managerial behavior and ensure strategic decisions such as the decision to invest in CSR to enhance a company’s quality of earnings. Table 4 also shows that the interaction term between CSR and board independence negatively influences earnings management at the 1% level. This means that the monitoring role of independent directors on CSR is effective and is likely to have a negative effect on earnings management practices. This finding suggests that the effect of CSR on earnings management is exacerbated in presence of more independent boards.

Regarding ownership characteristics, the results show that the ownership concentration has a significant negative impact on earnings management. This result is not confirmed for the interaction terms of CSR and the ownership concentration on earnings management because the interaction coefficient (CSR* Conc) is positive and significant at the 5%. This result suggests that large shareholders may exert pressure on managers by maximizing her own utility function. In this case, the CSR is used in pursuit of interests and to hide mismanagement of large shareholders.

We find also that there is a negative and statistically significant relationship between the percentage of shares held by institutional investors and earnings management, suggesting that these investors may decrease incentives of managers to engage in earnings management. This result is also confirmed for the interaction term between CSR and institutional ownership. Taken together, these results prove the negative effect of CSR on earnings management for firms highly held by institutional investors. Institutional investors control strategic decisions and investment in CSR so that these activities will be a form of construction and retention of the reputation of the firm by ensuring the confidence and the support of different stakeholders.

Table 5

Moderating role of corporate governance

Moderating role of corporate governance

-> Voir la liste des tableaux

Conclusion 

Corporate social responsibility has become an issue in the economic, political and social life. The first response of many companies consists of an extension of the communication policy about their responsibility. In this context, the present study investigates whether CSR has an impact on the earnings management within the French institutional setting and examines the moderating effect of corporate governance devices on this relationship.

Using a sample of 101 companies between 2010 and 2013, we find that there is a negative relationship between CSR and earnings management. Our results support the stakeholder theory and the reputational view arguing to that firms have incentives to act socially responsible to enhance their reputation. This finding suggests that French companies that implement CSR activities are more likely to constrain earnings management practices to comply with the ethical requirements and to satisfy stakeholders’ interests. In France, these firms are under shareholders, and press scrutiny.

The results of the moderating effect of corporate governance attributes on the relationship between CSR and earnings management show that CSR activities reduce earnings management particularly, in small and highly independent boards. We find also that institutional investors control strategic decisions and investment in CSR to mitigate earnings management. This helps ensuring the confidence and the support of different stakeholders.

Our study contributes to the literature by investigating the concept of “ethics”, which changes the behavior of leaders about earnings management practices. CSR encourages managers to be responsible by disclosing relevant and reliable information. Furthermore, our results help all market participants to better understand the role of CSR in the company’s transparency processes. This is why, French authorities should support the development of initiatives on CSR activities. Some guidelines can be introduced by French policymakers to emphasize that CSR is based on actual practice and not just a “green wash” statement to deceive stakeholders. In this context, the need for a more regulatory system of social responsibility for listed companies is crucial.

Future research may then focus on alternative corporate governance features likely to control the use of CSR activities and hence to negatively influence earnings management.