Earnings management, corporate social responsibility and corporate governance in Indonesian banking industry

This paper describes the research which investigates the moderating effect of corporate governance (CG) on the relationship of earnings management (EM) practices and corporate social responsibility (CSR) disclosure in companies listed on the Indonesia Stock Exchange. EM used in this study is the different between discretionary realized security gain or loss and discretionary loan loss provition. Data is obtained by purposive sampling process with a total sample of 138 out of 46 banking sector companies. The results showed that EM had not effect on CSR and audit committee has a negative effect on CSR, but independent commissioner does not affect CSR if tested together with EM.


INTRODUCTION
Awareness about disclosure of social responsibility has become a trend for companies today. This is because companies must have responsibilities in social, economic and environmental aspects. Disclosure of social responsibility is one of the efforts made by the company to meet the interests of stakeholders and ensure the long-term sustainability of the company. The manager as the party that is given responsibility by investors, trying to improve the welfare of investors and hopes to get compensation based on earnings management. The existence of corporate social responsibility activities that must be disclosed in the annual report can lead to conflicts of interest between management and stakeholders. On the one hand, stakeholders such as investors have an interest in increasing the welfare of the profits earned by the company. Whereas for managers as financial statement makers who must realize the desires of investors and their own interests, it is possible to misuse company resources that can benefit themselves (Chih et al., 2008). The practice of perverting resources by management is often referred to as earnings management practice.
Earnings management occurs when managers change financial reporting that can mislead some stakeholders about the economic performance of the underlying company or to influence the outcome of contracts that depend on reported accounting numbers (Abbadi et al., 2016). Earnings management practices can mislead stakeholders regarding the value of assets, transactions, or the company's financial position, and this has negative consequences for shareholders, the environment in which the company is located, creditors, employees, the reputation and career security of managers and society as a whole (Zahra et al., 2005).
Previous studies have tested the relationship between the quality of financial statements specifically related to earnings management practices and CSR (Chih et al., 2008;Prior et al., 2008;Oktavia, 2013;and Grougiou, et al., 2014). These studies show different results, such as the study of Chih et al. (2008) found a negative relationship between CSR and earnings management, when earnings management was proxied by income smoothing. Whereas the research of Prior et al. (2008) showed that there was a positive influence between earnings management practices and corporate social responsibility. Although the results of these studies are different, the researchers agree that earnings management practices are considered detrimental because they can reduce the value of financial statements, provide irrelevant information for investors and have been widespread and pervasive in every financial reporting submitted by the company (Chih et al., 2008;and Kim et al., 2012).
Some researchers such as Herawaty (2008), Riwayati et al. (2015), Achyani (2015), Rice (2016), Abbadi et al. (2016) and Supardi and Asmara (2018) stated the need for integration between the role of corporategovernance and the disclosure of social responsibility. Herawaty (2008) and Riwayati et al. (2015) states that corporate governance is needed to control the behavior of company managers so that they act not only to benefit themselves, but also to benefit the company's owners. Corporate governance is a concept that is based on agency theory, carried out for the improvement of company performance through supervision or monitoring of management performance and ensuring management accountability to stakeholders by basing it on a regulatory framework (Herawaty, 2008). Jamail et al. (2008) states that corporate governance will not be effective without encouraging the disclosure of sustainable social responsibility, because companies must be able to meet the needs of stakeholders and also be able to generate profits to be able to create value for shareholders.
Based on the explanation above and the inconsistency of the results of previous studies, this study will re-examine the effect of earnings management on corporate social responsibility by referring to research conducted by Grogiou et al. (2014) of the banking industry in Indonesia. The difference between this study and that of Grogiou et al. (2014) is that this study adds the corporate governance variable as a moderating variable. The researcher argues that corporate governance can be applied to improve the effectiveness of company performance and to increase value for shareholders, so that by implementing corporate governance can minimize fraud that may occur in the company (Rice, 2016;Abbadi et al., 2016). Therefore, this study aims to examine the effect of earnings management practices on corporate social responsibility and examine the moderating role of corporate governance on the relationship between earnings management and corporate social responsibility.

METHOD
This research is a type of quantitative research to determine the relationship or correlation and explain the relationship between one variable with another variable. This study is intended to determine the correlation and influence of earnings management variables on corporate social responsibility and the moderating role of corporate governance variables. By using agency theory and legitimacy theory, this research tests the hypothesis of the relationship between these variables based on secondary data. The data is processed using multiple regression analysis, The population in this study are banking sector companies listed on the Indonesia Stock Exchange (IDX) with a research period of 2015-2017. Our sample comprises 138 firm-year observations. In this study, the measurement of earnings management uses the difference between Loan Loss Provocation (LLP) and Realize Securities Gain and Losses (RSGL) used in the study of Grougiou et al., (2014). The dependent variable used in this study is the disclosure of corporate social responsibility. In this study, measurement of corporate social responsibility disclosure uses the Coporate Social Responsibility Disclosure Index (CSRDI) based on the Global Reporting Initiatives (GRI) indicators obtained from www.globalreporting.org. This study uses content analysis data analysis methods, namely by identifying CSR practices in the banking industry based on the GRI Index which consists of 79 items related to economic performance, the environment, labor practices and decent work, human rights, social or social, and responsibility responsible for the product.
As a moderating variable, Corporate governance is one tool that can be used to maintain and restore investor confidence in the company. In this study, the proportion of independent commissioners is a measure of corporate governance mechanisms that will affect the relationship between earnings management and corporate social responsibility. In addition to the proportion of independent commissioners, the frequency of audit committee meetings is also used in measuring corporate governance mechanisms. In line with research Xie et al. (2003), Djuitaningsih and Marsyah (2012) said that the number of audit committee meetings conducted by audit committee members in one year can be seen from the number of meetings reported in the company's annual financial statements included in the report corporate governance and audit committee reports.
In this study, researchers also used two control variables, namely leverage and intangible assets. Leverage is the ability that companies have in using assets or funds that have a fixed cost (fixed cost assets or funds) to increase income (return) for company owners. Leverage is calculated by dividing the total debt to equity. While intangible assets can be measured by dividing the amount of intangible assets to total assets. As presented in table 1, the lowest value of earnings management obtained from the discretionary difference between RSGL and LLP discretionary is 0.019 while the highest value is 10.876 so the range of earnings management data is not too diffuse, so the possibility of outliers is very small. Likewise with the dependent variable corporate social responsibility which shows that corporate social responsibility data is not too diffuse and the outliers that occur are small. Whereas the moderating variable, especially the audit committee, shows that the lowest and highest value of the AUD variable obtained from the activities of the company's audit committee in this case is the number of audit committee meetings means that the company conducts at least 2 audit committee meetings in one year, whereas most companies hold audit committee meetings 22 times in one year.

RESULTS AND DISCUSSION
Pearson correlation between variables is calculated and presented in Table  2. Testing of the correlation matrix for independent variables in Table 2 shows no correlation coefficient above 0.8. This shows that there is no multicollinearity problem (Gujarati, 2005). Variance inflation factors (VIF) are also tested and the results are still within acceptable limits.  Table 2 presents the results of the bivariate correlation test in the form of correlations between research variables and it appears that the EM and CSR variables are negatively correlated. This result is an early indication that earnings management decreases the quality of corporate social responsibility disclosure. This provides an initial answer to hypothesis 1. Table 2 also shows that the EM * INDP and EM * AUD variables are negatively correlated with CSR variables. These results indicate that corporate governance practices weaken the relationship between earnings management and corporate social responsibility. Although the results presented in Table 2 have indicated the results of the relationship of each variable, these results will be tested again in regression analysis. Table 3 presents the results of a regression analysis to test the hypothesis. The regression results show that model 9 is not significant (p >0.01) and explains about 0.4% of the relationship between the dependent variables and the independent variable. While, model 10 has a statistically significant F value (p <0.01) and explains approximately 13.3% of the relationship between the dependent variables and the independent variable. Therefore, F value and adjusted R 2 of the regression,especially model 10, shows that the multiple regression model is statistically significant. Thus, the research model (10) is eligible for use in the analysis. The main variable tested in this study and to prove H1 is earnings management (EM). The results presented in table 3, especially Model 9, obtained the coefficient of EM variable of -3,542 and was not significant. This shows that the earnings management variable does not have a positive effect on corporate social responsibility disclosure by the company. Thus, hypothesis 1 which states that CG has a positive effect on CSR is rejected and is not supported by empirical data on research.

Table3.Regression
In addition to earnings management variables, this study uses variables of interest to prove H2a and H2b, namely the audit committee (AUD) and independent commissioners (INDP) who interact with corporate governance (EM*INDP and EM*AUD variables). The coefficient of the EM * INDP variable in Model 10 Table 4 presents a figure of 0.018 and is not significant. While the EM*AUD variable coefficients in Model 10 of table 4 obtained a number of -0.027 and significant at the 5% level. The results obtained from hypothesis 2a show that corporate governance measured using the proxy proportion of independent directors cannot moderate the effect of earnings management on corporate social responsibility disclosure, so hypothesis 2a is rejected and is not supported empirically. While hypothesis 2b which states that the implementation of corporate governance using a proxy number of audit committee meetings can weaken the influence of earnings management on corporate social responsibility disclosure, proven and empirically supported.

A. Relationship Analysis of Earnings Management and Corporate Social Responsibility
Testing of hypothesis 1 which states that earnings management practices have a positive effect on disclosure of corporate social responsibility is proven unsupported. The findings in this test are not consistent with the results obtained by Prior et al. (2008), Oktafia (2013 and Grogiou et al. (2014) which they found that earnings management practices can increase the extent of corporate social responsibility disclosure. However, the results of this study are the same as the research conducted by Sun et al. (2010). Researchers suspect that opportunistic managers who do earnings management will not necessarily be driven to disclose corporate social responsibility more broadly.
Researchers also suspect that the banking companies that became the study sample still have a low awareness in disclosing corporate social responsibility. This is allegedly because banking companies in conducting their business are not directly related to nature and the environment, so many of the Global Reporting Initiatives (GRI) indicators used in this study are mainly related to the environment in the Corporate Social Responsibility Disclosure Index, which cannot be maximally fulfilled, thus causing disclosure of corporate social responsibility is still relatively low and may not be used as one of the main indicators in assessing the performance of a company. In other words, the level of corporate social responsibility disclosure in the banking sector is still lower compared to companies in other industrial sectors, especially those that operate directly related to nature and the environment.

B. Analysis of the Role of Corporate Governance Moderation in the Relationship of Earnings Management and Corporate Social Responsibility
Testing hypothesis 2a shows that the proportion of independent boards of commissioners is not proven to significantly moderate the effect of earnings management on corporate social responsibility disclosure. The results of this study are consistent with research conducted by Chintya (2012), Oktafia (2013) and Sari (2014) but different from studies conducted by Sun et al. (2010). In their research, Sun et al. (2010) states that the proportion of independent commissioners has a significant effect on disclosure of corporate social responsibility. This finding indicates that changes in the composition of independent directors cannot influence managers in carrying out earnings management practices by expanding the scope of corporate social responsibility disclosure.
Unsupported hypothesis 2a indicates that there are still a number of banks in Indonesia that have a percentage of independent commissioners to all members of the board of commissioners still below the provisions made by the Financial Services Authority Number 55/POJK.03/2016 regarding the Implementation of Governance for Commercial Banks (see Table 1). In addition, the professional independence of commissioners is highly dependent on the integrity of each of them and independent commissioners have not been able to show their independence, as a result the oversight function does not go well (Chintya, 2012). Therefore, no matter how large the composition of independent commissioners, it will not affect the disclosure of corporate social responsibility (Oktafia, 2013). This further strengthens the suspicion of researchers that the supervisory function carried out by the independent board of commissioners, especially in banking companies in Indonesia has not been functioning effectively.
Hypothesis 2b test results related to the number of audit committee meetings can significantly weaken the influence of earnings management on corporate social responsibility disclosure. The results of this study are the same as those of Rodgers et al. (2007), Sun et al. (2010), and Oktafia (2013) which state that the number of audit committee meetings has a significant influence in the relationship between earnings management and corporate social responsibility disclosure. This finding shows that although the establishment of a corporate audit committee in Indonesia is mandatory, a meeting conducted by the audit committee can guarantee oversight of management and raise awareness of the company's audit committee to carry out its supervisory functions effectively to managers.
Related to the support of hypothesis 2b, the researcher suspects that the role of the audit committee in banking companies can be carried out effectively and efficiently. This is evident from the tests whose results are presented in table 1. Although the number of audit committee meetings in several banking companies is still not in accordance with the Financial Services Authority Regulation number 55 / POJK.04 / 2015 which states at least 4 times a year, but the results of the audit committee meeting can be implemented optimally in overseeing managers, especially related to CSR disclosure in the annual report. In addition, the Financial Services Authority Regulation number 55 / POJK.04 / 2015 also regulates the involvement of the audit committee in preparing annual reports, so this can minimize managers to behave opportunistically by manipulating corporate social responsibility reporting.

Conclusion
This study was conducted with the aim of examining the relationship between earnings management and corporate social responsibility disclosure and the effect of moderating corporate governance mechanisms such as the proportion of independent boards of commissioners and the activities of the audit committee on these relationships in banking companies in Indonesia during 2015-2017. The results of the analysis of research that has been done, it is found that earnings management does not significantly influence the disclosure of corporate social responsibility. While the corporate governance mechanism used in this study obtained different results, namely the proportion of independent commissioners did not significantly influence the relationship between earnings management and corporate social responsibility disclosure. For the audit committee, it was supported negatively significantly and was proven to weaken the relationship between earnings management and disclosure of corporate social responsibility, so that hypothesis 1 and hypothesis 2a were rejected, while hypothesis 2b was accepted and supported by empirical data on research.
to the environment, they tend to have low levels of corporate social responsibility disclosure and are more related to community activities, so that the Global Reporting Initiatives indicators used are not met. Therefore, a suggestion that can be submitted for further research is to conduct a similar study in the company sector other than banking that operates directly in contact with nature and the environment, so that the practice of corporate social responsibility disclosure tends to be higher. In addition, research can also be done by adding the observation time.