CEO CHARACTERISTICS AND EARNINGS MANAGEMENT: EVIDENCE FROM INDONESIA

The CEO has an important role as a decision maker in the company and is responsible for the company's performance. This study was conducted to measure the effect of CEO on earnings management in 495 samples of non-financial companies listed on the Indonesia Stock Exchange in 2017-2019 using panel regression analysis with classical assumption testing. The results of panel regression examiners show that there is no significant effect between female CEOs and CEO turnover on earnings management proxies with discretionary accruals. Meanwhile, CEO tenure and company size show a significant negative relationship with earnings management. On the other hand, leverage and return on assets have a significant positive effect on earnings management. In addition, there is no significant influence between company age and Market to Book ratio on earnings management.


INTRODUCTION
CEO (Chief Executive Officer) is the "center" of policy in the company. Researchers, analysts, and the business media indirectly agree to view CEOs as the person who responsible and accountable for policies in a company (Dalton & Kesner, 1995). In Indonesia, CEO is a part of the Board of Directors who is assigned the task of the shareholders to lead and determine the company's strategy1. The appointed CEO can control the authority over company decisions and be responsible for performance within the company (Chou & Chan, 2018). However, the pressure from shareholders can encourage CEO to take opportunistic actions and impair the stock holder in a company. In Indonesia, cases related to manipulation in the presentation of financial statements have also caught the public's attention, one of which is the presentation of the 2018 financial statements of Garuda Indonesia which violated PSAK and regulations set by the OJK, so that on June 28, 2019 the Ministry of Finance finally sanctioned the Board of Directors, commissioners, and issuers (DetikFinance.com, Laucereno, 2019).
A CEO is selected and appointed in a different way from an ordinary employee, where the succession of members of the Board of Directors is carried out at the GMS (General Meeting of Shareholders), then the selection of a new CEO will consider the various interests of each shareholder by looking at the track record of the candidate. CEOs who work in the labor market will increase the incentive for them to demonstrate their abilities to market participants and avoid the penalties given. This situation can increase motivation for the CEO to sacrifice shareholder interests and defend his own interests (Jensen & Mecling, 1976). So that, CEO often practice earnings management (Baker, Collins, & Reitenga, 2009 Secretariat. Jakarta. Article 1 Number 5 explains that the Board of Directors has full responsibility to carry out the management of the company and achieve the company's goals. Bouaziz et al. (2020) state that the characteristics of CEO and earnings management are topics that need to be investigated further, this is due to the authority and controlling value that the CEO has in making decisions within the company which can affect the quality of financial reporting. In accordance with the Upper echelons theory which states that the characteristics of CEO can affect the value of the company, the selection of the company's strategy, and the company's financial reporting decisions (Hambrick & Mason, 1984). This is because the CEO's personality and experiences influence policy making on the interpretation of the situation they face (Hambrick, 2007). Research on the characteristics of the CEO, especially the gender of, the tenure, and turnover of the CEO can minimize earnings management and has become the subject of various studies with various results, but the concern is the lack of research related to the effect of CEO gender on earnings management (Alqatamin, Aribi). , & Arun, 2017). Therefore, in 2017, Alqatamin et al. conducted a study on how the influence of CEO characteristics on earnings management by considering the gender of the CEO at 1206 companies in Jordan. The result of this research is that the proportion of female CEO who lead a company does not affect the tendency of earnings management practices. CEO performance can be reflected in the performance of the company he leads, especially the amount of profit reported by the company, so this condition will affect the incentives that the CEO will receive in the future (Gibbons & Murphy, 1992). The CEO will tend to overstate profits in the first year, so as if he shows his ability to manage the company and indirectly increases the CEO's career potential and compensation in the future, and the result is the CEO can avoid dismissal when shareholders are still not confident in his abilities (Ali & Zhang, 2015). On the other hand, there are research results that show a non-linear pattern, namely an inverted U pattern between the tenure of the CEO and earnings management. Related to this, the CEO is being conservative in managing revenue while managing the company in the initial period, then the CEO to be aggressive in managing income after working a few years later, so that the earnings management conducted by the CEO reach the maximum level in fifth and sixth years. When the CEO has reached the maximum point of earnings management, the CEO will return to be conservative and less aggressive in reporting the company's earnings (Hu, Hao, Liu, & Yao, 2015).
In addition, controversy in research related to the influence of certain characteristics in CEO also occurs on the relationship between CEO turnover and earnings management. CEO turnover often have relevance to the poor performance of a company, as a result the new CEO tends to do " earnings bath ", which reduces corporate profits reported in the early period of his tenure with the aim of reducing the benchmark performance of these CEO (Bornemann, Kick, Pfingsten, & Schertler, 2015). On the other hand, the research results of Murphy and Zimmerman (1993) did not find the effect of CEO turnover on earnings management practices.
Therefore, the researcher wants to focus on the gender proportion, tenure and turnover of the CEO in non-financial companies listed on the Indonesia Stock Exchange (IDX) in 2017-2019. The results of this study are expected to expand previous research and assist companies in considering CEO characteristics, so as to reduce earnings management actions that have a negative impact on stakeholders in the company.

REVIEW OF RELATED LITERATURE
1. CEO Gender and Earnings management CEO gender will influence earnings management in a company. After the global financial crisis in 2008, gender diversity in key positions in companies became the public's attention (Lakhal, Nekhili, & Zouari, 2015). In a study conducted by Krishnan and Park (2005) by comparing companies run by male CEOs and those run by female CEOs, the results showed that male CEOs tend to manage companies properly like female CEOs. However, based on several research results that have been done, it shows that female CEOs tend to be more conservative and risk-averse, by avoiding risky investment and funding opportunities (Faccio, Marchica, & Mura, 2016;Schubert, 2006). Gavious et al. (2012) also found that female CEOs are more likely to behave ethically, so that the presence of female CEOs can increase company value. Not only that, it was found that women also have a significant impact on corporate financial reporting, because earnings management performed by female CEOs is significantly lower than male CEOs. In addition, in gender theory it is said that women are more unbiased and conservative than men, so it can be concluded that gender in this case p-1412-3789 www.journalmabis.org e-2477-1783 54 refers to gender, namely men and women, have different characteristics thus they can have different influences in a company, because there are basic characters' difference between men and women. Therefore, female CEOs tend to make more ethical decisions and as much as possible avoid managing accounting results.
Furthermore, Setyaningrum et al. (2019) conducted a study in Indonesia that examined the influence of female executives on earnings management using data from companies listed on the Indonesia Stock Exchange (BEI) from 2013 to 2017. The results of this study found that female CEOs had a significant negative effect on earnings management, then a woman can reduce earnings management behavior in the company. This is because women are more careful, avoid risk and have higher ethical standards, consequently they can reduce motivation to carry out earnings management (Setyaningrum et al, 2019). In opposite, the effect of CEO gender on earnings management in companies in France was examined by Bouaziz et al. (Bouaziz et al., 2020) and the results showed that there was a negative and insignificant relationship between female CEOs and earnings management. In addition, based on research conducted by Setyaningrum et al. (Setyaningrum et al., 2019), the researchers want to do research that is more focused on the gender of CEOs whether it has an impact on earnings management, if it is examined more thoroughly in all sectors (except finance) in companies listed on the IDX in the years 2017-2019. Therefore, the hypothesis is formulated as follows: H1: The presence of female CEOs negatively affects earnings management practices.

CEO Tenure And Earnings management
CEO tenure will affect earnings management in the company. The tenure of the CEO is stated in POJK No. 33 / POJK.04 / 2014, Article 3 paragraph (2) and (3), which states that members of the Board of Directors are appointed for a certain term, where one term of office of members of the Board of Directors is no longer than 5 (five). years or until the closing of the annual GMS at the end of the 1 (one) term of office concerned. In addition, Article 40 of 2005 concerning Limited Liability Companies Article 94 paragraph (3) states that members of the Board of Directors are appointed for a certain period of time and can be reappointed.
Along with the end of the CEO's tenure, it makes his career time horizon shorter. This causes the CEO to prefer a combination of explicit incentives (such as salary, pension package, last bonus, etc.) compared to implicit incentives (such as: the possibility of a post-retirement appointment) (Gibbons & Murphy, 1992). This shift in preference is caused by the CEO's desire to realize his incentives and be explicitly useful for his consumption after the end of his tenure As a form of bonding cost, shareholders use a compensation package that refers to the company's performance (Healy, 1985). Dechow and Skinner (2000) stated that the compensation package provided motivates managers to practice earnings management. Related research has also documented that compensation sensitivity to performance influences higher managers' earnings management practices (Healy, 1985). This research views profit as still one of the performance indicators that is often used as a measure of compensation, regardless of its short-term period. In addition to maximizing compensation payments, managers perform earnings management at the end of the period to increase implicit compensation also, such as the possibility of p-1412-3789 www.journalmabis.org e-2477-1783 55 holding a post-retirement position (Gibbons & Murphy, 1992). By performing well in the final year of his tenure, the CEO will increase the likelihood of a higher career after stepping down. Thus, motivated by two incentives simultaneously, a CEO will try to increase the market value of his managerial labor at the end of his tenure. One of them is by implementing earnings management practices so that the last reported earnings before the end of their working period look good in the eyes of shareholders. Davidson et al. (2007) found that CEOs who are approaching retirement tend to do earnings management. This aims to obtain incentives in the form of bonuses, pension packages that can be used for consumption after the CEO's tenure ends (Davidson et al, 2007). Based on previous research, it was found that there was a positive influence between the tenure of the CEO and earnings management as measured by discretionary accruals. Therefore, the hypothesis formulated is as follows: H2: CEO tenure positively influences management practice profit.

CEO Turnover and Earnings management
CEO turnover is one of the best ways to improve the performance of a company that is experiencing a decline. Choi et al. (2014) found that CEO turnover associated with poor company performance tends to affect earnings management. In these circumstances, the CEO tends to perform income increasing patterned earnings management to increase the nominal profit reported by the company with the aim of covering up the company's poor performance. In addition, Bornemann et al. (Bornemann et al., 2015) argued that the change of CEO resulted in a replacement CEO who had just served in a company undertaking an "earnings bath" pattern of earnings management, so the amount of discretionary expense in the first year tended to be higher than in subsequent years. This aims to blame the previous CEO's performance which is considered bad so that it can lower the CEO's performance benchmarks and also reduce the profit target that must be achieved for the next period.
In a study conducted by Wells (2002), it was found that non-routine CEO turnover caused by dismissal from office due to poor performance would be positively related to earnings management. This is because forced CEO turnover has implications for accounting policies within a company, including CEO incentives to manage accounting results (Wells, 2002). Based on the results of previous research, it was found that there was a positive influence between CEO turnover and earnings management as measured by discretionary accruals. Therefore, the hypothesis formulated is as follows: H3: CEO turnover positively affects earnings management practices.

RESEARCH METHOD
The research sample was all companies, except the financial sector, which were listed on the Indonesia Stock Exchange (IDX) from 2017 to 2019 with the following sampling criteria: (1) the company has a complete financial report and annual report during the research period, (2) the company uses rupiah currency. Based on these criteria, a sample of 744 companies was obtained.
Research Model: The dependent variable used was profit quality as measured by earnings management (DA). This study used discretionary accruals using the Raman and Shahrur (2008) model. The choice of this model was because the Raman and Shahrur models (Raman & Husayn, 2008) were an improvement from Kothari et al. (2005). Model modification was done by adding Book to Market Ratio to calculate total accruals. The absolute value of discretionary accruals was calculated by deducting total accruals from non-discretionary accruals. The model is written in the following notation: The CEO tenure variable (ACD) was measured by the number of years the CEO had held positions in a company. For example, if in a year there is a change of CEO and the effective date for the appointment of a new CEO is before the middle of the period (before 30 June), then that year is considered as the first year of the new CEO. Information regarding the effective date of the appointment of the CEO or the date of the GMS which decides the appointment of the CEO was obtained through the company's annual report. The CEO gender variable (GEN) measured using a dummy variable (Arun et al, 2015), ' 1 ' if the CEO concerned is female, '0' if the CEO is otherwise. The CEO turnover variable (TURN) is measured by a dummy variable (Cooper, 2017), ' 1 ' if the company changes its CEO in the previous year, ' 0 ' otherwise. The control variables used are company size (SIZE), company age (AGE), leverage (LEV), return on assets (ROA), and market to book (MTB). Table 1 shows the descriptive statistics of the study, with the number of research is 744 companies. In this study, the dependent variable used is discretionary accruals (DA). In the results of descriptive statistical calculations in the table above, the minimum and maximum values for DA are 0.0000667 and 0.615388, which are respectively found in the company Asiaplast Industries Tbk in 2019 and Lippo Cikarang Tbk in 2017. The independent variables in this study consists of the ACD, GEN, and TURN variables. The first independent variable is ACD which is measured using the number of years the CEO has held his position in a company with the descriptive results shown in table 1.1 above. From the table, the minimum and maximum values for ACD are 1 and 49. Of the 744 samples there is only 1 company with a CEO who has served for the longest term of 49 years, the company is Ultra Jaya Milk Industry and Trading Company Tbk. in 2019. In addition, during the 2017 to 2019 period, there were 87 companies with new CEOs who served for 1 year from the total sample. It can be concluded that most companies in the eight sectors except for the financial sector have CEOs with a relatively short tenure of 1 year. The CEO's tenure is in accordance with the provisions stated in POJK Number 33 / POJK.04 / 2014 and Law Article 40 of 2005. The second independent variable is GEN which is measured using dummy variables with descriptive results shown in table 1.2. From these results, it is known that from the GEN variable there are 61 samples or 8.2% of the total sample of CEOs who are female. Meanwhile, as many as 683 samples or 91.8% of the total sample of CEOs were male. So that, most companies outside the financial sector have relatively few female CEOs. The third independent variable is TURN which is measured using a dummy variable with descriptive results shown in table 1. 2. By seeing the result, it is known that from the TURN variable there were 87 samples or 11.7% of the total sample of companies that changed the CEO of the previous year. Meanwhile, as many as 657 samples or 88.3% of the total sample of companies that did not change the CEO in the previous year or the CEO in the previous period served in the next period. In conclusion, the number of CEOs who changed from the CEO of the previous period is smaller when compared to the CEOs who remained in office from the previous period. The Chow test is conducted to find a model that is more suitable between the Pooled Least Square (PLS) or Fixed Effect Model (FEM) models in panel regression. Based on the results of the Chow test which are reflected in table 3 above shows the value Prob. Cross-section F of 0.0000 is less than 0.05, so the Fixed Effect Model (FEM) is better to use than the Pooled Least Square Model (PLS).   Lagrange Multiplier test shows the value of Prob. Breusch-Pagan is 0.0 000 so that the model Random Effect is better used instead of the model of Common Effect. Therefore, it is most appropriate to test panel regression data using the Random Effect model.

RESULTS AND DISCUSSION
Based on Table 6, the independent variable TURN was found to have a significant positive effect at the 0.05 level on DA with a significance value of 0.0373. In addition, there are two control variables that also have a significant effect at the 0.05 level on DA, namely SIZE and ROA. There is a significant negative effect in the SIZE variable on DA with a significance of 0.0203. In opposite, the ROA variable has a significant positive effect on DA with a significance of 0.0010. The partial test also shows that the GEN variable and two control variables, namely LEV and MTB, have a positive but insignificant effect on DA.  The results show that the presence of female CEOs has no effect on earnings manipulation, because the significant value is 0.2429, meaning that hypothesis one (H1) is rejected. These results are in line with the research of Peni and Vahamaa (2010), where female CEOs are not able to play a significant role in limiting and reducing the tendency of earnings management in companies. There is no significant effect of gender CEO on earnings management due to pressure from external parties, which is a factor that can affect the characteristics of female CEOs to make decisions regarding corporate financial reporting (Puspa & Adhariani, 2017). Alqatamin et al. (Alqatamin et al., 2017) and Krishnan and Park (Krishnan & Park, 2005) reveal that there is no relationship between the gender of the CEO on the effect of the quality of earnings reported by the company. The research results of Krishnan and Park (Krishnan & Park, 2005) found that male CEOs tend to manage companies properly like female CEOs. Powell and Ansic (1997) also stated that women are more likely to have attitudes that are more persuaded, less confident, less aggressive, and have low leadership and problem-solving skills when making decisions under risk.
These results contradict the research of Bouaziz et al. (Bouaziz et al., 2020), Setyaningrum et al. (Setyaningrum et al., 2019), Faccio et al. (Faccio et al., 2016, Lakhal et al. (Lakhal et al., 2015), and Gavious et al. (Gavious et al., 2012) which revealed that CEOs gender differences will affect earnings management within the company. Basically, women's attitudes tend to be more ethical and risk -averse than men and have high morale, which can reduce earnings management.
Hypothesis 2 is rejected because the ACD variable has a value of t -0.5857 with a significance level above five percent. These results are in line with the research results of Bouaziz et al. (Bouaziz et al., 2020) which revealed there is no significance on tenure CEO towards earnings management. Bouaziz et al. (2020) stated that the longer the CEO has served in the company, the more likely he will be serious and focused on improving the company's situation and choose to contribute to the company's growth and strive to develop the business rather than the CEO with a short tenure. Bergh (2001) in (Bouaziz et al, 2020) stated that the longtime CEO of a company has a wide range of experience in managing the company, so the CEO with a long service life are likely to affect the success in running the company. Ali and p-1412-3789 www.journalmabis.org e-2477-1783 60 Zhang (Ali & Zhang, 2015) tested the tenure of CEOs on earnings management practices, then found that CEOs tend to do income maximization patterned earnings management in the first year the CEO served in the company, because new CEOs generally have the motivation to demonstrate their abilities to shareholders at the beginning of their working period. In this case, the new CEO of a company wants to influence the market's assessment of his abilities in the early years of his tenure. Gibbons and Murphy (Gibbons & Murphy, 1992) also revealed that shareholders do not know with certainty the ability of the newly appointed CEO in a company, thus it takes the ability of the CEO to convince market perceptions of the CEO's ability.
Results were contrary to research Solihah and Herawaty (2019) who find there is a positive influence between CEO tenure of the earnings management. Because the CEO's tenure is longer, indicating that the CEO has a good performance so he can build a good reputation in front of shareholders because of his skills in managing the company. Davidson et al. (Davidson et al., 2007) found that CEOs who are about to retire tend to do earnings management. This aims to obtain incentives in the form of bonuses, pension packages that can be used for consumption after the CEO's tenure ends.
The TURN variable shows a significant value of 0.0373, which means that CEO turnover in the company has a positive effect on earnings management at the 0.05 level, thus hypothesis three (H3) is accepted. In line with the research results of Bornemann (Bornemann et al., 2015) and Hazarika et al. (2012). Previous research conducted in 1.895 companies in the ExecuComp database from 1992 to 2004 found that forced CEO turnover was positively related to earnings management. Wells (Wells, 2002) conducted a study of 100 companies listed on the Australian Security Exchange (ASX) from 1984 to 1994. The results of this study found that CEOs who had recently served in companies tended to carry out earnings management with a taking a bath pattern.
These results contradict the research of Bouaziz et al. (Bouaziz et al., 2020) where there is no significant influence between CEO turnover on earnings management. Murphy and Zimmerman (Murphy & Zimmerman, 1993) suggest that CEO turnover tends to be done when the company's performance is poor or when the previous CEO is approaching retirement.
Choi et al. (Choi et al., 2014) argue that routine CEO turnover will cause new CEOs to be less involved in earnings management, because the successors to previous CEOs tend to have a variety of knowledge and experience in managing the company, and previous CEOs generally continue to monitor the performance of the newly appointed CEO, so as to minimize moral hazard and reduce opportunities for the new CEO to do earnings management. In addition, the forced dismissal of the CEO also has no effect on earnings management because the change tends to be related to financial problems or business restructuring, thereby increasing public supervision of the company. SIZE (company size) has a significant negative effect on earnings management, meaning that the larger the company size, the internal controls contained in the company will be more sophisticated. The company's age (AGE) has no effect on earnings management in the company. This result may be due to a better market reputation and compliance with applicable regulations compared to companies that are just starting out. Leverage (LEV) also did not give any effect on earnings management. The fourth control variable, Return on Assets (ROA) has a significant effect on earnings management. This can be explained because management will tend to carry out income maximization patterned earnings management when the company's performance is poor. Conversely, the practice of income minimization pattern earnings management is carried out when the company's performance is above the profit target expected by the stockholders. Finally, the Market to Book Ratio (MTB) affects the company's earnings management. Basically, a higher share price will encourage management within the company to carry out earnings management with the aim that investors can be attracted to invest in the company.

CONCLUSIONS
This study examines non-financial companies listed on the Indonesian Stock Exchange for the 2017-2019 period that match the specified sample criteria. This study did not find a significant effect of the presence of female CEO and CEO tenure on earnings management. This is based on the attitude of women who tend to be easily persuaded, lack self-confidence, less aggressive, and have low leadership and problem-solving skills when making decisions under risk. This causes women to tend to follow the flow in the company. Although women tend to be ethical, when there is pressure from external parties, this can affect the characteristics of female CEO to make decisions related to corporate financial reporting. There is no significant effect between the tenure of the CEO and earnings management. This is because CEOs with long tenure in the company tend to be more serious and focus on improving the company's situation and choose to contribute to the company's growth and strive to develop the business rather than CEO with short tenure. In addition, new CEO tend to have the motivation to demonstrate their abilities to shareholders at the beginning of their tenure, with the aim of influencing market judgments about their abilities in the early years of their tenure, because shareholders do not know with certainty the ability of the newly appointed CEO in a company. On the other hand, there is a significant influence between CEO turnover on earnings management. This is because the CEO who has just served in the company tends to do earnings management by taking a bath to blame the performance of the previous CEO and reduce the performance benchmarks and profit target that the CEO must achieve in the following year.