THE INFLUENCE OF COMPANY SIZE, MANAGERIAL OWNERSHIP, PROFITABILITY, AND INFORMATION ASYMMETRY ON EARNINGS MANAGEMENT IN LEADING MANUFACTURING COMPANIES IN INDONESIA

How to cite this article (APA): Reschiwati, Lestari, N. P. A. (2021). The Influence of Company Size, Managerial Ownership, Profitability, And Information Asymmetry on Earnings Management in Leading Manufacturing Companies in Indonesia. International Journal of Research GRANTHAALAYAH, 9(10), 37-52. doi: 10.29121/granthaalayah.v9.i10.2021.4295 37 THE INFLUENCE OF COMPANY SIZE, MANAGERIAL OWNERSHIP, PROFITABILITY, AND INFORMATION ASYMMETRY ON EARNINGS MANAGEMENT IN LEADING MANUFACTURING COMPANIES IN INDONESIA


INTRODUCTION
The means of communication used for financial reporting is crucial because it is a way to describe what managers do. One of the many essential parameters in financial statements that is a benchmark for management performance is profit. In addition, profit is also an essential parameter for stakeholders in investing in the company. From the creditor's point of view, profit information is also a tool to provide a company's credit. From the government side, profit information can be used as a parameter to determine the tax imposed on a company. Managers must have an obligation to maximize the welfare of stakeholders, but on the other hand, managers are interested in their welfare. The interests of these parties combined with agency problems are caused by the separation between stakeholders and company management. Sometimes management does not always act in the interests of stakeholders but instead tends to act in the interests of management itself.
In Indonesia, phenomena occurred at PT Garuda Indonesia, where the company's net profit is US$ 809,850 in 2018 and losses of US$ 216.5 million in 2017. The reason is that the management of Garuda Indonesia ensured Mahata's revenue of US$ 239.4 million, the details of which are: worth US$28 million. Part of Sriwijaya Air's revenue share. Garuda Indonesia's Revenue is still in the form of receivables or bills Www.Cnnindonesia.Com. (25 April 2019), (Editor) (2019) The case of PLN achieved a net profit of Rp 11.57 trillion throughout 2018. This profit increased from a profit of Rp 4.41 trillion in 2017. In fact, in the second 38 quarter of 2019, PLN lost Rp 18,480 billion due to a foreign exchange loss of Rp 17,320 billion. The first thing that affects earnings management is the item of remuneration income, where this income does not appear in the 2017 balance sheet but is recorded in the remuneration income account of Rp 23.17 trillion. The other two net revenues increased from IDR 3.40 trillion in 2017 to IDR 15.66 trillion in 2018. The increase in other net income came from government revenue Dwi Ayuningtyas (2019), (Editor) (2019). The case of Bank Bukopin, where its net profit was revised in 2016 from Rp 1.08 trillion to Rp 183.56 billion. The primary decrease was related to the allowance for income and commissions, namely credit card income and additional impairment losses on assets. As a result, the allowance for impairment of financial assets was revised from Rp 649.05 billion to Rp 797.65 billion. As a result, the company's expenses increased by Rp. 66.6 billion Donald Banjarnahor (2018).
The data above illustrates cases of earnings management in Indonesia that often undermine the level of external trust related to the quality of the financial statements presented (Febriani, 2014). The technique of applying earnings management that is much different from time to time requires that earnings management must be observed to prevent or minimize cases of earnings management in Indonesia Mutiarasari (2016).
According to research from Purwanti et al. (2021) and Giovani (2017) company size has no significant effect on earnings management. According to Agustia and Suryani (2018) and Taco and Ilat (2016) company size significantly influences earnings management practices.
Earnings management arises among managers determined by motivation; different motivations will undoubtedly bring about different earnings management practices-for example, shareholder managers with managers who do not hold shares. Ownership will affect decision-making and policies regarding the accounting method used. Earnings management practices will automatically affect because the policies on accounting methods applied in the company are also determined by managers Ujiyantho and Agus Pramuka (2007). Managers who have share ownership are a strategy to reduce agency in the company. Managerial ownership of shares is expected to harmonize the interests of owners and managers, which will have a direct impact on earnings management practices.
Another aspect that correlates with earnings management practices is profitability which describes the profits of a company. Profitability encourages managers to manage earnings management because if a company's profitability is low, a manager will practice income management to keep their results positive in the owner's sight Gunawan (2015). Earnings management is positively influenced by profitability, under the research conducted Aljana and Purwanto (2017), Agustia and Suryani (2018) and Giovani (2017).
In addition to the above factors, the information asymmetry factor also affects the occurrence of manipulative activities. Information asymmetry is the action of a manager having more access to company data that is not known to others. That way, the more data management has about stakeholders, the more opportunities the company controls to practice earnings management.

39
The cause of earnings management errors may be due to information asymmetry. When an information asymmetry is at a high level, and managers do not have enough incentives, resources, or access to reliable data to monitor managers' actions, this can provide an opportunity for managers to manipulate earnings management Desmiyawati & Y. F. (2009). According to Ni & Maria (2016) information asymmetry has a positive effect on earnings management.
LQ 45 is an index on the Indonesia Stock Exchange (IDX) calculated based on the liquidity criteria of well-performing companies. This index consists of 45 issuers. This research focuses on manufacturing companies listed in LQ 45. Although there are many studies on earnings management, the research results are different. It is what underlies why researchers want to examine what aspects affect earnings management practices. There are differences in research in the form of the year to be studied and the software used, namely Eviews 9.
Based on the description above, the researcher will examine "The Influence of Company Size, Managerial Ownership, Profitability, and Information Asymmetry on Earnings Management in Manufacturing Companies listed on the LQ 45 2016-2020".

Agency Theory
Agency theory aims to understand comprehensively and solve problems that arise when the information situation is incomplete when making a contract (engagement). Agency theory describes the relationship between shareholders and agents, namely management. As a shareholder-appointed entity, management works for the benefit of the company and shareholders. Because they are elected, management is required to benefit shareholders Jensen & Meckling (1976) in Reschiwati and Hasudungan (2020).

Agency Cost
Agency theory explicitly considers the costs and benefits that will result from solving an existing agency problem. There are two aspects of costs that can be found in the agency relationship. The first is pre-cost, where a fee is incurred by managers and shareholders entering into a contract. The second is ex-post costs, which are costs incurred when shareholders delegate shareholders and managers in business supervision. Agency-related research generally focuses on ex-post costs because these costs impact the theory of business life Harjito (2011).

Earnings Management
According to Scout (2015) in Sucipto and Zulfa (2021) earnings management is opening policies by managers to increase company value by misleading stakeholders. Financial earnings management can manage from a financial report based on management deliberately manipulating the report to better not describe the company's actual state. Financial statement users may make wrong decisions due to earnings management practices Arlita et al. (2019).

Company Size
Company size shows how big a company is calculated from the Ln of its total assets. According to (Febriyani, 2013) in Giovani (2017) assets are all company resources due to past events. From these assets, the entity expects future profits. Company size is an assessment to describe the company's condition in terms of sales and total assets. The greater the assets and sales mean that the company is classified as a large company Sucipto and Zulfa (2021).

Managerial Ownership
Managerial ownership is a requirement for managers to own shares in the company. This managerial share ownership is also stated in the report on the percentage ownership of the management company's shares. Managerial share ownership can balance managers' and shareholders' interests because managers also benefit from their decisions. However, managers also face the risk of loss as a result of bad decision-making. In a company, the higher the share ownership portion of management, the better the company's performance will be and can also unite the interests of managers and shareholders Sutojo and Aldridge (2005). Managerial ownership refers to the shares owned by the subsidiaries of the company in question and its affiliates. Management ownership of the company can equalize the positions of managers and shareholders, thereby reducing conflicts between managers and shareholders Jensen and Meckling (1976).
Profitability Profitability is an indicator from the management side of managing company assets in terms of the profits earned Sudarmadji and Sularto (2007). Profitability is the company's net profit from its operating activities Gunawan (2015). Profitability is a measure of operating income in a certain period Kieso (2011). The profitability ratio is a number to determine the company's ability to seek profit. The ratio describes a measure of the effectiveness of the company's management. It is evidenced by the profit from sales and investment income (Kasmir, 2013) in Agustia and Suryani (2018).

Information Assymetry
Information asymmetry occurs because the principal does not have adequate data about the agent's performance. The principal cannot choose the agent's business contribution to the actual company (Supriyono, 2000) in Wiryadi and Sebrina (2013). Information asymmetry will affect managers' earnings management practices in Richardson's (1998) research in Ni & Maria (2016. When information asymmetry is high, stakeholders do not have sufficient resources for reliable information to monitor manager behavior, leading to earnings management practices. The impact of information asymmetry can motivate managers to present irrelevant information related to the manager's performance measurement.

Variable Operations Earnings Management
To measure earnings management practices, discretionary accrual proxies can use the Modified Jones Model (Gupta & Suartana, 2018) in Reschiwati and Hasudungan (2020). The calculation model is as follows: Then calculate the estimated total accrual value (TAC) using the following regression equation: By using the regression coefficient, the nondiscretionary accrual (NDTA) value can be obtained using the formula: Discretionary accruals (DTA) are residuals that can be obtained from the estimated total accruals obtained as follows: Information: DTAit = Discretionary accrual of company i in period t NDTAit = Non-Discretionary Accruals of company i in period t Niit = Net profit of company i in period t Tait = Total accruals of company i in period t CFOit = operating cash flow of company i in period t TAit-1 = Total assets of company i in period t-1 REVit = revenue of company i in period t PPEit = total fixed assets of the company in period t

Firm Size
= ( ) Information: SPREAD: The difference between the ask price and the bid for the company's shares aski,t : the highest ask price of the company's stock bid i,t : the lowest bid price of the company's shares

Population and Sample
The population describes an area generalized to include subjects or objects with specific characteristics determined and concluded by the researcher Sugiyono (2013). Manufacturing companies listed on the LQ 45 index from 2016-2020 are the population in this study. This data was obtained through the website www.idx.co.id with 45 companies. The sample is part of the whole and the characteristics that the population has. The purposive sampling method used in this study is: (1). Companies listed on LQ 45 for the 2016-2020 period. (2). Manufacturing companies that are not listed consecutively in LQ 45 during the 2016-2020 period. (3). Manufacturing companies that do not present their financial statements in Rupiah for the 2016-2020 period.

Classic assumption test Normality Test
The normality test is intended to determine whether the data is normally distributed or not and show whether the sample being tested represents a normally distributed population. Usually, the values are assumed to be distributed if the values are standardized and the mean values are close. The normality test is carried out by paying attention to the level of spread of the data on the diagonal axis of the graph or by paying attention to the histogram residual.

Multicollinearity Test
According to Ghozali and Ratmono (2013) a multicollinearity test is used to test the high and low correlation between independent variables in a regression model. This research uses an alternative multicollinearity detector that can be executed by observing the correlation value between two independent variables, with a note that 0.80 is the maximum limit of the correlation value.

Heteroscedasticity Test
The heteroscedasticity test is used to test whether there is an imbalance between residuals or observations with other observations. If the residual variance is from one observation to another observation, then it is called homoscedasticity, and if it is different, it is called heteroscedasticity. The statistical test used in this study is the white test.

Autocorrelation Test
The purpose of the autocorrelation test is to test the residual correlation error (interference) from the linear regression model in period t against errors in the previous period (t-1). A good regression model is a model that does not occur. To test the presence or absence of autocorrelation, symptoms can be detected by the Durbin-Waston test (DW Test).

Hypothesis test Coefficient of Determination (R2)
The coefficient of determination (adjusted R2) was used to measure and explain the influence of the independent variable on the dependent variable. The value of the determinant coefficient in broad ranges between 0 and 1. A value close to 1 (adjusted R2) indicates that the independent variable will present almost all the information needed to explain the dependent variable. On the other hand, if the coefficient value is far from 1, the independent variable only provides limited information.

Partial Hypothesis Testing (t-Test)
The t-test model is used to measure and see the magnitude of the influence of the independent variable on the dependent variable individually. Testing the regression results was carried out using the t-test with the following provisions: H0: if the value of tcount > ttable (0.05), then H0 is rejected. H1: if the value of tcount < ttable (0.05), then H1 is accepted Simultaneous Significance Test (Statistical Test f) According to Ghozali and Ratmono (2013) the f test model can show how much influence the independent variables included in the f test model have on the dependent variable. With the provisions, among others: (1). If the probability (significance) > 0.05, it can be said that the independent variable simultaneously does not affect the dependent variable. (2). If the probability (significance) <0.05, it can be said that the independent variable simultaneously influences the dependent variable. Based on Table 1  Based on Table 2 the probability of a chi square cross-section of 0.0000 is smaller than alpha (0.05) so that Ha is accepted. So, the best estimation method to perform regression test is the Fixed Effect Model. Based on Table 3 shows that the probability of a random cross-section of 0.2868 is greater than alpha (0.05) so that Ho is accepted. Then the regression model used is the Random Effect Model.

Results of the Panel Data Regression Model Selection
Based on Table 4 obtained the value of Prob. Breusch-pagan, showed a result smaller than 0.05, which is 0.0000. So, it shows that the regression model used is the Random Effect Model.

Panel Data Regression Model Selection Results
The Chow, Hausman, and Lagarnge test results of the multiplier show that the best model in this study is the Random Effect Model. The results of the model selection are presented in the following table:   The coefficient value from Table 6, each multicollinearity test result shows no correlation coefficient between independent variables less than (<) 0.80. it can be said the regression model and the four independent variables do not experience multicollinearity problems. Based on the results of the heteroscedasticity test, Table 7 shows that the Prob between the independent variables has a value > alpha 0.05. So, the regression model used does not experience heteroscedasticity problems. Based on Table 8, it is known that the DW value is 1.7902, and the DU value in the Durbin Watson table is 1.7214, so that a comparison of 1.7214 < 1.7902 < 2.2786 (4 -1.7214) means that the data in the study does not experience autocorrelation problems.

Panel Data Multiple Regression Analysis
Multiple regression analysis was conducted to determine the relationship between Profitability, Managerial Ownership, Company Size, and Information Asymmetry on earnings management in manufacturing companies listed on LQ 45 in 2016-2020. The estimation results after going through the Chow test, Hausman test, and Lagrange multiplier test the model used in this study is a random effect model with the following estimation results: Based on Table 9, the equation model using the random effect method can be formulated as follows: DA = -2.6366 + 0.0484 SIZE + 0.3449 MOWN + 0.1211 ROA + 0.0068 SPREAD + e The coefficient determination results in Table 10 above show that r square is 0.2538, which means that the proportion of the influence of company size, managerial ownership, profitability, and information asymmetry on earnings management in manufacturing companies listed in LQ 45 for the 2016-2020 period is 25.38 percent while the rest is 25.38 percent. 74.62 percent (100 -25.38 percent) is influenced by other variables not examined in the study. Based on Table 11, the research hypotheses are: 1) Hypothesis 1: There is an effect of company size on earnings management

T-Test
The probability value of company size (SIZE) < alpha 0.05 is 0.0155 < 0.05 and t count > t table is 2.5164 > 2.0141, this shows that company size has an effect on earnings management. The regression coefficient of 0.0484 indicates a positive direction. It means that company size has a positive effect on earnings management in manufacturing companies listed in LQ 45 for the 2016-2020 period. The conclusion of the study is to accept the first hypothesis. 2) Hypothesis 2: There is an effect of managerial ownership on earnings management The probability value of managerial ownership (MOWN) < alpha 0.05 is 0.0457 < 0.05 and t count > t table is 2.0551 > 2.0141, this shows that managerial ownership has an effect on earnings management. The regression coefficient of 0.3449 indicates a positive direction. It means that managerial ownership positively affects earnings management in manufacturing companies listed in LQ 45 for the 2016-2020 period. The conclusion of the study is to accept the second hypothesis.

3) Hypothesis 3: There is an effect of profitability on earnings management
The probability value of profitability (ROA) > alpha 0.05 is 0.6814 > 0.05, and the t-count < t table is 0.4132 < 2.0141, this shows that profitability has no effect on earnings management in manufacturing companies listed in LQ 45 for the 2016-2020 period. The conclusion of the study is to reject the third hypothesis. 4) Hypothesis 4: Information asymmetry affects earnings management The probability value of information asymmetry (SPREAD) < alpha 0.05 is 0.0031 < 0.05 and t count > t table is 3.1223 > 2.0141, this aspect shows that information asymmetry affects earnings management. The regression coefficient of 0.0068 indicates a positive direction. It means that information asymmetry positively affects earnings management in manufacturing companies listed in LQ 45 for the 2016-2020 period. The conclusion of the study is to accept the fourth hypothesis. Based on Table 12, the probability value of F-statistic is smaller than alpha (0.05) which is 0.0016 < 0.05, and f count > f table is 5.17 > 2.58; this shows that Profitability, Managerial Ownership, Company Size, and Asymmetry Information simultaneously affects earnings management in manufacturing companies listed in LQ 45 for the 2016-2020 period. The study's conclusion is to accept the fifth hypothesis, namely that there is an effect of Profitability, Managerial Ownership, Company Size, and Information Asymmetry on earnings management.

Company Size to Earnings Management
From the results of this study, it can be concluded that the company's size affects earnings management, which means that large companies have more complex operational activities than small companies, making them more likely to carry out earnings management. This study is in line with the research results from Agustia and Suryani (2018); this research is inversely proportional to research results from Purwanti et al. (2021) which states that company size does not affect earnings management.

Managerial Ownership to Earnings Management
From the results of this study, it can be concluded that managerial ownership has a positive effect, which means that the higher the managerial ownership, the higher the level of earnings management, and the lower the managerial ownership, the lower the level of earnings management. This study aligns with research results by Arlita et al. (2019).

Profitability to Earnings Management
From the results of this study, profitability does not affect earnings management because it affects management performance and the high level of profitability. The results of this study are in line with Gunawan (2015). However, the results of this study are not in line with research conducted by Giovani (2017) which states that profitability has a significant effect on earnings management.

Information Assymetry to Earnings Management
Information asymmetry is a factor that affects earnings management actions. It can be concluded from this study that information asymmetry affects earnings management. Furthermore, these results align with research conducted by Ni & Maria (2016) and Mahawyahrti and Budiasih (2016) which state that information asymmetry significantly affects earnings management. It shows that the higher the level of information asymmetry, the earnings management will increase, as happened to PT Indofood CBP Sukses Makmur Tbk (ICBP).

CONCLUSION IMPLICATION AND SUGGESTION Conclusion
Based on the analysis and discussion in the research results conducted by the author is the following conclusions are as follows: (1) Company size partially has a significant positive effect with a coefficient of 0.049 and has a significance value of 0.015; (2) Managerial ownership partially has a significant positive effect with a coefficient of 0.345 and has a significance value of 0.046; (3) Profitability partially has no significant effect on earnings management with a coefficient of 0.121 and has a significance value of 0.681; (4) Partial information asymmetry has a significant positive effect with a coefficient of 0.007 and has a significance value of 0.003; (5) Company size, managerial ownership, profitability, and information asymmetry simultaneously significantly affect earnings management with an F-statistic value of 5.167 and an F-statistical profitability value of 0.001638.