THE INFLUENCE OF OWNERSHIP STRUCTURE ON AUDIT FEE

The aim of the research was to analyze the effect of managerial ownership, foreign ownership, government ownership, ownership concentration, and percentage of shares of multiple large shareholders on audit fees. The independent variables were managerial ownership, foreign ownership, government ownership, ownership concentration, and percentage of shares of multiple large shareholders; while, the independent variable was audit fees. All non-financial companies listed in the Indonesia Stock Exchange in 2013-2015 were sampled. Proportionate stratified random sampling was applied to 207 sampled companies verified using slovin formula and multiple regression analysis to test the hypothesis of the research. This research found that the managerial ownership and the percentage of multiple large shareholders were negatively affected audit fees. Moreover, government ownership positively and significantly influenced audit fee; while, foreign ownership and ownership concentration had no significant effect on audit fee.


INTRODUCTION
As part of corporate responsibility for the implementation of corporate activities, the management of the company is required to issue financial reporting to shareholders. Before issuing, the financial statements should be audited by auditor to ensure the quality of the financial statements. Beside internal auditors, external auditor to which the company must pay for the auditing services; the auditing fees associated with the auditor are also called audit fees (Wahab, Zain, & James 2011); all fees provided by corporate and paid to external auditors for conducting auditing. The audit fees exists because shareholders and company management are in conflict of interest. As management mostly tend to act according to their own interests, shareholders require a supervision by an independent party of the company's management, and the expenses incurred for such monitoring are called audit fees.
The audit fee spent by companies varied as a result of different ownership structures and different control mechanisms used by shareholders to oversee companies' affairs, including financial reporting processes. Previous studies identified audit fees paid to external auditors varied according to the client's ownership structure (Khan, Hossain, & Siddiqui 2011;Partner and Hossain 2007;Adelopo, Jallow, & Scott 2012). In addition, the belief of the external auditor on the company's internal control also varied according to its ownership structure. Therefore the difference level of complexity on audit work was used by external audit company to determine the amount of audit fees charged to clients.
This study used variables of ownership structure consisting of managerial ownership, foreign ownership, government ownership, concentration ownership, and shares percentage of multiple large shareholders as independent variables; while, audit cost was the dependent variable.
Agency theory is one of the relevant theories mostly used to explain how corporate ownership structures affect audit costs. Given a good corporate governance mechanism, shareholders increased the oversight upon the company's management to decrease the occurrence of mismanagement and financial reporting misstatement.
Principal and agent interaction was deeply discussed in agency theory (Jensen & Meckling 1976). In their relationship, shareholders and managers signed in a contract in which the managers had to perform its functions following the interests of the shareholders. In this case, mostly, the shareholders demanded the managers to increase the value of the company to generate maximum profit. However, in reality, the manager often disobeyed what the shareholders wanted them to do. As a result, agency problems raised resulted that the shareholders had to spend substantial fee called agency costs to oversee the managers.
The split between company ownership and control of a company authorized by a manager potentially affect to the company's value adversely. Managers have a tendency to use power and opportunities they have to meet their own interests. Although shareholders have an authority to choose company directors, they do not have direct control over the company because they do not the ones running the company. Therefore, the shareholders were in need of services from external auditors to oversee the company's management.

The Effect of Managerial Ownership on Audit Costs
In the agency theory, contract upon the relationship between shareholders and managers was explained (Jensen & Meckling 1976). The efforts to avoid agency problem -conflict of interest between shareholders and managers -insider ownership or managerial ownership were needed to reduce cost of supervision that should be issued by shareholders. Jensen & Meckling (1976) further explained that managerial ownership could reduce agency problems arising from the separation between management and corporate owners by reducing agency costs, which come from the direct expropriation of funds by managers. Therefore, managers having their own shares could motivate management control (Fleming, Heaney, & McCosker 2005). The higher the number of shares the managers has, the greater the value of the company should the managers increased.
In the study of Mustapha & Ahmad (2011) in Malaysian business environment about the influence of managerial ownership based on agency theory showed that agency theory had predicted the inversely proportional relationship between managerial ownership and monitoring costs. This finding showed that the amount of auditing fees associated with the auditor would be lower if their directors or management owned most shares of the company. This was because directors and management had the ability to access personal information and the ability to properly manage the corporate resources. Therefore, firms with greater organizational ownership required less audit work. Based on the description, the first hypothesis is formulated:

H1:
Managerial ownership negatively affects audit costs

The Influence of Foreign Ownership on Audit Costs
According to Nelson & Mohamed-Rusdi (2015), because of the complexity of financial reporting and a geographical separation, foreign-owned subsidiaries or companies with foreign ownership will pay more for auditing process. The complexity of the financial reporting is caused by the location of the parent company, which is stationed in deifferent country with different accounting rules. The previous literature on audit costs have identified the increasing cost of the audit following the financial report's complexity of the clients (Abdullah, Ismail, and Jamauddin 2008;Salleh, Stewart, and Manson 2006;Goodwin-Stewait & Kent 2006). The complexity of a company might cause agency problems based on which managers took advantages of geographical boundaries to meet their own interests. Meanwhile, as foreign investors expect to have highly auditing quality, the audit costs will also increase because auditor need to spend more time doing the audit details (Zureigat 2011). Therefore, the second hypothesis is formulated as H2: Foreign ownership positively affects audit costs.

Government Ownership Impact on Audit Costs.
The form of government ownership is slightly different from other forms of ownership. The state-owned companies are mostly financed by citizens' money; therefore, the government ownership is very widespread. This makes free riders problem more real than any other form of scattered ownership. In the government ownership, shareholders do not have a strong willingness to oversee directly the management since each shareholder has only a small portion of investment in the company. In fact, the control of governmentowned enterprises is actually conducted by persons in the government (Shleifer & Vishny 1997).
Although individuals within the government do not have the right to cash flow such as receiving dividends from the company, but their reputation will be at stake if the company fails. Most state-owned companies do not have enough costs for internal control such other types of ownership companies have. As a result, the internal control systems in the governmentowned companies tend to be weaker and the agency problems become higher. Consequently, external auditor is necessary to oversee the company's performance and operations. Shareholders in companies with government ownership rely more on audit as a means of controlling corporate management behavior (Chan et al., 1993). Thus, the third hypothesis was formulated as:

H3:
Government ownership positively affects audit costs

The influence of ownership concentration on audit costs
The more the stock block holders within a company are, the more concentrated the company will be. The number of stock block holders relate to oversight level to management as firms with high levels of ownership will have better levels of oversight resulting in lowering audit costs. A research in Malaysian business environment proposed the number of stock block holders as a measurement of the company ownership concentration. The study proved that highly concentrated ownership in firms have a significant effect on audit cost (Adelopo, Jallow, & Scott 2012). Therefore, the fourth hypothesis is formulated that H4: The concentration of ownership negatively affect the cost of audit

The Influence of Share Percentage of Multiple Large Shareholders to Audit Cost
As multiple large shareholders profoundly influence a company compared to common shareholders, the multiple large shareholders affect the cost of audit because they have more access to the company, such as assessing the financial statements before being published (El-gazzar 1998). Multiple large shareholders also tend to interfere in the company management with regard to earnings management done by managers. (Balsam, Bartov, & Marquardt 2002).
Previous studies used percentage of investor ownership as a measurement of the ownership structure and as a method to oversight the management of the company (Adelopo, Jallow, & Scott 2012;Nitisari 2015). Thus, the fifth hypothesis of this study is as follows: H5: The percentage of shares of multiple large shareholders negatively affect the cost of audit.

Dependent Variables
The dependent variable was audit cost (LNAFEE) defined as the auditor's remuneration paid to the external auditor for doing audit services (Che-Ahmad & Abidin 2008; Al-Ajmi 2008). The Audit fees were measured in Indonesian Rupiah and the data were obtained directly from the company's annual report.

Independent Variables
Five independent variables used in this research were managerial ownership, foreign ownership, government ownership, ownership concentration, and percent share of multiple large shareholders. Measurement of each variable were conducted based on the following explanation.

Managerial ownership
This study used a common measurement for managerial ownership based on the total percentage of shares board of commissioners and directors possess to get a better picture of managerial ownership as proposed by Mazlina Mustapha and Ayoib Che Ahmad (2011); Al-Fayoumi and Abuzayed, (2009).

Foreign Ownership
Foreign ownership was operationally measured by the percentage of share ownership by foreign individuals or nonindividuals. Foreign individual is defined as a non-Indonesian citizen or a foreign national; while, foreign non-individual is defined as any company or organization registered or established outside of Indonesia.

Government Ownership
Government ownership in a company was measured by the share percentage owned by government. In this study, the government ownership within listed companies was identified through annual report disclosure regarding the majority of shareholders in a company. The number of multiple large shareholders in a company shows the number of stock blocks holders owned by the company. The concentration of ownership can be measured by the number of investors within the company that have at least 5% of the total shares of the company in the annual financial statements.

Share Percentage of Multiple Large Shareholders
The percentage of shares of multiple large shareholders is the number of shares owned by multiple large shareholders (Adelopo, Jallow, & Scott 2012). Multiple large shareholders are shareholders with at least 5% ownership share.

Sample Determination
The population in this research was non-financial companies listed on the Indonesia Stock Exchange in 2013, 2014, and 2015. The non-financial companies in Indonesia consisted of eight business sectors, including agriculture, mining, basic industries and chemicals, various industries, consumer goods industry, property, real estate, and construction, infrastructure, utilities, and transportation; trade, services, and investment. The sample was determined by proportionate stratified random sampling method; the sampling method was calculated in comparison with the proportion of each sector.

Analysis Method
Methods of data analysis used was four data analysis techniques, which are:

Description of Research Sample
Among 207 samples processed, the disturbing sample data were singled out and the final sample was 144. Source: Secondary data processed 2017.  As the average value is greater than the standard deviation value, the variation in the data was relatively small as the value of each sample was about the average of the count value. In addition, the average value of the audit cost was relatively larger than that of the statistical test results, meaning that most of the companies sampled spent a considerable audit cost. Meanwhile, as the independent variable MOWN has an average value of 0.037974, a standard deviation of 0.1135822, a maximum value of 0.7, and a minimum value of 0, in the sample there were companies' shares not owned by the board of directors or the board of commissioners proven by the average value was smaller than the that of standard deviation. This explained that the variation of the data in the research sample was relatively large. In addition, the statistical test results showed that the relatively small average of the managerial ownership indicated that most companies sampled were not dominated by managerial ownership.

Variable Description
Furthermore, as the independent variable FOWN has an average value of 0.381367, a standard deviation of 1.1238277, a maximum value of 0.9631, and a minimum value of 0, in the sample there were companies that their shares were not owned by a foreign party proven by the average value was smaller than that of standard deviation. This showed that the variation of data in the research sample was relatively large. In addition, as value of the average foreign ownership was relatively small, most companies sampled were not dominated by foreign ownership.
As the independent variable GOWN has an average value of 0.103066, a standard deviation of 0.2399096, a maximum value of 0.8066, and a minimum value of 0, in the sample there were companies that their shares were not owned by the government proven by the average value was smaller than that of standard deviation. Thus, the variation of the data in the research sample was relatively large. In addition, the average value of the government ownership was relatively smaller than that of the statistical test results, indicating that most companies sampled were not dominated by government ownership.
Moreover, as the independent variable of MLS has an average value of 2.5556, a standard deviation of 1.52269, a maximum value of 7, and a minimum value of 1, the non-financial corporation in Indonesia sampled in this study did not have multiple large shareholders more than 8 to be categorized as highly concentrated firms. In addition, the results obtained that as the average value was greater than that of the standard deviation, the variation in the data was relatively for the value of each sample was around the average of the count. Meanwhile, the statistical test results showed that as the average value of ownership concentrations was relatively small, most companies sampled were widely held firms.
In addition, the independent variable SHARE has an average value of 0.716154, a standard deviation of 0.1581142, maximum value 0.9776, and minimum value 0.2048. As the average value was greater than that of the standard deviation, the variation in the data was relatively small because the value of each sample was around the average of the count. In addition, statistical test result revealed that the average percentage value of the multiple large shareholders was relatively large, indicating that most companies sampled were dominated by stockholders (above 5% ownership)

Discussion
The result of the classical assumption test revealed that the regression model used in this study passed the multicollinearity test with tolerance value above 0.1 and VIF value below 10 for all independent variables; while, the heteroscedasticity test using Glejser test was significance at 0.05 for all independent variables, and the normality test of by using one sample of Kolmogorov-Smirnov was significance at 0.05 level for the residual regression model. The result of the first hypothesis (H1) testing of MOWN against LNAFEE showed that the t value was -2.043 and the significance value was 0.043, meaning that MOWN significantly influenced LNAFEE; therefore, H1 is accepted. Thus, the magnitude of the percentage of the managerial ownership shares affected the audit cost of the company. This finding was in line with the one of Niemi (2005) and Mustapha & Ahmad (2011) stating that the greater the percentage of the managerial ownership shares is, the lower the company's audit costs will be.
The result of the second hypothesis (H2) testing of FOWN against LNAFEE showed that the value of t was 0.229 and of significance was 0.819, meaning that the foreign ownership had no significant effect on the audit cost; therefore, H2 was rejected. This was because the level of the complexity of the foreign ownership companies could not be determined by the large percentage of the foreign shares in the company. The complexity of the financial reporting in the foreign firms was caused by the inter-state accounting rules differences resulted from the geographic separation between the stockholders and the management of the company. In addition, the foreign investors from countries with similar characteristics to Indonesia could also be the cause, such as developing countries or Asian countries whose accounting rules were not much different from that of Indonesia. This result was in line with the finding of Niemi (2005) stating that the percentage of foreign ownership shares has no significant effect on audit cost.
The third hypothesis (H3) of GOWN against LNAFEE showed that that value was 2.599 and the significance value was 0.010, which meant that there was a significant influence between GOWN and LNAFEE; so that, H3 was accepted. It could be concluded that the large percentage of shares of the government ownership affected the company's audit cost. This finding supported the research result of Nelson & Mohamed-Rusdi (2015) stating that the greater percentage of the government ownership shares is, the more increasing the company's audit cost will be.
The fourth hypothesis (H4) of MLS against LNAFEE showed that the t value was 1.599 and the significance value was 0.112. These values could be interpreted that the concentration of the ownership did not significantly affect the audit cost, so H4 was rejected. The reason was that in the ownership of shares in Indonesia the majority shareholders had more authority to conduct supervision compared to minority ones. The majority shareholder was those whose ownership was above 50% or those whose ownership was obtained by merger the minority shares up to more than 50%. Therefore, it could be concluded that regardless the number of multiple large shareholders in the company, the significant effect would not exist as long as the multiple large shareholders could not achieve the majority share. The authority to supervise the management of the company was preferred to be the majority shareholders.
The result of the fifth hypothesis (H5) testing of SHARE against LNAFEE showed that that value was -7.757 and the significance value was 0.000, meaning that there was a significant influence between SHARE and LNAFEE, so H5 was accepted. Therefore, the large percentage share of the multiple shareholder shares affected the audit cost of the company. This was in line with the one of Adelopo et al. (2012) and Nitisari (2015) stating that the greater percentage of the shares of the multiple large shareholders the lower the company's audit costs will be.

Conclusion
The aim of the study was to analyze the effect of ownership structure on audit costs on non-financial companies listed on the Indonesia Stock Exchange (IDX) in the period 2013-2015 using 144 samples. In addition, managerial ownership and percentage shares of multiple large shareholders significantly and negatively affect audit cost. While, government ownership significantly and positively influences audit costs. Otherwise, foreign ownership and the concentration of ownership does not significantly affect audit cost.

Limitations
Some limitations or weaknesses of this research consist of factors affecting audit costs in this study consist of only five ownership structure variables; while, there are many other factors that can affect the cost of corporate audit. In addition, this research only takes 69 companies sample each year from 428 non-financial companies listed on BEI. Suggestions Based on the results and the limitations in this study, the suggestion is to add other variables that are expected to affect the company's audit costs; so that, the results would be even better by providing research results that explain factors outside of this research model that may affect audit costs. Furthermore, adding research samples so that data that will be used can be more accurate and more varied compared to this research.