THE EFFECT OF PERSONAL TENURE ON EARNINGS SURPRISE MANAGEMENT

The purpose of this study is to investigate the effect of personal tenure between Audit Partner and Client CEO on firm’s likelihood to do upward earnings management. Using non financial firms listed on Indonesia Stock Exchange period 2012 – 2014, this study finds that personal tenure between Audit Partner and Client CEO have significant positive influence on firms’ likelihood to do upward earnings management and downward forecast guidance. The results of this study indicate that personal tenure positively influences both mechanisms of avoiding negative earnings surprises. Furthermore, since the company's resources in doing both mechanisms are often limited, this study also examines the impact of personal tenure on firm’s choices between the two mechanisms, by testing whether companies choose to do downward forecast guidance, but not upward earnings management or using upward earnings management but not downward forecast guidance. The results show that personal tenure positively associated with the likelihood of firms doing downward forecast guidance without upward earnings management. The result indicates that lower independency of the audit partner due to longer tenure between Audit Partner and Client CEO encourage management to do earnings management by avoiding negative earnings surprise.


INTRODUCTION
Research on how the company performs earnings management continues to grow to date. Research in this area focuses not only on methods of earnings management (whether through accrual mechanisms or real mechanisms), but also on how firms perform earnings management to meet market expectations (earnings surprises management). Prior studies (Matsumoto 2002;Ho et al. 2010) suggest that companies tend to report positive earnings surprises instead of negative earnings surprises. Positive earnings surprises happen when actual earnings reported are higher than analyst forecast, while negative earnings surprises happen when actual earnings reported are lower than the analyst forecast. Because negative earnings surprises can negatively affect both firms performance and stock price, firms tend to avoid negative earnings surprises.
According to (Ho et al. 2010), there are two ways to avoid negative earnings surprises. First, upward earnings management which firms increase their reported earnings by creating accruals. Accruals earnings management is carried out through the selection of accounting policies to achieve a desired financial reporting result. Second, downward forecast guidance, also called as expectation management, which firm's manager, such as CEO, influence the analyst to make downward expectation to avoid negative earnings surprises. The relationship between CEO and analyst can caused firm's CEO to have a big influence on analyst decision. Companies often have limited resources to do both at the same time, or if managers do so, then the market is likely to perceive the behavior as an opportunistic behavior. In order to avoid that, then the manager will tend to choose between upward earnings management and downward forecast guidance, which sometimes called trade-off between the two mechanisms. So, trade-off is when manager avoid negative earnings surprises by using either upward earnings management or downward forecast guidance.
One of the mechanism to limit the earnings game is trough high quality of audit. Firm's annual audit is designed to constrain any material misstatement of earnings reported. Higher quality of audit will reduce the incidence of earnings management. Audit quality is affected by audit tenure. (Ho et al. 2010) suggest that longer firm's tenure will lead to a better audit quality. Upward earnings management will be higher in early audit tenure as (Brown and Pinello 2007) stated that since earnings management is subject to audit procedure, increasing audit tenure further, the upward earnings management will be substituted by guiding analyst forecast downward (downward forecast guidance). Thus, better audit quality might reduce the firms' likelihood to do upward earnings management, but it might increase firms' likelihood to do downward forecast guidance.
Contrast to substitution hypothesis, (Sankaraguruswamy and Sweeney 2005) support complementary hypothesis by suggesting that firms are likely to do both upward earnings management and downward forecast guidance in order to avoid negative earnings surprises. Their results show that company do earnings management using several methods and not just rely on one single method. Lower monitoring role by auditor due to independency impairment caused by longer personal tenure can provide incentives for managers to perform earnings management by doing several methods of earnings management, including earnings management methods to avoid negative earnings surprises such as upward earnings management and downward forecast guidance.
The objective of this paper is to examine the effect of personal tenure between Audit Partner and Client CEO on earnings management using earnings surprise mechanisms. Different from (Ho et al 2010), this paper uses different aproach to measure audit tenure. We use individual audit tenure approach suggested by (Ball et al. 2015) instead of firms audit tenure. Ball et al (2015) suggest that longer personal tenure between Audit Partner and Client CEO may reduce audit quality due to independence impairment between the Audit Partner and Client CEO. Overall, the purpose of this paper is to examine whether personal tenure between Audit Partner and Client CEO following (Ball et al. 2015), affect firms' likelihood to do upward earnings management, downward forecast guidance, and whether firm choose to do downward forecast guidance but not upward earnings management or by doing upward earnings management but not downward forecast guidance.
The rotation and restrictions on tenure of both Public Accounting Firms and Public Accountant (which refers to the signing Audit Partner) have become a much debated issue in Indonesia. The regulations regarding the rotation of Public Accountant and Public Accountant Firms have changed many times during last ten years. Yet, research on how tenure affect the quality of earnings in Indonesia mostly focus on tenure between firm and Public Accounting Firms (firm tenure), not individual tenure. Research on how Audit Partner and Client CEO affect manager's behavior to manage earnings is limited. In addition, the use of the Indonesian context in this study is interesting in view of the fact that the ownership structure of companies in Indonesia tends to be owned by families and has a high concentration of ownership. Under such ownership conditions, the CEO is usually dominated by the family and the CEO's incentives in earning management can be different from the context of the company with scattered ownership as in the US or UK or other countries.
Using the Indonesia context this study provides several evidences. This study shows that personal tenure between Audit Partner and Client CEO is positively and significantly affect firms' likelihood to do upward earnings management, downward forecast guidance, and firms' choices to do downward forecast guidance without upward earnings management. Our results suggest that firm's manager and analyst may develop certain relationship to avoid negative earnings surprise. This study shows that in the context of Indonesia, the results show that CEOs still have an incentive to conduct earnings management through negative avoidance earnings surprises and with increasing length of personal tenure will increase the incentive.
This study makes a number of contributions to the literatures and practice. For literature this study contributes several contributions. First, as far as our knowledge, this paper is among the first to study and provides evidence about the relation between individual Audit Partner and Client CEO with the firms' likelihood to manage negative earnings surprises mechanisms (upward earnings mana-gement and downward forecast guidance). Previous research on audit tenure always uses firm relationships between Public Accounting Firm and client entities. Brooks et al. (2013) examine the tenures of the relationship between Public Accounting Firm and client entities on audit quality as measured by discretionary accrual. While (Davis et al. 2000) measure the relationship between audit tenure using tenure of Public Accounting Firm, auditor independence and earnings management using discretionary accrual proxy. This study incorporates a new relationship model in audit tenure, an individual relationship between Audit Partner and CEO of the company developed by (Ball et al. 2015) and earnings management behavior which is manager's behavior in avoiding negative earnings surprise. Second, this study also provides contribution by using Indonesian context in examining above relationship. Indonesia as emerging market provides interesting context despite of the regulation on audit tenure. Furthermore, in contrast to (Brown and Higgins 2005) examining downward forecast guideline actions by managers in US firms, research in Indonesia has different contexts with US in aspects such as investor protection, market efficiency, and corporate ownership. The Indonesian context with family-dominated ownership structure is interesting to examine to see how CEOs in Indonesia have incentives to do earnings management by avoiding negative earnings surprises and how personal tenure can affect it. Brown and Higgins (2005) state that even in countries such as US earnings management behaviors such as the downward forecast guideline cannot be restricted and there is no regulation governing it, especially in countries with weak investor protection levels such as Indonesia. In addition, although many firms are owned by families with concentrated ownership structures, in practice the measurement of financial performance based on profit figures and stock prices is still very important. With the absence of strong external monitoring from auditors for companies in Indonesia and with the incentive of companies to conduct earnings management through avoiding negative earnings surprises, this study is expected to add literature on the relationship between personal audit tenure and earnings management mechanism with avoiding negative earnings surprises with Indonesian context. This research also gives contribution to the practice. It provides insights for regulator especially in Indonesia about individual relationship between Audit Partner and Client CEO with audit quality, which has not been regulated yet. Using Indonesia as the context, this study provides contribution by giving policy implication on the regulation of audit tenure in Indonesia which only regulate tenure of audit partner with the company.

Regulation on Audit Tenure in Indonesia
In Indonesia, the rotation Public Accountant (which refers to the signing Audit Partner) or Public Accountant Firms is an issue that is often debated. The regulations regarding the rotation of Public Accountant and Public Accountant Firms have changed many times during last ten years. In 2002, the Decree of the Ministery of Finance No. 423/KMK.06/2002 set the general audit services of financial reports of an entity can be done by the same Public Accountant Firms with maximum tenure of five years consecutively and by the same Public Accountant (signing partner) for maximum of three years consecutively.
In 2008, the Ministry of Finance published Decree No. 17/PMK.01/2008 which set an extension of the audit tenure period for Public Accountant Firms. Public Accountant Firms were allowed to perform general audit services for maximum period of six consecutive years. Public Accountant Firms were also allowed to provide audit services to the same client after a paused of time of one year. The same applies to a Public Accountant who were allowed to provide audit services to the same client for maximum of three years in a row with intervals of one year.
On April 6 th , 2015 the government of Indonesia issued the Government Regulation No. 20 which revised the period of rotation for Public Accountant and Public Accountant Firms. Public Accountant are allowed to perform audit services with maximum tenure of five years in a row and can gives audit services to the same client after period of two years. However, this only applies to the entities listed in the capital market, Bank, Pension Fund Insurance Company, and State Owned Enterprises. Meanwhile, the tenure of Public Accountant Firms who previously was limited to six years, now freed from the rotation, which means that the Public Accountant Firms can perform audit services to an entity with unlimited period of time, as long as the signing partner is rotated with maximum tenure of five years.
Recently in 2017 the Indonesian government issued the latest regulation regarding the appointment of auditors through Financial Services Authority Regulation (Peraturan Otoritas Jasa Keuangan/ POJK) No. 13 / POJK.03 / 2017. The regulation governs the limitation of the use of audit services from the same public accountant (signing partner) for the period maximum of 3 (three) consecutive reporting years. The signing partner may only give audit service after 2 (two) consecutive reporting years of cooling-off period. There is no limitation on the tenure for the public accountant firms.
The regulation above, opens the freedom to the Public Accountant Firms to audit with unlimited tenure, but gives the limitation only on the signing partner. This regulation was issued because many practitioners in Indonesia consider previous regulation was not effective because many Public Accountant Firms addressing the rotation rule by changing the partner and build new Public Accountant Firms partnership in order to change the name of the firms, even though most of the composition of an accountant public and the international affiliation are still the same. This is often called with pseudo rotation.
Despite of this is a frequently debated issue and the regulation oftenly changes, research on the relationship between tenure audit and the quality of financial statements in Indonesia is still very limited. Fitriany et al. (2015) state that in the pre regulation period (1999)(2000)(2001), increasingly length of audit tenure, the lower the quality audit. In the post-regulation period (2004)(2005)(2006)(2007)(2008), the empirical evidence shows a convex relationship between audit tenure and audit quality from the side of neutrality and timeliness. Different with this research, Fitriany et al. (2015) use firm tenure rather than individual tenure.
Up to now, there is no regulation in Indonesia, and in many countries that regulate the maximum tenure period of the relationship between individual party in the company and in the audit firms such as tenure between Audit Partner and Client CEO. The independency of the auditor can be impaired due to close relationship between audit partner and CEO. The longer the tenure between Audit Partner and Client CEO the higher the probability of independency impairment (Ball et al. 2015). Ball et al. (2015) states that personal tenure between signing partner and CEO can also impair independency from the auditor thereby decreasing audit quality.

The Relation between Upward Earnings Management and Downward Forecast Guidance
Several studies (Matsumoto 2002;Brown and Pinello 2007;Ho et al. 2010) state that upward earnings management and downward forecast guidance are mechanisms on earnings surprise game. Downward forecast guidance is also called as expectation management, by managers lowering investors' earnings expectation by providing guidance to analysts that are lower than earnings that can be achieved by the company. Downward forecast guidance may become alternative method over accrual based upward earnings management because of the following reasons. First, earnings management is subject to audit procedure and there is a higher requirement on the financial disclosure. Second, it is difficult for manager to continously upwardly manage reported earnings (Li et al. 2005). The choice of using the earnings management method shows that between the earnings management method can be either substitute or complementary to each other.
The substitution mechanism means that management will choose one method against other methods, resulting in a tradeoff between the two methods. For example, managers will tend to do accrual based upward earnings management, but because of the reasons above, the downward forecast guidance method becomes a substitution of management's upward earnings management method. In comparison with firms unaudited quarterly result (Brown and Pinello 2007) also find that firms annual audit will decrease the firms likelihod to do upward earnings management, but will increase firms' likelihood to do downward forecast guidance. Their evidence later concludes that there is substitution effect between upward earnings management and downward forecast guidance.
In contrast, (Sankaraguruswamy and Sweeney 2005) suggest that firms are likely to do both upward earnings management and downward forecast guidance in order to avoid negative earnings surprises. Their result supports complementary hypothesis rather than substitution hypothesis. It shows that company do earnings management using several methods and not just rely on one single method. Mikhail et al. (1999) show that analyst turnover is related to analysts' forecast accuracy, which suggests that an analyst is concerned about managers that involve the analysts in making large forecast errors. In response of this, firms managers and analysts may develop a symbiotic relationship regarding earnings announcements and earnings forecasts. If managers plan to announce earnings that are lower than analyst forecast, they have an incentive to manage analyst expectations downward, and they try to manage small gap differences between earnings reported and earnings forecast. In addition, to achieve certain level of earnings, firms managers will do upward earnings management, but with less extreme.
Until now, as long as the researcher's knowledge there is no research on the mechanism of earnings management through avoidance of negative earnings surprises in Indonesia or in emerging countries that have some institutional characteristics that are different from developed countries like US. However, given that firms in Indonesia still have incentives for earnings management, and the company's resources to perform all the methods of earnings management are limited as well as the absence of regulation in restrict the manager in doing so referring to (Brown and Higgins 2005) which state that in countries such as US, regulatory restrictions for earnings management behaviour trough method such as downward forecast guidance even does not exist), then the behaviour of earnings management by avoiding negative earnings surprises in Indonesia allegedly still occur.

The Relation between Audit Tenure and Audit Quality
There are two opposing views on how audit tenure affects audit quality (Ball et al. 2015). The first view is the auditor independence hypothesis. The auditor independence hypothesis maintains that auditor independence, and therefore audit quality, becomes impaired as the association between the auditor and the client lengthens. Three argument to support the hypothesis. First, auditors may develop a "learned confidence" or become too familiar with the client's operations. Second, longer auditorclient relationships could lead to the development of person-to-person relations which can impact an auditor's objectivity and therefore independence. Third, as auditor tenure increases, economic considerations could impact decisions and conduct as explained in the low balling practice (DeAngelo 1981).
The second view is the auditor expertise hypothesis. The auditor expertise hypothesis maintains that audit quality increases with auditor tenure as it allows client specific knowledge and expertise to develop and increase. This hypothesis is based on the degree of information asymmetry between the auditor and the client, which reduces over time as auditors acquire client specific knowledge. The knowledge and expertise are developed over repeated audits and create significant learning curves during the period of audit. Myers et al. (2003) documented that longer firms audit tenure associated with decreased Discretionary Accruals (DA) and interpret this result as longer audit tenure positively affect audit quality. While Johnson et al. (2002) found that longer firms audit tenure positively affect eanings management, thus reduce audit quality.
Prior research about audit tenure are mostly focus only on firms relation. Ball et al. (2015) extend the study of audit tenure by examining the individual relationship between Audit Partner and Client CEO. They found that individual relationship of Audit Partner and Client CEO also contribute to affect audit quality. They use client -CEO as benchmark since it is most likely that CEO involved in firms decision on selecting audit firms. The result suggest that longer personal tenure between Audit Partner and Client CEO will reduce audit quality due to independence impairment between Audit Partner and Client CEO. Dechow et al. (2003) argue that avoiding negative earnings surprise is the biggest concern for firms in 1999 until 2001. It is also supported by (Brown and Caylor 2005). Using sample from 1996 until 2002, they found that firms are most likely to avoid negative earnings surprise. There are two ways firms manage to avoid negative earnings surprise: upward earnings management and downward forecast guidance (Matsumot 2002;Ho et al. 2010). Upward earnings management is measured by positive discretionary accruals, while forecast guidance is refer to earnings forecast by analyst. Downward forecast guidance happens when the actual earnings reported is lower than earnings forecast by analyst.

Hypothesis Development
The first hypothesis of this paper is about the relation between personal tenure of Audit Partner and Client CEO with firms' likelihood to do upward earnings management. Ball et al. (2015) find that personal tenure between Audit Partner and Client CEO will reduce audit quality due to independence impairment of Audit Partner and Client CEO. Mautz andSharaft (1961 in Fitriany et al. 2015) state that with the length of the relationship between auditor and his client will affect the independence of the auditor because the auditor's objectivity will decrease over time. With the decrease in independence the auditor is no longer able to conduct audit in high quality, so that management can perform earnings management, which in most case by performing upward earnings management. Using the Indonesian context Fitriany et al. (2015) states that the audit tenure will reduce the quality of audits by using earnings management as proxy when the tenure audit has passed the optimal level. If lower audit quality associated with higher incidence of earnings management. Then our first hypothesis is: H1: Personal tenure between Audit Partner and Client CEO is positively associated with upward earnings management in order to avoid negative earnings surprise.
The second hypothesis of this paper is about the relation between personal tenure of Audit Partner and Client CEO with firms' likelihod to do downward forecast guidance. Brown and Pinello (2007) suggest that downward forecast guidance is a substitution mechanism for upward earnings management. They find that in comparison with unaudited quarterly result, annual audit reduces firms' likelihood to do upward earnings management, but will increase firms' likelihood to downward forecast guidance. In the contrary with (Brown and Pinello 2007), Sankaraguruswamy and Sweeney (2005) find that firms are likely to do both upward earnings management and downward forecast guidance in order to avoid negative earnings surprises. Their result supports complementary hypothesis rather than substitution hypothesis.
In a context where the role of external oversight by auditors is relatively low, and managers have the relatively higher freedom to give downward forecast guidance (considering there are no regulations that restrict it) the substitution relationship between the two methods may not necessarily applicable. In the Indonesian context, where investor protection is relatively weak, managers have relatively more freedom to provide a lower forecast guidance to analysts and the role of auditors to avoid it is relatively weak. In this condition, the length of the personal tenure will weaken monitoring role by the auditor on earnings management behavior through downward forecast guidance. Our second hypothesis is: H2: Personal tenure between Audit Partner and Client CEO is positively associated with downward forecast guidance in order to avoid negative earnings surprise.
Both hypotheses above assume that managers have the ability (and resources) to perform both the earnings management methods through the avoidance of negative earnings surprises ie upward earnings management and downward forecast guidance. If the manager has limited ability or if managers do both, then the market is likely to perceived the behavior as an opportunistic behavior, then the manager is faced with the choice of one method between the two methods.
In relation to managers' behavior in earnings management, the auditor plays a role in carrying out its external monitoring function to minimize the opportunistic behavior of the manager. Nevertheless, the role of auditors leads more to the context of financial reporting in which the auditor plays a role in limiting the manager's opportunistic behavior in upward earnings management. Auditors are often unable to supervise manager relationships with analysts so that the auditor's ability to limit downward forecast guidance is relatively small. Therefore, when managers are faced with a choice of methods, the longer personal tenure between auditors and CEOs will cause managers to prefer upward earnings management rather than downward forecast guidance. The impairment in auditor independence due to the length of personal tenure will cause management to avoid negative earnings surprises by choosing upward earnings management rather than downward forecast guidance. This argument shows that the length of personal tenure negatively impacts the choice of downward forecast guidance rather than upward earnings management. Thus the longer the personal tenure then the company's likelihood to do upward earnings management and not do downward forecast guidance is higher. Therefore, the relationship between personal tenure and tradeoff between downward forecast guidance and upward earnings management becomes negative. Our third hypothesis is as follow: H3: For firms conducting trade-off between upward earnings management and downward forecast guidance, personal tenure between Audit Partner and Client CEO is negatively associated with the choice of using downward forecast guidance but not upward earnings management in order to avoid negative earnings surprise.
Recognising that audit expertise may also be increasing with longer audit firm tenure based on the auditor expertise view, we control for the tenure of the relation between the audit firm and the client and audit firms size to determine whether there are further benefits arising from audit firm rotation, or whether this would impose additional costs in terms of lower audit quality.

Sample Selection
The data used in this paper is based on financial reports for all industry, excluding financial industry, listed in Indonesia Stock Exchange for period 2012 -2014. This period was taken to avoid any regulatory changes related to audit tenure, given the regulatory changes related to tenure audit in 2015. We use secondary data obtained from few sources. We obtain financial report data and earnings announcement date from Thomson Reuters, stock trading data from Datastream, and EPS forecast published by analyst was obtained from I/B/E/S Thomson Reuters.
Our initial sample for H1 contains 1,272 firm-year observations during 2012-2014. We exclude 240 observations from financial services industries because incentives differ across firms in regulated and non-regulated industries. We find 492 observations that do not have complete data to estimate discretionary accruals, individual tenure, earnings surprises and control variables. Finally, after applying the above criteria, our full sample for testing H1 consists of 540 observations. To test our second hypothesis, we exclude 228 observations that do not have complete data of EPS forecast published by analyst. So that, our full sample for testing H2 consists of 312 observations. To test H3, we further limit our sample from H2 by including only observations that either upward earnings management or downward forecast guidance but not both.
Finally, we get full sample consists of 158 data for our H3.

Empirical Model
To examine whether personal tenure between Audit Partner and Client CEO affect firms' likelihood to do upward earnings management, downward forecast guidance, and trade-off choices strategy to use downward forecast guidance but not upward earnings management in order to avoid negative earnings surprises we use logistic regression between personal tenure and probability of upward earnings management, downward forecast guidance, and trade-off choices. We developed the empirical model based on (Ho et al. 2010) that investigate the effect of audit tenure and earnings surprise management and modified based on (Ball et al. 2015) to investigate the relationship between personal tenure and negative earnings surprises avoidance. Our empirical model to test the relationship between personal tenure and the likelihood of firms avoiding negative earnings surprises by doing upward earnings management, downward forecast guideline and the choice of strategy trade-off are as follow: Hypothesis 1 is tested by looking at the significance of α1, where hypothesis 1 is accepted if α1 is positive. Hypothesis 2 is tested by looking at the significance of β1, where hypothesis 2 is accepted if β1 is positive. Whereas hypothesis 3 is tested by looking at the significance of γ1, where hypothesis 3 is accepted when γ1 is negative.
To measure the negative earnings surprises, we follow (Ho et al. 2010). UPWARD_EM is a dummy variable with the value of 1 for firms with positive discretionary accruals, indicating firms' likelihood to do upward earnings management and value of 0 for firms with negative discretionary accruals, indicating firms' likelihood to do downward earnings management. DOWNWARD_FG is a dummy variable with the value of 1 for firms with recent earnings forecast by analyst (FORE) is lower than expected earnings forecast (E[FORE]), and value of 0 for firms with recent earnings forecast by analyst (FORE) is higher than or equal with expected earnings forecast (E[FORE]). CHOICE is variable to identify whether the firms choose to use downward forecast guidance and not upward earnings management or vice versa. This variable tends to capture the method of company use in doing trade-off. CHOICE is measured by a dummy variable with the value of 1 for firms using downward forecast guidance, but not upward earnings management (DOWNWARD_FG = 1 and UP-WARD_EM = 0) and value of 0 for firms for firms using upward earnings management but not downward forecast guidance (DOWNWARD_FG = 0 and UP-WARD_EM = 1).
In empirical model we control the likelihood of firms to avoid negative earnings surprises by adding several control variables. First, we control for Audit Firm Tenure as (Myers et al. 2003) suggest that longer audit tenure is associated with decreased discretionary accruals, and interpret this result as longer audit tenure is positively associated with audit quality. In contrast, (Johnson et al. 2002) find that longer audit tenure is associated with increased discretionary accruals, and suggest that longer audit tenure is negatively associated with audit quality and the lower audit quality will increase the earnings management or in the other words low audit quality will provide an opportunity for managers to avoid negative earnings surprises through earnings management. Based on (Johnson et al. 2002), we predict the relationship between firm tenure and manager's behavior in avoiding negative earnings surprises is positive. Following (Ball et al. 2015) audit firm tenure is measured as the number of years audit firm (if the audit firm has foreign audit firm affiliation we measure based on the tenure of the foreign affiliation) and client firm relationship at the fiscal year end (FTENURE). The FTENURE are measure start from 2009. Due to the research period used in this study is from 2012 to 2014, the retrospective search to measure audit tenure over the past five years is considered adequate.
Second, we control for Growth Prospect. Compared with lower growth prospect firms, higher growth prospect firms have higher incentive to avoid negative earnings surprises since they suffer to higher market assymetric reaction (Skinner and Sloan 2002). Matsumoto (2002) find that firms with higher growth prospect are more likely to do earnings management and downward forecast. While (Brown and Pinello 2007) suggest that firms with higher growth prospect are tends to avoid negative earnings surprises but negatively associated with earnings management and forecast guidance. Growth prospect measured by market to book ratio (LTG).
Third, we control for Litigation Risk. A sudden drop in share price at earnings announcement will lead to shareholder litigation. Firms with higher shareholder litigation risk are more likely to avoid negative earnings surprises (Matsumoto 2002;Ho et al. 2010). Consistent with (Ho et al. 2010) and (Matsumoto 2002), we classify firms in biotechnology, computer, electronic, and retailing as firms with high shareholder litigation. We use dummy variable for Litigation Risk (LIT). Value one assigned for firms in biotechnology, computer, electronic, and retailing industry, and value zero assigned for otherwise.
Next we control for the Implicit Claim to Employee as Brown et al. (1995) and Matsumoto (2002) suggest that firms with higher dependence on implicit claim to employee are tends to beat the analyst forecast and avoid negative earnings surprises. Implicit claim to the employee is done when the employee executes the ownership of share of the company (e.g. trough employee stock ownership program). Therefore, firms that have high implicit claim amounts tend to try to beat the analyst forecasts and avoid negative earnings surprises in order for their stock prices to rise. This variable is measure by labor intensity, LABOR, computed as one minus the ratio of gross Propert, Plant, and Equipment to Total Asset.
We also control for Firms Size. Although large and small firms have the same incentive to avoid negative earnings surprises (Llukani, 2013), large firms tend to be less optimistic in making future financial report projection. Those firms are easier to do downward forecast guidance since they do not need to be involved with earnings management (Brown and Pinello 2007). In accordance with (Brown and Pinello 2007;Matsumoto 2002) found that larger firms are positively associated with downward forecast guidance. Firms size measured by using the log of market value of equity (LNMV).
To control the audit quality, we use Auditor Size as a proxy. Prior research (Becker et al. 1998;Francis et al. 1999) document that larger audit firms are positively associated with audit quality. We defined size of audit firms as Big 4 and non Big 4 membership. Value one assigned to firms using Big 4 auditor, and value zero assigned to firms using non Big 4 auditor (BIG4). Previous research also suggest that leverage is positively associated with firms earnings management. Betty and Weber (2003 find that firms with high leverage are more likely to do earnings management in order to avoid debt covenant. However, (Jellinek, 2007) examine the leverage increases towards earnings management and concludes that that leverage is negatively associated with earnings management. In order to control the effect of leverage in our estimation, we use the ratio of long term debt to total asset (LEV) as control variable.
Last, we also control for industry because the behavior of earnings management in order to avoid earnings surprises might different across industry. We include dummy industry in our model. Industry classification used in this paper is based on Indonesian Stock Exchange, consists of 8 (eight) industry : Mining Industry (MINING), Basic Chemicals (BASIC-CHEM), Consumption Industry (CON-SUMPTION), Service, Trading, and Investment (STI), Agriculture (AGRI), Property and Real Estate (PROP), Transportation, Infrastructure, and Utility (TRANS), and Other (OTHER). We use TRANS as industry base.

Measurment of Personal Tenure between Audit Partner and Client CEO
We measure the personal tenure between Audit Partner and Client CEO as the number of years Audit Partner -Client CEO relationship at the fiscal year end. We use sample for all industry, exclude financial industry, in 2012 until 2014. However, since the mandatory regulation for Audit Partner rotation in Indonesia is maximum for 3 years in a row, we trace the rotation of Audit Partner back from 2009. Personal tenure is measured using the number of years of assignment between the Partner Audit and the Client's CEO. The following is illustration to calculate the individual tenure between the Partner Audit and the CEO in this study.
Based on Table 1, it can be seen that the method of calculating the individual tenure of Partner Audit and Client's CEO is based on the relationship between the same people. For example in 2009, the relationship is between A and X. This relationship is calculated as 1 year. In 2010, the relationship is still the same as in 2009, which is between A and X so that the calculation of tenure increased to 2 years. In 2011, the Audit Partner changed to Y, so the relationship became between A and Y. This relationship is a new relationship, so the tenure is again counted as 1 year. Then in 2012 the relationship goes back to A and X, continuing the same relationship as in 2009 and 2010 so that its tenure is counted as 3 years.

Measurement of Upward Earnings Management
We measure upward earnings management by using discretionary accruals. To develop proxy for discretionary accruals, we use the (Kothari et al. 2005) model. Kothari et al. (2005) found that discretionary accruals estimated by Jones and modified Jones model are likely contain some errors regarding firms performance, so ROA is added in the model to control firm performance. Consistent with our first hypothesis, we use positive discretionary accruals as an indicator of upward earnings management. The Kothari et al. (2005) model is as follows: Where TAijt is total accrual for firm i, in industry j in year t. Total accrual is defined as earnings before extraordinary and discontinued operation minus by operational cash flow. Aijt-1 is total asset for firm i, in industry j in year t-1. ΔREVijt is change in revenue for firm i, industry j in year t. ΔRECijt is change in receivable for firm i, industry j, in year t. PPEijt is total gross property, plant, and equipment for firm i, in industry j year t. ROAijt-1 is return on asset for firm i, in industry j year t-1. We estimate the model for each firm year using all firm observation, exluding financial industry, from the same industry classification by Indonesian Stock Exchange. We assign value one for firms-year with positive discretionary accruals as an indicator of upward earnings management, and value zero for firms-year with negative discretionary accruals as an indicator of downward earnings management.

Measurement of Downward Forecast Guidance
We use following model to develop proxy for downward forecast guidance. Ho et al. (2010) also use this model to measure downward forecast guidance following (Matsumoto 2002). The model is as follows: Where ΔEPSijt is change in earnings per share for firm i, in industry j in year t. Pijt-1 is price per share for firm i, in industry j in year t-1. CRETijt is cumulative daily excess return for firm i, in industry j in year t. Cumulative return measured from 3 (three) days after earnings announcement in year t-1 until 20 (twenty) days before earnings announcement in year t. Earnings announcement date is based on publication date listed in Thomson Reuters.
We estimate the above model for each firm-year using all firm-years from the same industry, excluding the firm for which we are estimating the parameters. In this paper, we limit the parameters only for all firm-years with a complete earnings forecast data for period 2012 -2014.
To determine the expected change in EPS E[ΔEPS], we use following model suggested by Matsumoto (2002): We add the expected change in EPS to the earnings from the prior year to estimate the expected forecast of the current year's earnings (E[FORE]). To determine whether firms use downward forecast guidance, we compare the expected earnings forecast (E[FORE]) with the most recent earnings forecast prior to the early announcement date by analyst ([FORE]). We categorize firms as 1 if if FORE firm i, in industry j year t. ROAijt-1 is return on asset for firm i, in industry j year t-1. We estimate the model for each firm year using all firm observation, exluding financial industry, from the same industry classification by Indonesian Stock Exchange. We assign value one for firms-year with positive discretionary accruals as an indicator of upward earnings management, and value zero for firms-year with negative discretionary accruals as an indicator of downward earnings management.

Measurement of Downward Forecast Guidance
We use following model to develop proxy for downward forecast guidance. Ho et al. (2010) also use this model to measure downward forecast guidance following (Matsumoto 2002). The model is as follows: Where ΔEPSijt is change in earnings per share for firm i, in industry j in year t. Pijt-1 is price per share for firm i, in industry j in year t-1. CRETijt is cumulative daily excess return for firm i, in industry j in year t. Cumulative return measured from 3 (three) days after earnings announcement in year t-1 until 20 (twenty) days before earnings announcement in year t. Earnings announcement date is based on publication date listed in Thomson Reuters.
We estimate the above model for each firm-year using all firm-years from the same industry, excluding the firm for which we are estimating the parameters. In this paper, we limit the parameters only for all firm-years with a complete earnings forecast data for period 2012 -2014.
To determine the expected change in EPS E[ΔEPS], we use following model suggested by Matsumoto (2002)

Descriptive Statistic
The sample selection process can be seen in Table 2. For full sample, the descriptive statistic is displayed in Table 3. UPWARD_EM is variable that represent upward earnings management. Based on Table 3 the frequency of firm that has positive discretionary accrual or firms that conduct an upward earnings management is 45% which means as much as 45 percent of the total observations in this research is the company that do upward earning management. Slightly larger than UP-WARD_EM, variable DOWNWARD-FG shows that 47% of the sample conduct downward earnings forecast guideline practice. Table 3 also shows that 47% of the sample that conduct trade off choose downward forecast guidance but not upward earnings management and 53% do trade off by choosing upward earnings management but not downward forecast guidance.
Personal tenure based on Table 3 shows that the maximum tenure of Audit Partner and CEO of the client is five years. Table 3 which is consist of full sample shows that the average personal tenure is 1,9 years. From reduced sample to test hypothesis 2 and 3, the average of personal tenure is 2 years. The comparison between full sample and sample to test hypothesis 2 and 3 shows that the average of personal tenure between sample is similar. Compare to personal tenure the average tenure of firm tenure, tenure between audit firm and the company is 3,6 years, with the maximum period of 6 years. From reduced sample to test hypothesis 2 and 3, the average of firm tenure is also similar which is 4 years. This shows that in term of personal tenure the duration is shorter compare to the firm tenure. Table 3 shows that the average of LTG variable which control the growth rate of the company is 2,9381. By using market to book ratio, the average value of 2,9381 shows that the market value of the company in average exceed almost 3 times of the book value. This shows that the average companies in the sample are growing companies. Similar with the hypothesis 2 and 3 sample, the variable of LTG are 2.779 and 2.784 respectively, which show consistency with full sample, that both sample also from growing companies. LIT variable is the control variable that controls the litigation risk. Based on Table 3 for all sample, most of the company has little litigation risk. LABOR variable is the control variable that represents the value of the implicit claims to employees. The value of LABOR for full sample in average is 0.5295 which implies that the company in the sample is not highly labor intensive and face a moderate claim to employee. For hypothesis 2 and 3 sample the average LABOR is 0.5103 and 0.5124, which indicates the similar intensity with the full sample. MVOE variable is the value of the market value of equity that control size. In the Table 3, MVOE variable presented in the amount of billions of Rupiah. The value of the average MVOE based on full sample is 15.134 billion Rupiah. While the value of the average MVOE variable on the model 2 and model 3 is 21.474 and 21.743 billion Rupiah. The dispresion of the market value of equity is very high which represents wide deviation of the size of the companies in the sample. The full sample and hypothesis 2 and 3 sample shows similar size of the companies. The BIG 4 is a variable control that represents the classification of the auditor using auditor size, big 4 and non-big 4. Based on the descriptive statistics, the frequency of companies audited by Big 4 is 49% and companies audited by Non Big 4 is 51%. From reduced sample to test hypothesis 2 and 3, the companies audited by Big 4 is 61% and 58% respectively.
Variable LEV is a variable control that represents leverage. Based on the descriptive statistic of full sample, the average of leverage is 13,94%, which shows that 13,94% of the asset was funded by debt. Based on the industry classification, the de scriptive statistic shows that companies from Service, Trading, and Investment holds largest proportion of the sample, 23%, followed by property and mining. Based on the Table 3, we can conclude that the characteristic of the full sample and sample for hypothesis 2 and 3 is quite similar and there is no significant differences between sample for model 1, 2, and 3. Table 4 shows correlation between variables. Based on the Table 3, Personal Tenure has positive correlation with the upward earnings management, downward forecast guidance, and choice of trade-off strategy by choosing downward forecast guidance, but not upward earnings management. Initial indication from correlation test shows that the higher tenure between CEO and auditor the higher earnings management in avoiding negative earnings surprises. Table 4 also shows that Firm Tenure also has positive association to the upward earnings management, downward forecast guidance, and choice of trade-off strategy by choosing downward forecast guidance, but not upward earnings management. This result also provides initial indication that the firm tenure increases the likelihood of firm doing earnings management by avoiding negative earnings surprises.

Regression Result
This research uses multiple logistic regression where the dependent variable of this research is categorical or binary that is discretionary accrual, downward forecast forecast guidance, and choice of trade-off strategy by choosing downward forecast guidance, but not upward earnings management. While the independent variables in this study include audit tenure Partners and CEOs, tenure clients and KAP, growth companies, companies categorized as UPWARD_EM: upward earnings management measured by dummy variable with the value of 1 for firms with positive discretionary accruals, indicating firms' likelihood to do upward earnings management and value of 0 for firms with negative discretionary accruals, indicating firms' likelihood to do downward earnings management.; DOWNWARD_FG: downward forecast guidance measured by dummy variable with the value of 1 for firms with recent earnings forecast by analyst (FORE) is lower than expected earnings forecast (E[FORE]), consistent with downward forecast guidance and value of 0 for firms with recent earnings forecast by analyst (FORE) is higher than or equal with expected earnings forecast (E[FORE]), inconsistent with downward forecast guidance; CHOISE: Trade-Off Choice Strategy measured by dummy variable with the value of 1 for firms using downward forecast guidance, but not upward earnings management (DOWNWARD_FG = 1 and UPWARD_EM = 0) and value of 0 for firms for firms using upward earnings management but not downward forecast guidance (DOWNWARD_FG = 0 and UPWARD_EM = 1); PTENURE: Personal tenure between Audit Partner and CEO of the clients; litigation companies, implicit claims against employees, company size, and auditors BIG 4. Logistic regression testing is done by using pool test. We show our result examining the relation between personal tenure between Audit Partner and Client CEO with earnings surprise management in Tables 4 -6 below. Table 4 shows the relation between personal tenure of Audit Partner and Client CEO with firms' likelihood to do upward earnings management, while Table 4 shows the relation between personal tenure of Audit Partner and Client CEO with firms' likelihood to do upward earnings management, while Table 4 shows the relation between personal tenure of Audit Partner and Client CEO with firms' likelihood to do downward forecast guidance. Finally, we show the relation between choice of trade-off strategy by choosing downward forecast guidance, but not upward earnings management with personal tenure of Audit Partner and Client CEO in the Table 5. Table 4 shows that PTENURE, FTENURE, LABOR, and BIG4 are significantly related with the likelihood of firms have positive discretionary accruals. PTENURE is positively and significantly affect the likelihood of firms do upward earnings management, suggesting that the longer period of personal tenure between Audit Partner and Client CEO increase the firms' likelihood to do upward earnings management. Our result support (Ball et al. 2015) that longer relation between Audit Partner and Client CEO may impair the independence between Audit Partner and Client CEO. FTENURE is also positive and significant. This means, consistent with the result of PTENURE, longer relation between audit firms and client firms increased firms' likelihood to do upward earnings management, characterized by the increase of discretionary accruals (Johnson et al 2002). The magnitude of both variable is nearly the same, suggesting that the important of personal tenure and firms' tenure is quite the same.
LABOR as control variable is positive and significant, shows that firms with higher dependence on employee implicit claims are tend to do upward earnings management. This result is consistent with  stated that firms with higher dependence on stakeholder implicit will try to avoid negative earnings surprises because the firm's ability and reputation are assessed by their employees.
BIG4 as control variable is negative and significant, suggesting that firms using Big 4 auditor have lower possibility to do upward earnings management compared with firms using non Big 4 auditor. Notes: *, **, and *** is a significance level on 10%, 5%, and 1% respectively. UPWARD_EM: upward earnings management measured by dummy variable with the value of 1 for firms with positive discretionary accruals, indicating firms' likelihood to do upward earnings management and value of 0 for firms with negative discretionary accruals, indicating firms' likelihood to do downward earnings management.; PTENURE: Personal tenure between Audit Partner and CEO of the clients; FTENURE: Firms Tenure between Public Accountant Firms and Client; LTG: the level of growth company; LIT: companies that have a high risk of litigation, dummy variable 1 is for companies in the industry that has high risk of litigation and 0 otherwise; LABOR: implicit claims against employees; MVOE: market value of equity (in billions of Rupiah); LNMV: company size; BIG4: dummy variable 1 for companies audited by Big 4 and 0 otherwise; LEV: level of leverage; MINING, BASICCHEM, CONSUMPTION, STI, AGRI, PROP, TRANS, OTHER is dummy variable to classify industry, with TRANS as our industry reference.
The regression result of downward forecast guidance model is presented in Table 6. PTENURE and BIG4 are significantly related to downward forecast guidance. PTENURE variable is positive and significance. It shows that longer personal tenure between Audit Partner and Client CEO increases firms' likelihood to do downward forecast guidance. This result shows that that the company is also doing downward forecast guidance. The result is consistent with what was suggested by (Sankaraguruswamy and Sweeney 2005) that management is responsive to analyst careers. Forcing analysts to lower their earnings estimates can help corporate management avoid negative earnings surprises. But if the gap between the estimated earnings of analysts and profits announced by the company is too far, it will be dangerous analyst position. Gaps that are too far can be interpreted by investors as a lack of professionalism of analysts. Investors no longer believe in earnings estimates by analysts, which will then have an impact on the career of the analyst. So to overcome this, management performs a combination of upward earning management and downward forecast guidance. Company management may ask the analyst to lower the company's earnings estimate reasonably, or in other words the profit published by the analyst is not too low to maintain the career of the analyst. Then to meet the earnings estimates from analysts, management will also make upward earnings management, but with lower intensity than without downward forecast guidance.
While BIG4 variable is positive and significant, indicating that firms using Big 4 auditor are more likely to do downward forecast guidance. Other variables are consistent with previous result. In the Indonesian context, the results of this study indicate that although most firms in Indonesia are dominated by family firms with concentrated ownership structures, managers still have an incentive to earn earnings management by avoiding negative earnings surprises if there is low independence of auditors viewed from tenure personal between partner and CEO. The results of this study indicate that with the increasing length of personal tenure managers have an incentive to perform avoidance of negative earnings surprises by upward accrual earnings management and downward forecast guidance.
The regression result of Choice model is presented in Table 7. Table 7 shows that PTENURE and BIG4 are significantly related to the choice of using downward forecast guidance, but not upward earnings management. PTENURE variable is positive and significance. Inconsistent with our hypothesis, the result shows that longer individual of Audit Partner and Client CEO will increase firms' likelihood to do downward forecast guidance without using upward earnings management. BIG 4 as control variable is positive and significant, suggesting firms using Big 4 auditor are more likely to do downward forecast guidance without using upward earnings management.
In overall, consistent with the first two hypotheses our results show that PTENURE is positively and significantly associated with the likelihood of firms do upward earnings management and downward forecast guidance. PTENURE also positively significant associated with the likelihood of firms do downward forecast guidance but not upward earnings management. Our result indicates that the longer personal tenure between Audit Partner and Client CEO increase the likelihood of both mechanisms of earnings management by avoiding negative earnings surprises. For firms that do trade-off between the two mechanism, this study shows that personal tenure increases the likelihood firms to choose downward forecast guidance but not upward earnings management.
Our result support (Sankaraguruswamy and Sweeney 2005) suggesting that firm's manager can influence analyst forecast. Forcing analyst to lower the earnings forecast may help firms to avoid negative earnings surprise, but if the gap between the actual earnings reported and earnings forecast by analyst too huge, it could harm analyst position. Huge gap of actual earnings reported and earnings forecast can be interpreted by investor as analyst error and lack of professionalism. To address this issue, firm's managers and analyst may develop a symbiotic relationship regarding earnings announcements and earnings forecasts (Ho et al. 2010). If managers plan to announce earnings that are lower than analyst forecast, they have an incentive to manage analyst expectations downward, but they try to manage small gap differences between earnings reported and earnings forecast. In addition, to achieve certain level of earnings, firms managers will do upward earnings management, but with less extreme.
FTENURE is significant and positive for upward earnings management. However, we do not find any relation between FTRENURE with downward forecast guidance and firm's choice to do downward forecast guidance, but not upward earnings management. This means that longer audit firm tenure only increase firms' likelihood to do upward earnings management. The same result also goes to LABOR, indicating that firms with high dependence on employee implicit claim are more likely to do upward earnings management only.
BIG4 is negative and significant for upward earnings management, but shows positive and significant for both downward forecast guidance and firm's choice to do downward forecast guidance, but not upward earnings management. It indicates Notes: *, **, and *** is a significance level on 10%, 5%, and 1% respectively. DOWNWARD_FG: downward forecast guidance measured by dummy variable with the value of 1 for firms with recent earnings forecast by analyst (FORE) is lower than expected earnings forecast (E[FORE]), consistent with downward forecast guidance and value of 0 for firms with recent earnings forecast by analyst (FORE) is higher than or equal with expected earnings forecast (E[FORE]), inconsistent with downward forecast guidance; PTENURE: Personal tenure between Audit Partner and CEO of the clients; FTENURE: Firms Tenure between Public Accountant Firms and Client; LTG: the level of growth company; LIT: companies that have a high risk of litigation, dummy variable 1 is for companies in the industry that has high risk of litigation and 0 otherwise; LABOR: implicit claims against employees; MVOE: market value of equity (in billions of Rupiah); LNMV: company size; BIG4: dummy variable 1 for companies audited by Big 4 and 0 otherwise; LEV: level of leverage; MINING, BASICCHEM, CONSUMPTION, STI, AGRI, PROP, TRANS, OTHER is dummy variable to classify industry, with TRANS as our industry reference. that firms using Big4 auditor are likely to use downward forecast guidance and tradeoff between downward forecast guidance and upward earnings management.

Sensitivity Test
We perform sensitivity analysis by replacing the measurement of expected earnings to determine the downward forecast guidance with expected earning using naïve random walk suggested by (Ball and Watts 1972). Random walk model suggest that firms' expected EPS in year t will be the same with EPS of previous year.
The formula of random walk model is as follows: EPS[FORE-RW] t = EPS t-1 EPS[FORE-RW] t is the earnings per share estimation of firm i in the year of t using random walk model. EPS t-1 is the EPS of firm i in the year of t-1. DOWNWARD_FG is valued at 1 if EPS[FORE] t < EPS[FORE-RW]t, which means that current period of analyst earnings forecast is smaller than earnings expectation using random walk model. DOWN-WARD_FG is valued at 0 if EPS[FORE] t ≥ EPS[FORE-RW]t, which means that current period of analyst earnings forecast is larger than than earnings expectation using random walk model. By using the random walk assumption, the result from model 2 and model 3 test show consistent result. PTENURE is positively affect the likelihood of firms Notes: *, **, and *** is a significance level on 10%, 5%, and 1% respectively. CHOSE: Trade-Off Choice Strategy measured by dummy variable with the value of 1 for firms using downward forecast guidance, but not upward earnings management (DOWNWARD_FG = 1 and UPWARD_EM = 0) and value of 0 for firms for firms using upward earnings management but not downward forecast guidance (DOWNWARD_FG = 0 and UPWARD_EM = 1); PTENURE: Personal tenure between Audit Partner and CEO of the clients; FTENURE: Firms Tenure between Public Accountant Firms and Client; LTG: the level of growth company; LIT: companies that have a high risk of litigation, dummy variable 1 is for companies in the industry that has high risk of litigation and 0 otherwise; LABOR: implicit claims against employees; MVOE: market value of equity (in billions of Rupiah); LNMV: company size; BIG4: dummy variable 1 for companies audited by Big 4 and 0 otherwise; LEV: level of leverage; MINING, BASICCHEM, CONSUMPTION, STI, AGRI, PROP, TRANS, OTHER is dummy variable to classify industry, with TRANS as our industry reference. doing downward forecast guidance and the likelihood of firms doing downward forecast guidance without doing upward earnings management.

CONCLUSION
In this paper, we examine the relation of personal tenure between Audit Partner and Client CEO with the underlying ways firm manage their earnings surprise mechanisms. Since firms tend to avoid negative earnings surprise, we use two mechanisms which are upward earnings management and downward forecast guidance. We also examine whether personal tenure affect the option of trade-off behaviour between the two mechanisms.
Our results show that longer personal tenure between Audit Partner and Client CEO is significantly related to firms' likelihood to do upward earnings management and downward forecast guidance. The result shows that the longer the personal tenure between Audit Partner and Client CEO, firms tend to avoid negative earnings surprise by using upward earnings management and downward forecast guidance. For firms that perform trade-off, this study shows that personal tenure positively associated with the likelihood of firms choosing downward forecast guidance without upward earnings management.
Several limitations should be noted from this study. One of limitation of this study is not considering the personal tenure between audit partner and CFO. This study use Audit Partner -Client CEO personal tenure relationship following (Ball et al. 2015) since CEO has important role in involving in auditor appointment decision.
CFO has also important role in the appointment of auditor decision. Further study might also explore the role of personal tenure between audit partner (or key person in the audit team) with CFO. Next, this study limits the measurement of personal tenure and firm tenure from 2009 for observation period of 2012-2014. Further study can extend backward period in order to measure the personal tenure. Third, this study use Litigation as variable control and measured by dummy variable with value one assigned for firms in high litigation industries which are bio-technology, computer, electronic, and re-tailing industry, and value zero assigned for otherwise. In Indonesia, these industries may not have the same level of litigation risk compare to in US or other countries.
Last, despite of using Indonesia as the context, this study does not consider the differences of the ownership structure in the companies. This study does not test whether the relationship between personal tenure and earnings management mechanisms is different between companies with family ownership or between high and low concentration of ownership or between companies that have CEO from family or not. Further study might consider the effect of ownership and whether CEO is part of the family or not in considering the relationship between personal tenure and earnings management mechanisms.