“Do family-controlled and financially healthy firms manage their reported earnings? Evidence from Indonesia”

This paper examines whether family-controlled and financially healthy firms practice earnings management. The data collection focuses on non-financial firms listed on the Indonesia Stock Exchange for the fiscal year 2017–2019. Family and financially healthy firms are key predictor variables for predicting earnings management behavior. Jones’s (1991) modified cross-sectional model measures discretionary accruals (the earnings management indicator). This study reveals a negative relationship between family entities and earnings management practices, suggesting that family-controlled firms are more likely to report a higher quality of earnings. This study also documents that family entities with financial difficulties have more incentive to practice earnings management. Additionally, the study indicates that the involvement of a family member in executive positions leads to lower financial reporting quality. Finally, this study reports a nonlinearity association between family share ownership and the magnitude of earnings management. The study’s findings may assist policymakers in considering the costs and benefits associated with various levels of ownership concentration, especially in the hands of family members.


INTRODUCTION
Earnings management has become one of the relatively central research issues in the accounting literature as a result of an increasing number of high-profile accounting scandals across the globe.These unexpected scandals undermine investors' faith in the accuracy of financial reporting and may even obstruct the efficient flow of capital through financial markets (Jackson & Pitman, 2001).Earnings management occurs due to management's inherent discretion in choosing accounting methods for preparing financial reports (Atik, 2009).
Past studies examining the determinants of earnings management have explored various aspects of a company and corporate governance attributes using a sample from different countries.Yet, limited research has been conducted on the relationship between family businesses and earnings management practices (Prencipe & Bar-Yosef, 2011).Family-owned enterprises play an essential role in many parts of the world.They are critical to economic growth and job creation (Claessens et al., 2006).According to Claessens et al. (2006), 73% of firms in Indonesia belong to family members.They contribute 25% of the country's GDP and generate millions of jobs, making them central to Indonesia's economy (PriceWaterhouseCoopers, 2018).Indonesia may be a good example of an emerging economy with some unique institutional frameworks.First, Indonesia has a history of weak corporate governance and audit environments, as evidenced by the many busineses collapes in the 1997 Asian crisis.Second, according to Joni et al. (2020b), most enterprises in Indonesia are owned by family business groups, making it the country in East Asia with the highest concentration of ownership.Finally, several studies (e.g., Chaney et al., 2011;Joni et al., 2020b) claim that many Indonesian businesses have political ties.As a result, the impact of a firm's political connections on its performance, its financing costs, and the quality of its earnings have become the attention of researchers, academics, and policy-makers (Chaney et al., 2011;Joni et al., 2020aJoni et al., , 2020b)).With this in mind, this study poses the following research question: How do Indonesian family-controlled and financially healthy firms engage in earnings management?

LITERATURE REVIEW, AIM, AND HYPOTHESES
This study adopts agency theory to explore how family-controlled and financially healthy firms manage their earnings figures.According to Jensen and Meckling (1976), there are two primary agency relationships: the agent-principal relationship (agency problem type I) and the majority-minority shareholder's relationship (agency problem type II).
Conflicts between managers (agents) and owners (principals) typically arise in widely held companies where ownership and control are segregated.Professional managers are more prone to present financial data that differs from the actual economic transactions that underlie it to increase their benefits at the owners' expense (Wang, 2006).This relationship is significantly different from most developing countries, where many firms are highly concentrated in family ownership.Fan and Wong (2002) state that most firms' equity in many developing countries is held by a small group of shareholders, often with strong family ties.This phenomenon creates a new agency conflict between the majority and minority shareholders.Majority (family) owners can use their controlled power to confiscate the profits of minority owners (Salvato & Moores, 2010).This problem is commonly referred to as a principal-to-principal conflict or agency problem type II (Jensen & Meckling, 1976).
There are two views on how family and non-family ownership interact.The pessimistic viewpoint contends that agency conflicts arise more frequently in family-run companies than non-family companies.Due to their interests, family management frequently expropriates the wealth of minority owners (Faccio et al., 2001;Salvato & Moores, 2010).The optimistic view of family businesses suggests that family businesses face fewer agency issues than their counterparts (Anderson & Reeb, 2003).This study assumes that Indonesian family businesses have fewer agency issues (Anderson et al., 2003;Andres, 2008;Joni et al., 2020a).Anderson and Reeb (2003) note that concentrated ownership in family enterprises reduces the traditional problems of managerial expropriation (agency problem type II), since the family's wealth is tightly tied to the company's success.Family owners put family members in crucial positions, which can potentially monitor and control the company (Demsetz & Lehn, 1985).They also have relatives within the company who are more knowledgeable about the business because they have a more extended period in the industry (Faccio et al., 2001).Moreover, Prencipe and Bas-Yosef (2011) and Andres (2008) suggest that families have invested considerable private capital in the firm, so they tend to be particularly concerned about its survival, hence providing strong incentives to control management activities.These arguments infer that companies under family control are less likely to act opportunistically and, as a result, have higher earnings quality.According to Wang (2006) Moyer, 1990) report that managers of troubled firms systematically take income-increasing actions to keep their jobs and reduce potential intervention by the firms' boards.Likewise, managers of companies at risk of breaching debt covenants have more incentives to practice upward earnings management (Iatridis & Kadorinis, 2009).Similarly, Jaggi and Tsui (1999) argue that companies in precarious financial positions are more prone to report optimistic information in their financial statements.
This paper aims to examine the effect of family-controlled and financially healthy firms on the earnings management practices of Indonesian non-financial listed companies.Consequently, the current study proposes the following hypotheses: H1: Family-controlled firms are less likely to practice earnings management.
H2: Financially healthy firms are less likely to practice earnings management.

Sample selection
This study employs non-financial listed family companies on the IDX for the fiscal year 2017 to 2019 1 .The study identifies family businesses from the Globe Asia Business Magazine article "Family businesses: Maintaining relevance in the modern era" (GlobeAsia, 2019).The study then traced the websites of each group of companies and found 91 firms that continuously published their annual reports throughout 2017-2019.However, 29 entities are financial corporations or do not provide comprehensive information for measuring the variables studied; therefore, the useable sample is 62 corporations or 186 observations.One observation exhibits unique and extreme features (outlier) and is excluded from the research sample.The final sample is 185 observations.

Proxy for the dependent variable
The modified Jones (1991) model's discretionary accruals are used as a measure for the manage-1 To ensure data homogeneity, this study focuses on non-financial firms due to these companies are dominant in the Indonesian economy (Craig & Diga, 1998).
ment of earnings.In line with the existing literature in earnings management (e.g., Francis et al., 2016), this study emphasizes the absolute instead of the real sign of the discretionary accruals.Total accruals must be calculated before constructing the value of discretionary accruals: where TA t = total accruals in year t; ∆CA t = the difference between year t and t -1 in current assets; ∆Cash t = the difference between year t and t -1 in cash balance; ∆CL t = the difference between year t and t -1 in current liabilities; ∆LTd t = the difference between year t and t -1 in long-term debt balance included in current liabilities; ∆IT t = the difference between year t and t -1 in income tax payable balance; DA t = the depreciation and amortization expense in year t.
The TA is then broken down into normal and discretionary accruals using Jones's (1991) modified estimation as follows: , where TA t = a company's total accruals in industry k for year t; TA k,t-1 = a company's total assets in industry k for year t-1; ∆REV k,t = the difference between year t and t -1 in net sales for a company in industry k; ∆AR k,t = the difference between year t and t -1 in accounts receivables balance for a company in industry k; PPE k,t = a company's gross property, plant & equipment balance in industry k for year t; α j , β j , γ j = the industry-specific estimated coefficients; ε j = the error term.
Normal accruals are the fitted values from Equation 2, and discretionary accruals are the residual value.

Proxy for independent variables
This study employs family ownership and financially healthy firms as predictors for earnings management.Following Andersen and Reeb (2003), the present study identifies family firms based on equity ownership proportions and the involvement of family members in executive or supervisory roles.The assessment of the financial health of firms is of particular importance in this study.Numerous bankruptcy models have been previously developed in the literature (e.g., Altman, 1993 where Z shows a firm's propensity to bankruptcy, a firm with a high Z-score has a low probability of going bankrupt.Therefore, per the proposed hypothesis (H 2 ) and the Altman score, this study would suggest the larger (smaller) the Altman index of a firm, the less (more) likely the firm is to practice earnings management.
Table 1 presents the study samples based on the IDX industry sector and firms' financial health.The study uses an Altman Z-score above 2.99 as the threshold for splitting the sample firms into financially distressed and healthy firms sub-samples, as has been well-established in the literature (Altman, 1993;Casillas et al., 2019).

Proxy for control variables
This study controlled for variables that might impact earnings management practices.Corporate governance structures (board independence, board meeting frequency, audit committee) are included in the regression analysis as it is designed to encourage efficiency and promote accountability of resources used by management.In the literature, effective corporate governance is frequently linked to reliable financial reporting and business performance ( According to earlier research (Frankel et al., 2002), companies audited by Big4 auditors report higher-quality earnings than those of their counterparts.

Empirical model equations
Multiple regression of the ordinary least squares is the main statistical method for testing hypotheses.
The regression models are defined as: where DAC = the absolute discretionary accruals measured by Jones's (1991) modified model.FAM = the proportion of publicly traded shares that family members own.ALTMAN = takes a value of one if a company's Altman Z-score is higher than 2.99 and zero otherwise.BOIND = the percentage of independent board members.BOMEET = the annual number of supervisory board meetings.AUDCOM = the total members of the audit committee.BIG4 = takes a value of one if the company auditor is a Big4 accounting firm and zero otherwise.ROA = the net income to total assets ratio.AGE = the years since the company was founded, expressed as a natural logarithm.it = the firm i in year t.YEAR FIXED EFFECT, INDUSTRY FIXED EFFECT = the fixed effects for different years and industries.

Correlations
This study uses Pearson's correlation matrix to test for multicollinearity across the variables.Table 3 reports that correlation results fully support the study hypotheses.As hypothesized, the finding shows that family ownership (FAM) and financially healthy firms (ALTMAN) are significant negative correlated (both at p<0.01) with DAC (a proxy for earnings management).The results also report a low correlation coefficient (0.319) between the independent variables.All variable correlation coefficients are within the crucial range of 0.80 (Cooper & Schindler, 2003).Thus, the regression models show no evidence of multicollinearity across variables.Moyer (1990) note that managers of troubled firms have incentives to engage in income-increasing accounting practices to maintain their financial performance.

RESULTS 2
Regarding control variables, this study finds that BMEET, BIG4, and ROA help explain earnings management behavior.Coefficients on BMEET and ROA are positive, while BIG4 is significantly negative across all models.These findings confirm that compared to companies that use the audit services of non-Big4 firms, the level of earnings management is much lower among companies audited by the Big4 auditors.Interestingly, this study reveals that companies are more likely to manage their earnings figures if they (1) hold board meetings more frequently and (2) have better financial performance.

Robustness check
Previous studies document that earnings management varies across industries and in different periods (Al-Shaer & Salama, 2015; Salehi et al., 2019).Therefore, this study undertakes a further analysis (see Table 5) that considers industry and year effects.The results of the Hausman test suggest adopting fixed-effect estimates to account for the industry and year-specific characteristics.
The multiple regression analysis performed utilizing the fixed effect test (Table 5), and the main findings in Table 4 are generally consistent.A difference is that Table 5's explanatory powers (adjusted R 2 ) are marginally higher than those found in Table 4's data.

Additional analyses
This study performs two additional analyses to increase the main findings' reliability.First, the study considers the role of the family members as one of the family business characteristics.The impact of active versus passive family involvement in the enterprise is therefore examined in this study.Family members with senior management roles can more easily align a company's interests, enhancing the company's performance and reputation (Anderson & Reeb, 2003;Davids et al., 1997).However, they also can negatively affect a firm's performance.Family management tends to ex- propriate minority shareholders' wealth for their interest (Salvato & Moores, 2010).This study uses two separate active family controls to examine the role of family directors or executives (FAM AS DIR) and the family involvement on the supervisory board (FAM AS BOD).If a family controls the company and one or more family members hold director (executive) positions, the value of FAM AS DIR is one; otherwise, it is zero.If a family controls the company and one or more of its members serve in a supervisory capacity on the board, the value for FAM AS BOD is one; otherwise, it is zero.Table 6 summarizes the results of the additional test.
Panels B and C report that coefficients on FAM AS BOD are positive but statistically insignificant, inferring that family members on the board of directors have no significant impact on earnings management behavior.The findings in Panels A and C demonstrate that the coefficients on FAM AS DIR are positive and significant at p<0.01 and p<0.05, respectively.These results imply that family members sitting in management or executive positions enhance earnings management and, thus, lower financial reporting quality.A possible explanation is that family management experiences hardship in financial performance.Further analysis (not tabulated) confirms that the financial performance of family managers is significantly lower than the profitability of firms managed by non-family members.This finding aligns with other research (Frankel et al., 2002;Kothari et al., 2005), which suggests that less profitable companies are linked to more aggressive earnings management.
Second, the current study considers the impact of nonlinearities in the effects of family ownership on earnings management practices.Familyowned enterprises may have favorable and unfavorable consequences on their performance (Anderson & Reeb, 2003).The performance of a company improves with greater ownership concentration.Yet, the relation between the two variables might be negative at lower share ownership levels.These two hypotheses predict a nonlinear relationship between family-controlled firms and the magnitude of earnings management.Table 7 displays the test results for nonlinear relationships between family firms and management of earnings behavior.
The results reveal that family ownership (FAM) has a positive coefficient, while its square (FAM-Square) has a negative value (at p<0.01).These findings suggest a nonlinear relationship between family share ownership and the management of earnings practices.At low ownership levels, family owners have less power to control management actions and, thus, increase the magnitude

CONCLUSION
The paper mainly explores whether family-controlled and financially healthy firms practice earnings management and, thus, reduce earnings quality.Using a sample of 82 non-financial companies listed on the IDX from 2017-2019, the study documents a significant negative association between family ownership and earnings management.This finding infers that family-controlled businesses are less likely to manage their earnings.The study also shows that financially healthy businesses tend to practice earnings management less.In other words, the result documents that companies without a propensity to experience financial issues exhibit lower levels of earnings management.
In conclusion, the study's main findings indicate that family-controlled and financially healthy businesses report higher-quality earnings.Additionally, this study finds that the impact of the family executive members on the earnings management measure is positive and statistically significant.The result infers that the involvement of family members in the management or executives leads to lower financial reporting quality.Finally, the study indicates a nonlinear association between family share ownership and the management of earnings behavior.In the first stage (up to a 40.6% ownership concentration), the increased family equity causes earnings quality to decline.Still, family businesses begin to report higher-quality earnings as family ownership exceeds this cut-off point of ownership.
The findings of this paper have implications for policymakers who appear to control the ownership of company shares by the family members as such matters might reduce the practices of earnings management.The study argues this concern can be inappropriate due to families' limited ability to maximize control within the firm, which could provide limited benefits.

Table 1
shows that most firms (53 observations or 28.65%) in the sample are in the Property, real estate & building classification.At the same time, the smallest percentage is companies in the Agriculture sector and the Infrastructure, utilities & transportation industry with nine observations (4.86%), respectively.The largest proportion of firms classified as financially distressed is listed in the Property, real estate & building group, with 45 observations or 42.45% of the total distressed firms.In comparison, three observations of the listed firms in the Consumer goods industry were assigned a numeric value indicative of financial difficulties.Interestingly, all firms in the Miscellaneous sector are more likely to experience bankruptcy.

Table 1 .
Arora & Dharwadkar, 2011; Martinez-Ferrero et al., 2020).To control the potential Sample by the IDX industry sector and financially healthy versus distressed firms

Table 4
2We also perform the tests using an alternative proxy of earnings management (performance-matched discretionary accruals purposed byKothary et al. (2005).These results are qualitatively the same as those reported in Table4.

Table 3 .
Pearson correlation matrix

Table 5 .
Fixed effect regression results

Table 6 .
Active family control and earnings management The inflection point is 40.6%.If family equity increase beyond this cutoff point, family owners achieve excessive power to control management activities, causing a lower level of earnings management.In summary, the quality of earnings declines as family ownership up to the inflection point.However, when family equity rises above this ownership threshold, family businesses begin to generate higher-quality earnings.

Table 7 .
Nonlinearities between family ownership and earnings management Note: * and ** indicate significance at 1% and 5% levels, respectively.