The Impact of Family Control on Dividend Policy : Evidence from Indonesia

S A R I P A T I Vol. 9 | No. 3 ISSN: 2089-6271 | e-ISSN: 2338-4565 9-3.indd 147 3/31/17 5:33 PM 148 International Research Journal of Business Studies vol. IX no. 03 (2016 2017) INTRODUCTION This paper examines how family control affects dividend policy using data from publicly listed firms in Indonesia traded over 2003 to 2009. The finance literature suggests that family control may have both positive and negative impact on firm dividend payouts. On one hand, classical agency theory posits that controlling families may choose to expropriate minority shareholder wealth by preserving firm cash flows that can be misused, thus paying lower dividends (expropriation hypothesis) (Easterbrook, 1984; Jensen, 1986; Faccio et al., 2001). On the other hand, controlling families may opt to build up a reputation of treating minority shareholder well by paying higher dividend payouts (reputation hypothesis) (Gomes, 2000; Myers, 2000). In addition, in general controlling families have a considerable amount of their wealth invested in their firm. Since controlling families do not want to reduce their shareholdings and lose their control, dividend payments are the only possible way for them to obtain an income (family income hypothesis) (Isakov and Weisskopf, 2015). Both arguments predict a positive impact of family control on dividend payouts. Extant studies which have examined the relationship between dividend policy and family control have produced mixed results. For example, Gugler (2003) find that family controlled by families in Austria do not engage in dividend smoothing, choose lower target payout levels, and are less reluctant to cut dividends compared to other firms. Villalonga and Amit (2006), using a sample of Fortune 500 firms, find that family firms in U.S. tend to have significantly lower dividend payout ratios. Faccio et al. (2001) find that East Asian firms have significantly lower dividends, compare to Western European firms. The authors claim that the results indicate that firms operating in countries with weak legal shareholder protection are more likely to exhibit expropriation by controlling families. In contrast, Setia-Atmaja (2010) report that family controlled firms in Australia pay higher dividends than non-family firms. Yoshikawa and Rasheed (2010) who study Japanese firms document higher dividend pay-outs for family firms. Pindado et al. (2012) who examine firms in nine Eurozone countries find that family firms tend to have higher dividend pay-outs and that they tend to smooth their dividends more. Schmid et al. (2010) who study German listed firms find that family firms have higher pay-outs and also a higher likelihood to pay dividends. Finally, Isakov and Weisskopf (2015) indicated that family firms in Swiss display significantly higher dividend pay-outs relative to companies with other ownership structures. Meanwhile, Silva et al. (2004) indicate that, family control in Germany has little impact on dividend policy. Similar result is also reported by Chen et al. (2005) who study firms in Hong Kong. There is a main reason why study of the impact of family control on dividend policy in Indonesia is important. Family controlled firms are prevalent in Indonesia (Claessens et al., 2000). Empirical studies on this issue have been mainly conducted in countries with strong legal shareholder lebih menyukai pembayaran dividen yang rendah dengan tujuan memperoleh arus kas yang bias mereka gunakan untuk kepentingan mereka (the expropriation hypothesis). © 2017 IRJBS, All rights reserved. 9-3.indd 148 3/31/17 5:33 PM 149 Lukas Setia Atmaja / The Impact of Family Control on Dividend Policy: Evidence from Indonesia/ 147 156 protections, well-developed capital market and lower ownership concentration such as U.S. and U.K. This research provides evidence on this issue from a country with relatively weak legal shareholder protections, under-developed capital market and higher ownership concentration. To the best of my knowledge, this research is the first to examine the relationship between family control and dividend policy in Indonesia. Therefore the research results should contribute to the dividend and ownership structure (family business) literature development, as well as provide practical contribution for regulator and investor in capital market. Literature Review The finance literature suggests that dividends can be used as a mechanism to mitigate the conflict of interest between managers and shareholders (i.e., agency problems) because it decrease free cash flows that can be misused otherwise (Easterbrook, 1984; Jensen, 1986; Faccio et al., 2001). Agency theory has a mixed perspective on agency problems in family firms. In one hand, family controlling shareholders can potentially mitigate agency conflicts between owner and manager agency (Agency Problem I) through direct involvement in top management (alignment effect or argument). Indeed, in majority of family firms, family members sit in firm’s boards. On the other hand, there is also an counter argument that controlling families may expropriate minority shareholders wealth (Agency Problem II). Controlling families may represent their own interests, which need not coincide with the interests of minority shareholders. The divergence of interests between majority and minority shareholders may ultimately lead to the expropriation of minority shareholders by controlling shareholders. Illustrations of Agency Problem II are provided by Johnson et al. (2000a) and Johnson et al. (2000b). They describe the transfer of firm resources to controlling shareholders as “tunneling”. Furthermore, Shleifer and Vishny (1997) argue that controlling shareholders, including families, may extract private benefits at the expense of the minority shareholders (expropriation argument). Meanwhile La Porta et al. (2000) posit that dividends can play an important role in mitigating Agency Problem II. Like Jensen (1986), they suggest that dividend payments guarantee a pro-rata payout for all shareholders and reduce corporate wealth from controlling shareholders. As such, dividends are ideal mechanism for limiting minority shareholder wealth expropriation. Therefore, the literature suggests that the presence of Agency Problem II is associated to lower dividend pay-outs in family controlled firms. In contrast, Myers (2000) argues that managers tend to pay dividend pay-outs just large enough to avoid conflicts with shareholders. Furthermore, Gomes (2000) develops this idea and argues that large shareholders such as families may choose to build up a reputation of treating minority shareholders well. His model assumes that controlling families or other large shareholders will not expropriate minority shareholders. The author posits that family members may attempt to pay just enough dividends to minority shareholders to keep them satisfied. The family will build a reputation for treating them well by paying higher dividend payouts. As a consequence, this would limit the misuse of the firm excess cash. Isakov and Weisskopf (2015) argue that this reputation building behavior can also be justified by the «substitution model» of La Porta et al. (2000) that posits that firms tend to pay higher dividends when they plan to issue new equity in the near future. Since family firms tend to have weaker governance than non-family firms due to ineffective internal governance mechanisms, investors may hesitate to buy new stocks offered by family firms. Higher dividend payouts can therefore act as a substitute for the weaker internal 9-3.indd 149 3/31/17 5:33 PM 150 International Research Journal of Business Studies vol. IX no. 03 (2016 2017) governance mechanisms. Thus, the literature suggests that the willingness to build a good reputation by controlling families leads family firms to pay higher dividend payouts. Family ownership has two unique features (Isakov and Weisskopf, 2015; Anderson and Reeb, 2003). First, in general controlling family members have a significant portion of their wealth invested in the firm they control and, second, the family wants to preserve control of their firm. Therefore, controlling family members cannot sell shares to diversify their wealth or to fund their consumption. They should rely on dividend payments from the firms, and this should create a desire for firm higher dividend payout policy. Considering the relatively weak legal minority shareholder protection in Indonesia, I argue that the impact of family control on dividend payouts in Indonesia can be more explained by expropriation hypothesis than reputation hypothesis or family income hypothesis. Expropriation hypothesis/ argument predicts that family control has negative impact on dividend payouts. This leads to the following two hypotheses. Hypothesis 1: Family firms pay lower dividends than their non-family counterparts. Hypothesis 2: Family firms are less likely to pay dividends than their non family counterparts. METHODS Sample The study examines annual panel data over a seven-year period from 2003 to 2009. The sample is based on Indonesia Stock Exchange (IDX) firms. Following prior studies (e.g., La Porta et al., 2000), financial firms (218 firms) are excluded because their dividend policies are influenced by government regulations. The sample is further restricted to firms with annual reports available for 2003 2009 and those firms that are eligible to pay dividends1. This removes the possibility that zero dividends simply result from a firm’s inability to pay dividends. After excluding bank and financial services as well as incomplete data, the final sample comprises of a panel data of 1,945 firm-years from 336 firms. Financial data is obtained from firm annual reports and ownership information is obtained from firm annual reports, prospectus, company’s websites and magazines. Research Model Family controlled firms are defined as those in which the founding family or family member or private individual cont

This paper examines the relationship between family control and dividend policy in Indonesia.There are three possible explanations for the relationship.The expropriation hypothesis predicts that family control has a negative impact on dividend payouts.Meanwhile the reputation hypothesis and the family income hypothesis predict that family control has a positive impact on dividend payouts.Using a panel data of Indonesian publicly listed firms in the period of 2003-2009, the results shows that family control has a significant negative impact on dividend payouts, dividend yields and likelyhood to pay dividends.The results control for other variables that may potentially affect dividend payments such as growth opportunity, debt, profitability, firm size and firm age.From agency theory perspective, the finding is consistent with the argument that family controlling shareholders prefer lower dividends, in order to preserve cash flows that they can potentially expropriate (the expropriation hypothesis).

INTRODUCTION
This paper examines how family control affects dividend policy using data from publicly listed firms in Indonesia traded over 2003 to 2009.The finance literature suggests that family control may have both positive and negative impact on firm dividend payouts.On one hand, classical agency theory posits that controlling families may choose to expropriate minority shareholder wealth by preserving firm cash flows that can be misused, thus paying lower dividends (expropriation hypothesis) (Easterbrook, 1984;Jensen, 1986;Faccio et al., 2001).
On the other hand, controlling families may opt to build up a reputation of treating minority shareholder well by paying higher dividend payouts (reputation hypothesis) (Gomes, 2000;Myers, 2000).In addition, in general controlling families have a considerable amount of their wealth invested in their firm.Since controlling families do not want to reduce their shareholdings and lose their control, dividend payments are the only possible way for them to obtain an income (family income hypothesis) (Isakov and Weisskopf, 2015).
Both arguments predict a positive impact of family control on dividend payouts.
Extant studies which have examined the relationship between dividend policy and family control have produced mixed results.For example, Gugler (2003) find that family controlled by families in Austria do not engage in dividend smoothing, choose lower target payout levels, and are less reluctant to cut dividends compared to other firms.Villalonga and Amit (2006), using a sample of Fortune 500 firms, find that family firms in U.S. tend to have significantly lower dividend payout ratios.Faccio et al. (2001)  In contrast, Setia-Atmaja (2010) report that family controlled firms in Australia pay higher dividends than non-family firms.Yoshikawa and Rasheed (2010) who study Japanese firms document higher dividend pay-outs for family firms.Pindado et al. (2012) who examine firms in nine Eurozone countries find that family firms tend to have higher dividend pay-outs and that they tend to smooth their dividends more.Schmid et al. (2010) who study German listed firms find that family firms have higher pay-outs and also a higher likelihood to pay dividends.Finally, Isakov and Weisskopf (2015) indicated that family firms in Swiss display significantly higher dividend pay-outs relative to companies with other ownership structures.
Meanwhile, Silva et al. (2004) indicate that, family control in Germany has little impact on dividend policy.Similar result is also reported by Chen et al. (2005) who study firms in Hong Kong.
There is a main reason why study of the impact of family control on dividend policy in Indonesia is important.Family controlled firms are prevalent in Indonesia (Claessens et al., 2000).Empirical studies on this issue have been mainly conducted in countries with strong legal shareholder lebih menyukai pembayaran dividen yang rendah dengan tujuan memperoleh arus kas yang bias mereka gunakan untuk kepentingan mereka (the expropriation hypothesis).protections, well-developed capital market and lower ownership concentration such as U.S. and U.K.This research provides evidence on this issue from a country with relatively weak legal shareholder protections, under-developed capital market and higher ownership concentration.
To the best of my knowledge, this research is the first to examine the relationship between family control and dividend policy in Indonesia.
Therefore the research results should contribute to the dividend and ownership structure (family business) literature development, as well as provide practical contribution for regulator and investor in capital market.

Literature Review
The finance literature suggests that dividends can be used as a mechanism to mitigate the conflict of interest between managers and shareholders (i.e., agency problems) because it decrease free cash flows that can be misused otherwise (Easterbrook, 1984;Jensen, 1986;Faccio et al., 2001).Problem II are provided by Johnson et al. (2000a) and Johnson et al. (2000b).They describe the transfer of firm resources to controlling shareholders as "tunneling".Furthermore, Shleifer and Vishny (1997) argue that controlling shareholders, including families, may extract private benefits at the expense of the minority shareholders (expropriation argument).
Meanwhile La Porta et al. (2000) posit that dividends can play an important role in mitigating Agency Problem II.Like Jensen (1986), they suggest that dividend payments guarantee a pro-rata payout for all shareholders and reduce corporate wealth from controlling shareholders.As such, dividends are ideal mechanism for limiting minority shareholder wealth expropriation.Therefore, the literature suggests that the presence of Agency Problem II is associated to lower dividend pay-outs in family controlled firms.
In contrast, Myers (2000) argues that managers tend to pay dividend pay-outs just large enough to avoid conflicts with shareholders.Furthermore, Gomes (2000) develops this idea and argues that large shareholders such as families may choose to build up a reputation of treating minority shareholders well.His model assumes that controlling families or other large shareholders will not expropriate minority shareholders.The author posits that family members may attempt to pay just enough dividends to minority shareholders to keep them satisfied.The family will build a reputation for treating them well by paying higher dividend payouts.As a consequence, this would limit the misuse of the firm excess cash.Isakov and Weisskopf (2015) argue that this reputation building behavior can also be justified by the «substitution model» of La Porta et al. (2000) that posits that firms tend to pay higher dividends when they plan to issue new equity in the near future.Since family firms tend to have weaker governance than non-family firms due to ineffective internal governance mechanisms, investors may hesitate to buy new stocks offered by family firms.Higher dividend payouts can therefore act as a substitute for the weaker internal governance mechanisms.Thus, the literature suggests that the willingness to build a good reputation by controlling families leads family firms to pay higher dividend payouts.
Family ownership has two unique features (Isakov and Weisskopf, 2015;Anderson and Reeb, 2003).To examine the impact of family control on dividend policy, I develop the following model and then estimated by using random effects panel regression.The random effects technique addresses the possibility of a spurious relationship between the dependent and independent variables.This may arise due to the exclusion of unmeasured explanatory variables that nevertheless still affect firm behaviour.Our family control (dummy) variable is relatively stable over the period and consistent with the notion that families generally control their firms for long periods.Therefore, the random effects model is considered more appropriate than the fixed effects one in this study (Kennedy, 2003).Dividend = f (family control, debt, profitability, firm size, growth opportunity, firm age, industry, year) (Equation 1) Dividend is measured by dividend payout ratio, dividend yield and paying dividend dummy.
Consistent with prior research (e.g., La Porta et al., 2000;Faccio et al., 2001), Dividend Payout Ratio is measured as total ordinary dividends divided by net income before extraordinary items.Dividend yield is measured by dividend per share divided by share price.A binary variable that equals one for paying dividend firms and zero otherwise (denoted as paying dividend dummy) is used to differentiate paying and non-paying dividend firms.
Family control is the key exogenous variable of interest, while I control for several firm characteristics explained as follows: Debt (measured by total asset divided total liability) -A negative relationship between debt and dividend is expected.Extant research indicates that dividends and debt financing are substitute mechanisms for reducing equity agency costs (e.g., Rozeff, 1982;Jensen et al., 1992).Profitability

(measured by net income divided by total assets).
A positive relationship between firm profitability and dividend is expected as dividend is paid from net income.Firm Size (measured by the natural logarithm of total asset) -Larger firms tend to have better access to capital markets, which reduces their dependence on internally generated funding and allows for higher dividend-payout ratios (Aivazian and Cleary, 2003).Growth Opportunity (measured by annual sales growth in the last 3 years) -A negative relationship between Dividend and Growth Opportunity is expected as high growth firms may have lower dividend payouts due to their larger investment requirements and a tendency to retain funds to avoid external financing with its attendant costs (Rozeff, 1982;Fama and French, 2001).Firm age (measured by the natural logarithm of the number of years since the firm's incorporation) -A positive relationship between firm age and dividend payouts is expected.Firms that have reached the maturity stage in their firm life cycle tend to pay higher dividends.In addition, a two-way fixed effects model is used to assess variation in the dependent variable due to industry differences3 , while year dummies remove any secular effects among the independent variables.

RESULTS AND DISCUSSIONS Descriptive Statistics
Table 1 presents the descriptive statistics (i.e., means, medians, standard deviations, maximum and minimum values) for the sample.On average, firms report dividend payout ratio and dividend yield of 10.1 per cent and 1.15 percent, respectively.Percentage of firms paying dividends is 30.9 percent.Among family firms, the controlling family holds an average of 47.8 per cent of equity.
Family firms represent 81.4 percent of the sample.*** significant at the 0.01 level ** significant at the 0.05 level * significant at the 0.10 level to debt, family firms have significantly higher debt levels in their capital structure than non-family firms (63.4 per cent versus 52.2 per cent).Family firms are also less profitable (ROA of 0.9 per cent versus 3.7 per cent) and younger (25.1 years versus 33.5 years) than their non-family counterparts.

MANAGERIAL IMPLICATIONS
The study's findings have several important implications for policy makers and corporate decision makers.These implications are not only specific to Indonesia, but could possibly be expanded to other countries with weak legal shareholder protection.For policy makers, the results that family firms tend to pay lower dividends and are less likely to pay dividends could serve to justify initiatives to encourage higher dividend Agency theory has a mixed perspective on agency problems in family firms.In one hand, family controlling shareholders can potentially mitigate agency conflicts between owner and manager agency (Agency Problem I) through direct involvement in top management (alignment effect or argument).Indeed, in majority of family firms, family members sit in firm's boards.On the other hand, there is also an counter argument that controlling families may expropriate minority shareholders wealth (Agency Problem II).Controlling families may represent their own interests, which need not coincide with the interests of minority shareholders.The divergence of interests between majority and minority shareholders may ultimately lead to the expropriation of minority shareholders by controlling shareholders.Illustrations of Agency First, in general controlling family members have a significant portion of their wealth invested in the firm they control and, second, the family wants to preserve control of their firm.Therefore, controlling family members cannot sell shares to diversify their wealth or to fund their consumption.They should rely on dividend payments from the firms, and this should create a desire for firm higher dividend payout policy.Considering the relatively weak legal minority shareholder protection in Indonesia, I argue that the impact of family control on dividend payouts in Indonesia can be more explained by expropriation hypothesis than reputation hypothesis or family income hypothesis.Expropriation hypothesis/ argument predicts that family control has negative impact on dividend payouts.This leads to the following two hypotheses.Hypothesis 1: Family firms pay lower dividends than their non-family counterparts.Hypothesis 2: Family firms are less likely to pay dividends than their non family counterparts.METHODSSampleThe study examines annual panel data over a seven-year period from 2003 to 2009.The sample is based on Indonesia Stock Exchange (IDX) firms.Following prior studies (e.g., La Porta et al., 2000), financial firms (218 firms) are excluded because their dividend policies are influenced by government regulations.The sample is further restricted to firms with annual reports available for 2003 -2009 and those firms that are eligible to pay dividends 1 .This removes the possibility that zero dividends simply result from a firm's inability to pay dividends.After excluding bank and financial services as well as incomplete data, the final sample comprises of a panel data of 1,945 firm-years from 336 firms.Financial data is obtained from firm annual reports and ownership information is obtained from firm annual reports, prospectus, company's websites and magazines.Research Model Family controlled firms are defined as those in which the founding family or family member or private individual controlled 35 per cent or more equity, and was involved in the top management of the firm 2 .A binary variable that equals one for family firms and zero otherwise (denoted as family control is used to differentiate family and non-family controlled firms.My control sample, therefore, comprises non-family firms which include closelyheld firms controlled by non-family blockholders as well as widely-held firms. Random effects regressions are employed to estimate Equations (1) separately primarily to compare the results with previous standard regression studies.When dividend is measured by dummy variable, I use random effect logit regression to estimate the model.

Table 2
* This indicates proportion of firms, rather than the mean proportion for associated variables.† Based on family firms (1.585 firm-year observations).

Table 2 .
Comparison of Family and Non-family Firms: Dividend Policy and Firm Characteristics

Table 3 .
Random Effect Regression Results of the Relationship between Family Control, Dividend Payout Ratio and Dividend Yield

Table 4 .
Random Effect Logit Regression Results of the Relationship between Family Control and Likelihood to Pay Dividends payouts.For example, the issuance of the capital market regulation that forces profitable firms to pay dividends.CONCLUSION This paper examines the relationship between family control and dividend policy.Using a panel data of Indonesian publicly listed firms in the period of 2003-2009, my random effect regression results indicate that there is a significant negative relationship between family control and dividend policy.Specifically, Indonesian family controlled firms have a lower dividend payouts and dividend yields, as well as are less likely to pay dividends.
2001).Future research on this issue may consider specific family characteristics such as generation of family (founder and descendant).