DOES CEO’S HUBRIS AFFECTING DIVIDENDS PAYOUT?

The purpose of this research is to examine the effect of CEO’s Overconfidence (KDB) on nonfinancial firm’s dividend that listed in Indonesia Stock Exchange (IDX) from 2004 to 2013. It is being said that almost every person has overconfidence, similarly with a CEO of a company, particularly because of the scale of power that they hold. CEO with overconfidence has a tendency to lower their dividend payout because they are driven by their subjective consciousness of the possibility of getting an investment opportunity in the future. The samples of this study are 327 companies listed in Indonesia Stock Exchange (IDX). The model is using a logit regression. The result shows that company being lead by CEO with overconfidence will have a tendency to hold or reduce their dividend payout. Info Article History Article: Received September 2015 Approved November 2015 Published March 2016


INTRODUCTION
In the context of the company, managers are the individuals that have the biggest exposure to biases such as overconfidence (Li & Tang, 2010).There are several reasons why managers in a company have overconfidence.First of all, an individual will be very confident when they feel they have a strong control of the outcome (Langer, 1975).Second, individuals would also be vulnerable to this tendency to overestimate a result if they have a very high commitment to achieve it (Weinstein, 1980).Third, individuals will assess themselves highly or much better than any individuals around them when a comparison is hard to find (Ben-David et al., 2007).
At least, there are two reasons why the relationship between managerial bias and dividends are very tempting to be tested.First, although many researches related to this dividend decision have already done, but results from these studies in the past have not been able to explain why companies distribute dividends and why investors like dividends (Deshmukh et al., 2013).In its development, there is a wave of researchers who doubt about the assumptions made by traditional finance such as the theory of signalling like Benartzi et al. (1997), Grullon et al. (2005), Prasetyo (2013) as well as Yulianto (2013) that found that changes in dividend, cannot be used to predict the increasing future profit of a company or the increasing in the company' s operating performance.
The second reason is there are little numbers of research that examined the relationship between overconfidence and dividend.Early literature of overconfidence only focus on testing of company decisions like the investment decision, merger and acquisition decisions as well as funding decisions, as conducted by Malmendier and Tate (2005a, 2005b, 2008), Ben-David et al. (2007), Campbell et al. (2011) and Malmendier et al. (2011).
In their research, Li and Tang (2010) said that there are actual variables or factors moderating the relationship between overconfidence of CEO' s with managerial decisions.There are two interacting variables that will be tested in this study that are the life-cycle hypothesis and managerial discretion.According to Grullon et al. (2005), the company' s dividend policy contains information about the changes that occur in the life-cycle of the company.Lease et al. (2000) proved that the company' s dividend policy will follow the company life-cycle stages.The research of Denis and Osobov (2008) confirmed the results of research from DeAngelo et al. (2006) that stating the dividends policy of American will follow the financial age from the company itself.Financial age as a proxy for the age of the company is then tested by Von-Eije and Meggison (2008) as well as the Bulan et al. (2007).The result was supporting the hypothesis that age may explain the company' s dividend policy, which means that the assumption of the life-cycle hypothesis was proved.
The second interaction variable is managerial discretion.Managerial discretion, which is defined as a form of discretion which possessed by CEO to make a decision in a company, could explain the role of CEO of a company is stronger than other CEO in another company (Hambrick & Finkelstein, 1987).When the CEO has strong discretion, then the impact of the decisions they take for the company will also be stronger (Finkelstein and Hambrick, 1990;Finkelstein and Boyd, 1998).A simple conclusion can be drawn that the managerial discretion may become an important factor that moderating the relationship between overconfidence of CEO and dividends.
Research on managerial bias and managerial discretion are more often in the context of developed countries (Hambrick & Finkelstein, 1987;Malmendier and Tate, 2005a;Malmendier and Tate, 2005b;Ben-David et al., 2007;Goel & Takhor, 2008;Deshmukh et al., 2013).Not many researches that test the managerial discretion in developing countries such as in Indonesia.Whereas companies in the developing countries have the possibility of having different problems or rules with companies in developed countries.This may caused by the difference in economic conditions.The market in developing countries does not have strong regulatory and corporate governance which is different than developed countries.It could ultimately affect dividend policy that applies in developing countries.

Hypothesis Development
Confidence is the probability or subjective degree of belief associated with what we ' think ' will happen (Kahneman & Tversky, 1982).Subjective probability is an individual subjective on the accuracy of decisions taken.Confidence usually measured by comparing the subjective probability of individuals with objective probability/accuracy.When these two factors are not balanced, the miscalibration can cause the individual to have less self-confidence or excess (under or overconfidence).Miscalibration is the cause of the individual over or underestimates the truth or accuracy of something they believe will happen.
According to DeFinetti (1962), miscalibration would be an issue when the subjective beliefs cannot be compared with the accuracy of the reality of that happening because to be able to know that there is bias, and then the discrepancy should be seen between the subjective confidences and objective accuracy.If a person' s subjective confidence far outweighs the objective accuracy, then there was such a thing called as overconfidence bias effect.

Dividend Policy in Indonesia
The usage of net profit of companies in Indonesia has been set in some Acts.Study on the Act that governing the use of net profit needs to be done in order to know for sure whether the CEO of a company have the authority of the ideas of policyrelated net profits, which will be related to discretion belonging to each of the CEO.In Indonesia, there are two references to the Act that can be used: Limited Companies Act (PT) No. 40 year 2007 and ACT of State-owned enterprises (SOEs) No. 19 of 2003.
Net profit that earned by a company partly used to fulfil a mandatory reserve fund established by the general meeting of shareholders.The rest of the net income will be distributed to shareholders in the form of a dividend unless the general meeting of shareholders determines the decisions of others.The condition that must be met for a company can distribute dividends is a company must have positive net income.The verse states indirectly that companies have no restrictions related to net income allocation decision because in the act of decisions of the general meeting of shareholders is held by so that it is not impossible the company not to share the profit with the shareholders.This principle is somewhat different in a company owned by the State, or in the context of Indonesia referred to as BUMN (State-owned enterprises).A company owned by the State because all or part of the capital is owned by the State.The Minister may specify that some or all of the net profit will be used for dividend distribution to the owners of capital, or other divisions such as allowance for directors and supervisory board, employee bonuses, Social Fund reserves and others, or the placement of the net profit for the expansion of BUMN.
Malmendier and Tate (2005a) use the compensation received by executives in the form of a personal portfolio choices.The basic idea is the rational manager will execute stock options they have in the position of the in-the-money.Malmendier and Tate (2005a) argues that managers who hold their shares the option when the condition is in-the-money is a manager who is optimistic about the prospects of the company and feels confident that they can gain a greater advantage in waiting and do not do exercise those options.Malmendier and Tate (2005b) test with two goals, the first is to prove that managers tend to be biased on the company, and the second to test another proxy for overconfidence, using the perception of the media against the CEO of a company.Using the method of a survey to measure the level of investor confidence, Ben-David et al. (2007) give a new measurement tool using the experimental method.In his research, Ben-David et al. (2007) found that the confidence intervals are narrow, indicating a high level of confidence.Measurement instrument for testing the overconfidence is growing from time to time.Starting from Malmendier and Tate (2005a), which measures the overconfidence use the stock options owned by CEO, followed by Malmendier and Tate (2005b) using information from the media perception of the image of CEO, then measurements using survey methods such as that done by Ben-David et al. (2007).And the most recent is the researchers Campbell et al. (2011), they measure the confidence level using a level of investment made by the CEO as a company agent.In his research, Campbell et al. (2011) testing the assumption that says, the board of director more often to dismiss CEO or manager who has excessive confidence as well as excessive diffidence.
Manager or CEO with these characteristics tend to overestimate (underestimate) the accuracy of the information, thus causing overinvest (underinvest) on a project, which ultimately lowers the value of companies.The conclusion of Campbell et al. (2011) are consistent with the research of Goel and Thakor (2008) as well as Hackbarth (2008) that the level of bias in the medium, still able to provide a benefit to the company.

CEO Overconfidence to Dividend
CEO' s overconfidence is caused self-attribution bias (Baker et al., 2004) and the illusion of control bias (Langer, 1975).The first bias led to overrate his ability in taking decisions because of the successful experience in the past while the second causes the bias-CEO overestimate his ability to control the outcome of an event in the future.The biases can affect the decision that made by CEO.
Research conducted by Wu andLiu (2011) andBen-David et al. (2007) found that CEO's overconfidence, will lead the company lowering or even not distribute dividends, because they are convinced that the company has the opportunity to invest in the future, they believe can make a bigger profit from investments, although such investment decisions can destroy the value of the company.Cordeiro (2009) also mentioned that the CEO who has overconfidence would rather follow the pecking order theory assumptions; they believe that the use of external funds, such as the issuance of equity, would cause the investment made has a negative NPV.That assumption is also supported by research conducted by Deshmukh et al. (2013), which found a negative relationship between excess CEO' s confidence against dividends.H1: CEO' s overconfidence effect negatively to dividends

CEO's Overconfidence and Managerial Discretion
Life-cycle theory says that growing firms need more funds to make investments.One of the sources of funding for the company is retained earnings.The greater the investment needs of the company, the higher the company reduced the dividend distribution.Contrary conditions at a time when the company is already in the mature condition, where companies have investment opportunities in the stagnant conditions (DeAngelo et al., 2006).Hambrick and Finkelstein (1987) added that on condition of high managerial discretion, Board of Directors hands over the company decision to CEO, as a party that is considered to have more detailed information on good and bad decisions related to the company' s performance.On the research of Haleblian and Finkelstein (1993), managerial discretion has another term which is CEO dominance, defined as the ability of individuals to exert (impose, implement) his desire.H2a: Negative influences of CEO' s overconfidence on dividends will be weaker if the company is in the mature condition.

BUMN (State Owned Enterprise)
One of the things that can reduce the discretion of managers is State-owned Enterprise (BUMN) (Tsui, 2007;Li & Tang, 2010).In Indonesia, a company owned by the State company known as BUMN (Badan Usaha Milik Negara).BUMN companies making decisions taken by managers based on not only the interests of the company itself, but also consider the interests of the government as the owner.So even though the manager acts as a decision-maker, but decisions taken should consider an order from the Government as the owner.
Ben-David et al. (2007) say that CEO who owns overconfidence can make a bigger profit from investments they do for distribution to the owners of the company, although such investment decisions can destroy the value of the company.CEO on the BUMN cannot be as easily as it executes the decision because the State as an owner should consider.These conditions will cause the CEO lost the motivation to pursue investment opportunities they have.H2b: Negative influences of CEO' s overconfidence on dividends will weaken in BUMN

Data and Sample
The data used in this study is secondary data obtained from the Indonesian Capital Market Directory (ICMD) as well as annual reports of non-financial companies listed on the Indonesia stock exchange from 2004 to 2013, and finally obtained samples as much as 327 companies.
Dividends (DIV) as the dependent variable in this study using a dummy variable, where the value will be 1 if the company share the dividend and value of 0 when the company does not share the dividend (Cordeiro, 2009).Two interaction variables that used as a proxy of maturity and BUMN in this research are based on research DeAngelo et al. (2006) as well as Li and Tang (2010).Maturity is measured using the ratio RE/TE or RE/TA where RE is retained earnings or income withheld, TE is the total equity and TA is total assets.The larger the value of RE/TE or RE/TA, the more matures the company.The proxy of BUMN (OWN) uses a dummy variable as corresponding to the research of Li and Tang (2010).If the enterprise is state-owned company, then the sample will be given a value of 1 and value of 0 when otherwise.
The control variables used are the variables that are considered to have its influence on the dividend policy.The control variable is obtained from the model of Fama and French (2001) as well as the addition of Huang et al. (2011).Company size (SIZE) in this study, measured using natural logarithms of the total assets of the company.This variable is expected to associate positively with dividend as the dependent variable.Profitability (PROF), i.e. the ratio between operating profits compared to total assets.Tobin' s Q (Q), measured from the market value of equity coupled with a book value of liabilities divided by the book value of assets.Top1 is a percentage of shares owned by the owners of the largest stocks.Group (GR) is the dummy variables which value 1 if the parent company is a conglomerate, and 0 if otherwise There are two proxies of CEO' s overconfidence which will be used in this research, first developed by Richardson (2006), and the second proxy used by the Scrand and Zechman (2012).The proxy will be tested separately against the dividends to test the robustness of the CEO' s overconfidence proxy.Both the proxy will hopefully have a negative effect on the dividends.
Proxy modelled by Richardson (2006) used to see the level of investment from a firm in line with overconfidence measurements in this study.In accordance with research of Malmendier and Tate (2005) as well as Campbell et al. (2011) stated that there is a relationship between CEO' s overconfidence with investment decisions, theoretically and empirically.So a simple conclusion can be drawn that the level of investment of the company contains related information of CEO' s overconfidence.Model of Richardson (2006)  From the regression model above, the residual value will be obtained from each sample.According to Richardson (2006), a sample of companies that produce positive residual will fall into the category of companies' overinvestment; while for a sample of firms that have negative residual will go in a group of enterprises that underinvestment.The company entered the category of overinvestment will be worth 1, which means it is assumed to have the CEO with overconfidence while for companies categorized as underinvestment will be worth 0.
Schrand and Zechman (2012) using two indicators of overconfidence on firm-level related to investment decisions and funding decisions made by the company, which is already proven in the previous empirical studies.Schrand and Zechman (2012) assuming if both indicators are met, then the company is considered to have a CEO with overconfidence.So the overconfidence would be worth 1 and overconfidence would value 0 when both indicators are not met.Individual will be very confident when they feel they have a strong control of the outcome.Individuals will assess themselves highly outcomes.

Hypothesis Testing
The models are using a logit regression.Data is the data panel.Models to test hypotheses 1 are as follows: And below is model to test the hypothesis 2a and 2b Explanation: DIV it = dummy variable of the dividend, val ue of 1 when the company's share dividend and 0 if otherwise.KDB it = a dummy variable of CEO's overconfidence, which proxy is generated from overinvestment, valued at 1 when the company entered the category of overinvestment, value 0 if otherwise.RE/TE it and RE/TA it = the maturity of the company proxy.OWN it = variable for this type of company ownership dummy of value 1 if the company is BUMN, and the value 0 if otherwise.

Descriptive Statistics
The population in this research is all the nonfinancial companies listed on the Indonesia stock exchange (IDX) of the years 2004-2013.The data on this research is unbalanced data panel with some cross-section as much as 327 with ten years of observations from the 2004-2013 so that the resulting observations as much as 3270.The data will be processed using a logit regression because the dependent variable in this study is binary data.
The primary variables in this study (KDB1 and KDB2) have an average value of 0.1 and 0,228.Because the average value of both variables was below 0.5, then it can be concluded that more than half of the company' s CEO examined, which in this context is a registered company in IDX, have rational behavior, and most of them don't have such irrational behaviour of overconfidence.DIV variable has a value of 1 if the company share dividends and 0 if it does not share the profit.Then it is obtained an average value of 0,233 which means that more companies do not share dividends than share dividends.
Average of RE/TE is -0,061 while RE/TA is-0.17.The larger the value of RE/TE and RE/TA the more mature the company because the value of the business and retained earnings are high (Von-Eije & Meggison, 2008).The sample of this research showed that the average value of a company owned by the State is only 0.042 shown with variables OWN.The next variable is the PROF with an average of 0,069, SIZE with an average of 13,724 and Q with an average of 1,421.And the last is a variable of TOP1, which shows the percentage of ownership that is owned by the biggest owner stock, has the average of the 0,56 and GR variable, which indicates the status of the company that owns the parent company of a group of conglomeration with an average value of 0.142.

Hypothesis Testing
Hypothesis 1 shown in model 2, where KDB1 has the negative effect of dividends that indicated by a coefficient of KDB1, which is negative and significant at the 5% level of confidence so that the test results support the H1 that CEO' s overconfidence effect negatively to dividends for model KDB1.
H2 the interaction between KDB1 with maturity and country ownership of the dividends can be seen in the next model.Model 3-7 test the influence of the interaction of KDB1 with a maturity of dividends or hypothesis 2a.Variable of maturity in this research is measured using two proxies: RE/TE and RE/TA.In thus model can be seen that RE/TE and RE/TA are positive and significant for the dividend, which means the company will divide the dividend when mature conditions because they don't have a lot of investment opportunities.Hypothesis 2b that tests the interaction between KDB1 and OWN against dividends shown in models 5 and model 8.The results of the testing were expected to demonstrate in support of allegations that the company owned by the State will give discretion or less authority to the CEO at the time of decision-making, especially the decision to divide the dividend due to differences between the interests of the State and the interests of the CEO as an agent of the company.With a note that, interaction between KDB1 and OWN to DIV on the model of 5 or 8 is not significant.Table 3. shows some models indicating the results of the hypotheses testing.Model 2 indicates that the negative effect of the KDB2 on DIV, it is indicated by the coefficient of KDB2 which is negative and significant at the 5% level of confidence so that the test results support the H1 that CEO' s overconfidence effect negatively to dividends for model KDB2.The interaction of KDB2 with the maturity towards DIV (dividend) is shown by the model 3, 4, 6 and 7. Variable of maturity is measured using two proxies: RE/TE and RE/TA.Only two models in model 3, 4, 6 and 7 that showed that maturity has a significant positive influence towards DIV, namely model 4 and model 7 where both models using a proxy RE/TA to test maturity against DIV, while model 3 and 6 which uses RE/TE to test maturity against the DIV does not show significant results.This result is inconsistent with a previous test by using KDB1, where both RE/ TE and RE/TA showed positive results against DIV.On the model 5 and 8 can be seen that the hypothesis 2b is proven, where the negative influences of the KDB2 towards DIV will be weaker in the conditions of the OWN.The results indicated by the value of the coefficient of the interaction of KDB2 and its OWN greater than the coefficient of KDB2 towards dividends.On the model 5, the interaction of KDB2 and OWN against DIV has a coefficient of 0,269 and KDB1 coefficient is-0,497.While on the model 8, the interaction of KDB2 and OWN towards DIV has a coefficient of -0,173 and coefficient of KDB2 towards DIV is -0,309.
From hypothesis testing using the KDB2, there are some things that are not consistent with hypothesis testing using KDB1.Including a few variables that showed significant results when tested using a KDB1, showed no significant results when tested using KDB2.But neither KDB1 nor KDB2 is consistently able to prove that hypothesis 1, 2a and 2b are acceptable.

CONCLUSION AND RECOMMENDATION
Companies led by the CEO who have overconfidence proved to have a negative influence on the company' s dividend.The results of this research are consistent with findings from Cordeiro (2009) as well as Deshmukh et al. (2013) stating that companies that do not share dividend are the company led by CEO who has overconfidence.
Hypothesis 2a stated that the negative influences of CEO' s overconfidence towards dividends will be weakened if the company is in mature condition.Testing the effect of the interaction of KDB with the maturity towards dividend, either use KDB1 or KDB2 as a proxy can prove that hypothesis 2a is supported.The influence of KDB towards dividend is weaker if the company is in the condition of mature.It is caused due to a mature company does not have much of a chance when compared with the company' s investment in the growth stage.This result consistent with research from Lease et al. (2000) and DeAngelo et al. (2006).
The hypothesis 2b tested the influence of managerial discretion to the relationship of overconfidence with dividends.The test results indicate that the influence of overconfidence against dividend is weaker and not significant if the company is led by CEO who have overconfidence is a company owned by the State.These results are consistent with Chen et al. (2011) and Rasheed et al. (2012).
For further research, it is necessary to test the other overconfidence (KDB) constructs, because there is still no fixed construct to measure overconfidence.From some constructs of overconfidence tested, then it would be found one model to measure the overconfidence which is the most robust.
is as follows: t -Sales of PPE t )-(De preciation & amortization t ).Growth = measured by Tobin's Q (market value/book value of equity).Lev = measured using total debt divided by total assets.Cash = ln (cash + short-term invest ments).Size = ln (total assets).

Table 2 .
KDB1 CEO to Dividend