Bank needs for good corporate governance practice increase as the complexity of risk as faced by the bank escalates. The study of risk-taking behaviour is inseparable from psychological studies that assess the individual aspect as the main actor in decision-making. Adams, Almeida and Ferreira, [2] suggest that the high risk will increase the CEO's power to make decisions that impact the occurrence of greater risk. The purpose of this research is to create a model of risk taking behavior in Indonesian banks. 

We propose a model for bank risk taking behavior in Indonesia due to the limited number of previous studies related to the effect of CEO Overconfidence on bank risk taking behaviour. Prior studies showed the effect of CEO overconfidence on bank risk taking, and among them were carried out by Niu [3-4]. Studies on the effect of CEO Overconfidence have been conducted on various variables, including on the investment decisions [5], company acquisition decisions in Australia [6], merger decisions [7], corporate innovation [8], capital structure [9], the accuracy of financial reporting (Schrand and Zechman [10], corporate debt growth [11] and the value of the company's cash holding [12]. With these various considerations, it is still necessary to study the effect of CEO Overconfidence on bank risk taking in Indonesia. 

In Attribution Theory; Self-Serving Biases, it is argued that success is associated with factors from the individual himself when there is a positive evaluation of job performance. However, it will be associated with factors from outside the individual if there is a negative evaluation of the low success of a job. Several previous studies have contributed to self-serving bias theory, including the study of overconfidence and self-serving bias in developing the social and experimental psychological literature [1], Svenson [13], Larwood and Whittaker [14], and [15]. Next, Svenson [13]. The development of studies on bias from overconfidence continues to develop and become a concern for researchers in the field of economics and finance Brown and Sarma [6], Malmendier and Tate, [7], Galasso and Simcoe, [8] Schrand and Zechman [10], Hribar and Yang [16] Aktas, Louca, and Petmezas [12] Niu, [3-4] and Ho, Huang, Lin, and Yen [11].

We used a sample of 1,007 bank-year observations in Indonesia from 2000-2018, and have the banks grouped into 4 categories of BUKU (Commercial Banks Based on Business Classifications). BUKU 1 consists of 114 observations, BUKU 2 has 475 obs, BUKU 3 comprises of 304 obs, and BUKU 4 has 114 obs. Bank categorization is calculated by considering the amount of core capital of the bank in the last year of research (2018). Thecalculation of bank core capital is calculated based on the definition of categorization as stated by the financial service authority regulation (POJK Number 6 / POJK.03 / 2016, 2016).

Our results show various findings a positive and no effect on bank risk in BUKU 1, BUKU 2 and BUKU 3 and BUKU 4. First, we find that CEO Overconfidence has no effect on NPLLoan in banks classification in BUKU 1, BUKU 2, BUKU 3 and BUKU 4. However, CEO Overconfidence has a positive effect on NIMs in banks classified as BUKU 2, BUKU 3 and BUKU 4; on BOPO in BUKU 1, BUKU 2 and BUKU 3; and on LDR in BUKU 1 and BUKU 2. 

Literature Review

Each individual has various aspects of selfish bias where their self-confidence is seen as more capable than the reality or tends to be overconfident. Bradley [17] experimentally examined the success or failure of a task that will significantly affect individuals where the positive effect is in the form of success and the negative effect is failure. If it is extended to the organizational level, it will lead to overly optimistic and risky planning for the future [14]. The study of risk taking has also examined the determinants of individual behavioral motivation. Atkinson, [18] explained how the motives for achieving success and avoiding failure can influence risk-taking behavior.

The concept of overconfidence is used to distinguish overestimation bias against a country's performance measures such as economic growth [7]. Another study shows a new measure of CEO overconfidence related to corporate innovation [8]. Hribar and Yang [16] found that overconfidence CEOs tend to make inaccurate forecasting decisions in the future. Aktas, Louca, and Petmezas [12] reveal that overconfidence CEO affects cash holding value and has implications for company investment policies. Studies that have examined the direct effect of CEO overconfidence on bank risk taking, including Niu, [3] have proven that overconfidence CEO has a significant effect on risk taking, especially for banks with large total assets. Previous studies have examined the effect of manager overconfidence on the sensitivity of the influence of cash flows on investment decisions of manufacturing companies listed on the Indonesia Stock Exchange [19]. Most of the studies on the effect of CEO overconfidence with regard to bank risk taking were also carried out on banks in the United States [3-4] and [11].

 Hypothesis Development and Research Model

It is assumed that each individual has various aspects of egocentric bias [1], where their self-confidence is seen as more capable than the reality or tends to be overconfident [14], by calibrating abilities, competencies and the need to maintain a positive image that indicates a person's behavior to be better [20].

Based on the assumptions that have been developed from the above studies, it shows that the overconfidence behavior has a "better than average" effect due to the presence of more expertise than other individuals. This causes the CEO to become overconfidence, thus affecting company decisions, including a significant effect on the sensitivity of investing in cash flow [5], influencing company decisions related to acquisitions [6], and influencing cash flow and company capabilities to go into debt without having risk, so that it may have an impact on the company's decision to merge [7].

With regard to risk taking, there is a strong tendency for individuals to be more confident in showing their skills than other people, causing greater risk taking [13] and overconfidence behavior that may cause individuals to prefer risk because it is considered to be able to beat the chance of this risk [21].

CEO overconfidence tends to be more willing to take risks when assessing projects than rational managers [22] and more concerned about company-specific risks [23], more daring to take project risks with consequences for doing different things so that they are more promoted to CEO but are too invested and will benefit the company and commit to using high costs because they have stronger and motivated efforts to working on risky projects (Gervais, Heaton, and Odean, [24] Niu, [3-4]. It also develops the assumptions and the results of previous research [23,22] shows that overconfidence CEO has a positive effect on banking risk taking. Based on the assumptions and results of previous research, the following hypothesis is proposed:

Hypothesis

CEO Overconfidence has a positive effect on bank risk taking       

Data and Sample

The data and samples used in our research were taken from the annual and financial reports obtained from ThombsonOneBaker, Bloomberg, and banks’ websites in Indonesia from 2000 to 2018. The number of annual reports obtained depends on the completeness of data availability of bank’s websites. The data requirement refers to the financial data of bank risk (credit risk, market risk, operational risk and liquidity risk) and banking asset data. Other data refers to CEO profiles, bank share ownership, data regarding the composition of the company's board of commissioners, and board of directors from the year. 2000 to 2018.

The population of this research is commercial banks in Indonesia with a total of 94 commercial banks. The sample with complete data to be processed is reported to be 53 banks, classified according to the group of Commercial Banks based on Business Activities (BUKU) which are divided into 6 banks in the category of BUKU 1, 25 bank in BUKU 2, 16 banks in BUKU 3, and 6 banks in BUKU 4. Bank grouping including BUKU 1, BUKU 2, BUKU 3 and BUKU 4 were performed by calculating the amount of core capital of the bank in the last year of research (2018). The calculation of the core capital of banks in Indonesia is calculated based on the reference of the Financial Services Authority Regulation (POJK): Number 6 / POJK.03 / 2016, 2016)

Table 1: Operationalisation of variables

Operationalisation Variables

This study uses 11 variables in total. The dependent variables are NPL, NIM, BOPO, LDR. The independent variables are CEO OVERCONFIDENCE Meanwhile, the control variables are BANKSIZE, BOARDSIZE, OWN1, OWN1-2, OWN1-5 and COM.

Regresi Model Data Panel

Panel data analysis is employed by firstly describing the descriptive statistics information of each variable. Also, in the next step, we test the model specifications. We performed the test using Eviews (Pooled Least Square) to prove the research hypothesis. The statistical models developed in this study are available as follows:

The effect of CEO Overconfidence on bank risk taking without involving control variables: 

Rit = μ + β1Oi,t +εi,t

 (1a)

The effect of CEO Overconfidence on bank risk taking by involving control variables:

Rit = μ + αi + β1Oi,t ∑j=6 βjZˡj,i,,t-1 +εi,t

(1b)

Ri, t represents four variable measures of bank risk taking, namely credit risk (NPLi, t), market risk (NIMi, t), operational risk (BOPOi, t), and liquidity risk (LDRi, t) for bank i and year t. Oi, t is an overconfident dummy variable which has a value of one if bank i is an overconfident bank, and zero if bank i is a non-overconfident bank at time t. Zˡ is the six characteristics of the controlled bank in this study: BANKSIZE, BOARDSIZE, OWN1, OWN1-2, OWN1-5.

Summary of Statistics

First, we present the general information for the dependent variable and the independent variables. The sample of our study consists of banks classified as BUKU 1, BUKU 2, BUKU 3 and BUKU 4. This classification aims to prove the proposed hypotheses 1, 2 and 3 towards bank risk behaviour in the four BUKUs of categorizations. The sample classifications included in the four BUKUs are presented in Table 2.

Table 2: Banks in Indonesia based on Commercial Bank Business Classification (BUKU)

Descriptive Statistics

Descriptive statistics of all the variables we used in this study are shown in Table 3. The risk description of banks from BUKU 1 to BUKU 4 classification can be explained as follows: the average NPL / Nplloan is still below 5%, the mean NIM is above 2%, the BOPO of all banks is seen to be less than 95%, and the average value -The average LDR ratio ranges from> = 75%, to> = 85%. Based on Bank Indonesia decrees, on average bank risk can be categorized as healthy banking [25]. Not all banks in BUKU 1 to BUKU 4 are led by overconfidence CEOs. In detail, 47% of all banks in BUKU 1 were led by CEO Overconfidence, 54% of all banks in BUKU 2 were led by CEO Overconfidence, 66% of all banks in BUKU 3 were led by CEO.

Overconfidence and the largest, namely: 90% of all banks in BUKU 4 led by CEO Overconfidence The highest average size of the board of directors, the largest bank size and the largest average shareholding ranked 1 and 1-2 were in BUKU 4 banks.

Table 3. Summary of descriptive statistics

Dependent variables are NPLLOAN, NIM, BOPO and LDR. OB4 and OB5 are the measures of independent variables: CEO Overconfidence. Control variables, namely: BANK_SIZE is the size of the bank from the log of total assets, BOARD_SIZE is the size of the board (directors and commissioners). OWN1 is the largest shareholding. OWN2 is the share ownership of Rank 1 to 2. OWN5 is the share ownership of Rank 1 to 5. COM is the percentage of independent commissioners. BUKU stands for Commercial Bank based on Business Classification

The Impact of CEO Overconfidence on Bank Risk

We provide the regression results of CEO Overconfidence on bank risk in Table 4. In detail, the findings of hypothesis 1 will be translated into the following two main findings. First, from model 1a, it is found that CEO Overconfidence has a positive and significant effect on NIMs in banks classified as BUKU 2, BUKU 3 and BUKU 4; on BOPO in BUKU 1, BUKU 2 and BUKU 3; and on LDR in BUKU 1 and BUKU 2. Model 1b found that CEO Overconfidence had a positive and significant effect on NPLLOAN only for banks classified as BUKU 4; against BOPO in BUKU 3; and on LDR in BUKU 1, BUKU 2 and BUKU 4. The findings of hypothesis 1 which show that CEO overconfidence has a positive and significant effect on the four bank risks are consistent with the results of Niu [3] and Niu [4] 's research, proving that overconfidence CEO has a significant effect on decision making, especially for banks with large total assets.

Second, from the results of hypothesis testing 1, equation 1.a, it is found that CEO Overconfidence has no effect on NPLLOAN in banks classified as BUKU 1 to BUKU 4; against NIM in BUKU 1; to BOPO in BUKU 4, and to LDR in BUKU 3 and BUKU 4. While the findings from the results of hypothesis 1 test of equation 1.b are that CEO Overconfidence has no effect on NPLLOAN in banks classified as BUKU 1, BUKU 2 and BUKU 3; against NIM in BUKU 1 to BUKU 4; against BOPO in BUKU 1, BUKU 2, BUKU 4; and on LDR in BUKU 3. The findings of CEO Overconfidence that do not affect the four bank risk variables are not consistent with the findings of Ho, Huang, Lin, and Yen [11], where in their findings that overconfidence CEO tends to have a significant effect on Non-Performing Loans ( NPL) as indicated by a more aggressive lending decision making. The findings of CEO overconfidence that do not affect NPLLOAN in all BUKUs can be said that overconfidence reflects overestimation and is assumed to be more resistant or insensitive to risk. This finding supports the findings of Moore, [26] Camerer and Lovallo, [21], Gervais, Simon., Heaton, and Odena, [23,5].[22] And Gervais, Heaton, and Odean [24].

Table 4: Regression of CEO Overconfidence on Bank Risk

Table 4: Continue

*, **, and ***indicate significance at the 10%, 5%, and 1% levels, respectively

Study Limitation

This study still has several limitations. First, the appointment of a CEO is generally selected individuals with a good track record from previous experience as a CEO in banking or a CEO in a non-bank company. CEO practice obtained from two types of business entities with different characteristics, between non-bank and banking companies, will certainly impact the quality of decisions a CEO will take. For this reason, the experience factor has not been studied in more depth, and this could be an opportunity for further research.

This study aims to create a model of risk taking behavior for banks in Indonesia. Specifically, we do it in three stages of testing. We prove the effect of CEO overconfidence on banking risk taking in Indonesia. We used a sample of 1007 banks in Indonesia with observations from 2000-2018 in 4 categories of BUKU (Commercial Banks Based on Business Classifications), namely BUKU 1 with 114 observations, BUKU 2 with 475 obs, BUKU 3 for 304 obs and BUKU 4 for 114 obs

We find that CEO Overconfidence has a positive effect on bank risk although the positive influence of CEO Overconfidence on bank risk in Indonesia is uneven across the four types of risk in banks classified as BUKU BUKU. 2, BUKU 3 and BUKU 4. 

This study provides the following theoretical contributions. The results of this study can prove one of the behavioral biases, namely overconfidence in influencing risk taking. Furthermore, overconfidence theory predicts that CEO overconfidence behavior tends to be stronger in risk taking. The prediction of overconfidence theory can be proven from the positive influence of CEO Overconfidence in bank risk taking.This study found that the strong tendency of overconfidence CEOs to be more courageous in taking risks and the importance of the role of a diverse board of directors will determine the quality of bank risk taking.

1. Adams, B. et al. “Powerful CEOs and their impact on corporate performance.” The Review of Financial Studies, vol. 18, no. 4, 2005, doi:10.1093/rfs/hhi030.

2. Aktas, N. et al. “CEO overconfidence and the value of corporate cash holdings.” Journal of Corporate Finance, vol. 54, 2019, pp. 85–106, doi:10.1016/j.jcorpfin.2018.11.006.

3. Alicke, M.D. “Global self-evaluation as determined by the desirability and controllability of trait adjectives.” Journal of Personality and Social Psychology, vol. 49, no. 6, 1985, pp. 1621–1630.

4. Atkinson, J.W. “Motivational determinants of risk-taking behavior.” Psychological Review, vol. 64, no. 6, 1957, pp. 359–372, DOI:10.1037/h0043445.

5. Bank Indonesia. Regulatory of Bank Indonesia Number. 13/1/PBI/2011. Bank Indonesia, 2011.

6. Bradley, G.W. “Self-serving biases in the attribution process: A reexamination of the fact or fiction question.” Journal of Personality and Social Psychology, vol. 36, no. 1, 1978, pp. 56–71, DOI:10.1037/0022-3514.36.1.56.

7. Brown, R. and N. Sarma. “CEO overconfidence, CEO dominance and corporate acquisitions.” Journal of Economics and Business, vol. 59, 2007, pp. 358–379, DOI:10.1016/j.jeconbus.2007.04.002.

8. Camerer, C. and D. Lovallo. “Overconfidence and excess entry: An experimental approach.” American Economic Review, vol. 89, no. 1, 1999, pp. 306–318, doi:10.1257/aer.89.1.306.

9. Galasso, A. and T. Simcoe. “CEO overconfidence and innovation.” Management Science, vol. 57, no. 8, 2011, pp. 1469–1484, doi:10.1287/mnsc.1110.1374.

10. Gervais, S. et al. “Overconfidence, compensation contracts, and capital budgeting.” The Journal of Finance, vol. 66, no. 5, 2011, pp. 1735–1777, doi:10.1111/j.1540-6261.2011.01686.x.

11. Gervais, Simon et al. Overconfidence, investment policy, and executive stock options. Issue 312, 2003.

12. Hackbarth, D. “Determinants of corporate borrowing: A behavioral perspective.” Journal of Corporate Finance, vol. 15, no. 4, 2009, pp. 389–411, doi:10.1016/j.jcorpfin.2009. 02.001.

13. Ho, P.H. et al. “CEO overconfidence and financial crisis: Evidence from bank lending and leverage.” Journal of Financial Economics, vol. 120, no. 1, 2016, pp. 194–209, DOI:10.1016/j.jfineco.2015.04.007.

14. Hribar, P. and H. Yang. “CEO overconfidence and management forecasting.” Contemporary Accounting Research, vol. 33, no. 1, 2016, pp. 204–227, DOI:10.1111 /1911-3846.12144.

15. Larwood, L. and W. Whittaker. “Managerial myopia: Self-serving biases in organizational planning.” Journal of Applied Psychology, vol. 62, no. 2, 1977, pp. 194–198, doi:10.1037/0021-9010.62.2.194.

16. Malmendier, U. and G. Tate. “CEO overconfidence and corporate investment.” The Journal of Finance, vol. 60, no. 6, 2005, pp. 2661–2700.

17. Malmendier, U., and G. Tate. “Who makes acquisitions? CEO overconfidence and the market’s reaction.” Journal of Financial Economics, vol. 88, no. 1, 2008, pp. 20–43, DOI:10.1016/j.jfineco.2007.07.002.

18. Malmendier, U. and G. Tate. “Behavioral CEOs: The role of managerial overconfidence.” Journal of Economic Perspectives, vol. 29, no. 4, 2015, pp. 37–60.

19. Malmendier, U. et al. “Overconfidence and early-life experiences: The effect of managerial traits on corporate financial policies.” The Journal of Finance, vol. 66, no. 5, 2011, pp. 1687–1733.

20. Miller, D.T. and M. Ross. “Self-serving biases in the attribution of causality: Fact or fiction?” Psychological Bulletin, vol. 82, no. 2, 1975, pp. 213–225.

21. Moore, P.G. “The manager’s struggles with uncertainty.” Journal of the Royal Statistical Society, vol. 140, no. 2, 1977, pp. 129–165.

22. Niu, J. “The effect of overconfidence on the sensitivity of CEO wealth to equity risk.” Journal of Financial Services Research, vol. 38, 2010, pp. 23–39, DOI: 10.1007/s10693-010-0081-8.

23. Santi, F. The effect of manager overconfidence on the sensitivity of the effect cash flow on investment decision. Dissertation, Gajah Mada University, 2016.

24. Schrand, C.M. and S.L.C. Zechman. “Executive overconfidence and the slippery slope to financial misreporting.” Journal of Accounting and Economics, 2012, DOI:10.1016/j.jacceco.2011.09.001.

25. Svenson, O. “Are we all risky and more skillful than our fellow drivers?” Acta Psychologica, vol. 47, 1981, pp. 143–148.