At present, the academic community analyzes the risk-taking behavior of banks, mainly from the four perspectives, like market structure, capital supervision, macroeconomic factors and internal governance. There are a lot of literatures on how internal governance affects bank risk behavior. Gorton and Rosen (1995) [1] first proposed that managers as the controllers of the company and were the direct decision makers of risk-taking, but management was limited by information asymmetry and high supervision costs (Anderson and Fraser, 2000 [2] ). There is a difference in the views of how managers’ salaries affect the level of bank risk-taking. The research of Zhuang Yu, Zhu Jing and Sun Yanan (2013) [3] shows that the higher the salary of manager, the smaller the risk-taking of banks. But Xu Xin and Cheng Chunlin (2016) [4] believes that the salary incentives of bank managers are positively related to bank risks. Bao Huiling’s (2018) [5] study shows that the heterogeneity of management in terms of gender, age, professional experience, and education level will all affect the bank’s risk-taking level.

Kim, Jeong-Bon, Wang, Zheng, Zhang, Liandong (2014) [6] find that firms with overconfident CEOs have higher stock price crash risk than firms with non-overconfident CEOs. Tim R. Adam, Chitru S. Fernando, Evgenia Golubeva (2015) [7] found that managerial overconfidence, which has been found to influence a number of corporate decisions, also affects corporate risk management decisions.

Specifically for banks, Wang Wei (2005) [8] found that overconfidence can lead managers to favor more investment in more volatile projects, especially during the financial crisis. Bank managers’ overconfidence behavior will increase bank risk, and for the larger scale banks, correlation between managerial overconfidence and bank risk are more positive. Xu Jia’s (2018) [9] study shows that the higher the optimism of bank managers, the smaller the negative impact of investor sentiment on bank risk exposure. The empirical evidence of Wang Xiaoxu, Zhang Cheng and Zhao Lijiang (2015) [10] also shows that there is a stable quantitative relationship between managerial optimism and risk-taking levels.

Through the above combing, it can be seen that the management has an important impact on the level of risk-taking of the enterprise. For the bank is the same. Some scholars have studied the impact of management’s overconfidence on bank-taking. Scholars in China already have paid attention to the impact of managerial over-optimism on the risk-taking of non-listed banks, but there are few studies related to listed banks in China. The overall volume of listed banks accounts for a large proportion of the Chinese banking industry. And the listed banks are subject to stricter supervision. Whether these banks have excessive management optimism leads to an increase in the level of bank risk-taking, which is the significance of this article.

According to the combing of the literature, the paper puts forward the hypothesis

H₀: the higher the level of optimism of managers, the higher the level of risk-taking of enterprises. The more optimistic managers tend to underestimate risks and overestimate the benefits.

The sample for this article is 29 commercial banks that have been listed on the A-share market, and the research period is from 2006 to 2017. The data sources are the wind database, the annual bank reports and the National Bureau of Statistics. Excluding those with unsound data, a total of 201 samples from 28 banks were obtained.

The choice of data is primarily due consideration of availability, reliability and integrity. The sample of this paper includes five state-owned banks, seven national joint-stock banks and 16 city commercial banks and rural commercial banks. The deposits and loans of these banks are more than 70% of the bank in China. So the samples are representative.

In order to test the research hypothesis H₀ and analyze the impact of manager optimism on bank risk-taking, this paper establishes the following measurement model:

Risk i t = β 0 + β 1 O p i t + γ Control i t + ε i t (1)

Risk_it represents the bank’s risk-taking. This paper chooses bankruptcy risk as a proxy variable. At present, the agent variables of bank-taking mainly include expected default rate, non-performing loan ratio, bankruptcy risk Z-score and risk-weighted assets as a percentage of total assets. Among them, the data of the expected default rate is poorly available. The calculation of non-performing loan ratio has great subjectivity and operability. And the non-performing loan ratio can not accurately reflect the bank’s risk exposure. At the same time, with the continuous expansion of the bank’s intermediary business and off-balance-sheet business, only considering the risk of the loan business cannot fully reflect the overall risk exposure of the bank. The Z Index is a comprehensive measure of a bank’s profitability, leverage ratio and earnings volatility. And related financial data is easier to collect. Therefore, this paper chooses to use the Z index as a proxy variable for bank risk-taking in the empirical study.

The formula for calculating the Z index is as follows:

Z i t = R O A i t + ( E / A ) i t ( σ R O A ) i t (2)

OP represents manager optimism. For manager optimism and overconfidence, the existing measurement methods include the continuous increase of management shareholding, the growth rate of total bank loans and the proportion of management compensation. Considering the availability and comparability of data, this paper refers to Wu Zhaohui’s (2018) [11] methods to select the executive compensation concentration, that is, whether the ratio of the top three managers’ compensation to the total managers’ compensation is greater than the median.

In addition, based on theoretical analysis and previous research r, the empirical model of this paper also introduces the relevant control variables, namely Control_it. The specific definition is shown in Table 1.

Table 2 shows the stationarity test result of all variables. All variables expect rm2 are stable and can be used directly in the regression. Since the growth rate of the broad money is not stable, the paper used HP Filter to smooth the data.

Table 3 shows the descriptive statistics of the main variables.

The table reports the descriptive statistics of the main research variables under the full sample. It can be seen from the table that the bankruptcy risks of 28

Table2. Stationarity test result.

listed banks range from 0.395 (Pingan Bank, 2009) to 343.1 (China Construction bank, 2013), with an average of 21.99. This shows that the difference in risk-taking between listed commercial banks varies greatly. The over-optimistic average of the dummy variable is 0.401.

Also, it is generally believed that when the correlation coefficient reaches 0.8 or more, there may be multiple collinearity problems. By performing correlation analysis, it can be seen that the model does not have serious multi-collinearity problems. So this paper chooses to use OLS to regress the model.

By doing Hausmann’s test, Prob > chi2 = 0.0615 which means we cannot reject the null hypothesis at the 5% confidence level, so this paper uses the random effects model for regression. Table 4 shows the regression result of the relationship between managerial optimism and bank risk-taking.

This paper examines the sensitivity of bank risk-taking to managerial optimism. The regression results are shown in the Table 3. It can be seen from the table, the bank’s Z-value is significantly negatively correlated with management optimism at a 5% confidence level. The higher the Z value, the more stable the bank, the less risk the bank will go bankrupt. Therefore, the empirical results of this paper show that the more optimistic of management of commercial banks are, the higher risk the banks take. Hypothesis is confirmed.

In addition, in terms of the influencing factors of bank risk exposure, the empirical results of the table also show that the core capital adequacy ratio will reduce the bank’s risk exposure. Capital adequacy can reduce the willingness of banks to take risks through the value of the concession, thereby reducing risk. At the same time, as the loan-to-deposit ratio (DL) increases, bank liquidity reserves decrease and risks increase. The larger the bank, the higher the probability of receiving government support in times of crisis and the lower the risk of bankruptcy. In addition, the more adequate the money supply (RM2), the increased market liquidity, and the lower the bank risk.

t-value in parentheses, *, **, *** indicate the level of significance of 10%, 5% and 1%, respectively.

In order to test the robustness of the research, this paper also uses Zhang Cheng et al.’s (2014) [12] method calculate management optimism. If the bank’s loan growth rate is above 30% of the ranking, OP equals to 1, indicating optimism. If the ranking is below 70%, OP equals to −1, standing for pessimism. And for the rest of banks, OP is defined as 0. The regression results of managerial optimism and bank risk-taking calculated by this method are not significantly different from the regression results above. In addition, this paper uses the weighted risk assets to account for the proportion of total assets to represent the bank’s risk-taking, and there is no significant difference in the regression results either. Therefore, the research results in this paper can be considered to be robust.

By the regression analysis of the annual data of 28 listed banks in China from 2006 to 2017, we empirically analyzed the impact of management optimism on bank risk exposure. The results show that the excessive optimism of the management of listed banks in China will have a significant impact on their risk exposure. Excessive optimism in management can lead to an increase in bank bankruptcy risk.

Since a large number of city commercial banks and rural commercial banks are not listed, their financial data are difficult to obtain or there are a lot of missing. This paper only studies the relationship between the managerial optimism and risk-taking of the listed banks. But for small banks, the lack of supervision makes the management of these banks more likely to be overly optimistic. In the future, if the banks in the sample can be enriched, the research will be more convincing.