Military service can change the behavior of people in different manners that could influence behavior and decision after they become CEOs in the future (Benmelech and Frydman, 2015). In contrast, psychological literature shows that MTB CEOs are related to boldness, aggressiveness, and adventuresome behavior (Elder, 1986). A couple of ongoing studies found that organizations operating by CEOs with MTB are less inclined to perform better (Benmelech and Frydman, 2015; You et al., 2020). For instance, Benmelech and Frydman (2015) find that CEOs with an MTB put less in Research and Development (R&D) and follow less organization performance. Continuously, You et al. (2020) show that organizations led by military CEOs are more likely to make small R&D expenditures than their non-MTB counterparts in China. This proof can be clarified by the logic that military training and service values subordination to political authority, obligation, commitment, and kindness, prompting corporate literature that inspires little risk-taking and conservative investment behavior (Franke, 2001; Benmelech and Frydman, 2015). Therefore, we hypotheses that:

However, when an organization cannot use the available resources efficiently, it may negatively affect organizational performance. Political intervention of a CEO in the organization and a weak administrative ability of politicians can be decrease organizational performance (Boubakri et al., 2012). For example, Asquer and Calderoni (2011) found a negative impact of political associations on Italian corporation performance. Similarly, Bertrand et al. (2007) show that organizational and political connection prompts a negative impact on ROA. A study by Faccio (2010) indicates that politically connected organizations have lower performance than non-political connected firms. Crook et al. (2011) found that the effect of Politically connected CEOs (PCCEOs) is low investment effectiveness and organizational performance. Furthermore, Boubakri et al. (2008) indicate that politically connected organizations show a weak financial performance compared to their non-connected counterparts. Finally, Disli et al. (2013) found that politically associated CEOs in Turkish corporations slow down investors' motivating monitor and control of their banks. Briefly, it concluded that politically connected CEOs might decline bank performance. Therefore, we suggest the following hypothesis:

Human capital comprises the arrangement of abilities and information obtained through education and experiences that empower workers to play out their obligations, which deliver their economic values (Jansen et al., 2013). Therefore, the education level of the CEOs is vital for the corporation as it influences performance and decides its gain or loss (Pennings et al., 1998). The personnel of a firm is also called intangible assets that an organization cannot copy or substitute. They give a base to an organization's competitive advantages (Barney, 1991; Carpenter et al., 2001). Therefore, the education level of a director and training are comprised of their knowledge, aptitudes, and capacities—which include an intangible asset for the corporation (Crook et al., 2011). Individuals with higher formal education levels have more critical information and learn organization-specific knowledge (Hitt et al., 2001). Higher education levels also grow individuals' cognitive difficulty and are connected with innovation (Hambrick and Mason, 1984). Furthermore, education increases the self-confidence of an individual and positively impacts the degree of risk tolerance and the capacity to manage uncertainty (Mitchell et al., 2005; Jansen et al., 2013). The education level is considered a crucial part of upgrading an organization's performance (Makhlouf et al., 2017). Past investigations, e.g., (Carpenter and Westphal, 2001; Darmadi, 2013a), show that the education level of a CEO considers a personal cognitive capacity and abilities. For example, when CEOs have an advanced education level, they obtain good strategic ideas and decision-making skills because this advanced education quit access to original ideas. In addition, executive bodies with an advanced level of education have a more intellectual, scholarly influence, creative thoughts, and distinctive perspectives, all of which enable them to manage various problems effectively. This shows that an increase in education level enhances organization management competence and increases competition in the industry. Therefore, we hypothesize that the higher level of education, the higher the performance of the bank.

According to the study of Hambrick and Mason (1984), experience means skills, and it is regardless of education. In his evaluation, the effect of the board of directors' education on corporate performance, he also recommends that experience is a more potent variable in the board of director's capabilities. This idea may increase from the realistic experience compared with the conceptual and workbook methodology of formal education. It is found that the experience of an outside director positively affects the financial and market performance of an organization—previous experience of the CEO positively affects the financial performance of a bank in an unstable industry. Experience positively impacts the performance of a bank in a stable industry (Hambrick and Mason, 1984). Specialized experience of a CEO has been found to affect a CEO's activities positively and efficiently in leading organizations (Hambrick and Mason, 1984; Herrmann and Datta, 2006; Crossland et al., 2014). In summing up, CEOs would be empowered to get more knowledge and ability with more experience, which would interpret more tactical decision-making and enhanced bank performance. Therefore, we hypothesize that the higher the experience of the CEO will be higher the performance of the bank.

Chief Executive Officer global experience helps organizations in making international competitiveness through worldwide diversification. Such experience grooms directors for adapting to unexpected problems and new issues. Therefore, global experience has become a requirement for a CEO (Bass and Bass Bernard, 1985; Black, 1999). With increasing globalization, the organization will change the framework of its management by having more foreigners on its top management (Sanders and Carpenter, 1998; Oxelheim et al., 2013). In such a manner, organizations try to draw foreign directors to add managerial abilities and particular skills (Sanda et al., 2008). Suppose that CEOs with different backgrounds, especially nationality, can encourage organizations to understand comprehensive sensitivities, which help them in entering new international markets (Carter et al., 2003; Hillman and Dalziel, 2003; Pfeffer and Salancik, 2003; Abu et al., 2016; Scheppink, 2018). As a result, both CG and corporations prefer to assign directors with overseas nationality or experience (Oxelheim and Randøy, 2003; Masulis et al., 2012; Rose, 2016; Ware, 2016; Rahman, 2018). Past studies found that overseas executives have a positive correlation with the financial performance of an organization (Rosenstein and Wyatt, 1997; Carter et al., 2003; Oxelheim and Randøy, 2003; Müller, 2014) in the Netherland (Overveld, 2012), Nigeria (Abu et al., 2016), Korea (Choi et al., 2007), and Kenya (Karani, 2015). Conversely, it is found that foreign directors have low participation and a feeble monitoring role because of their home abroad. Furthermore, language differences and newness to or little learning of the neighborhood culture, market, and economy also diminish their effectiveness. Therefore, foreign CEOs have no binding impact on the stock exchange and organization market value in rising economies, such as Turkey, Indonesia, and Pakistan. As a result of the dissimilar and rare literature that concentrates on the developed economies, this study further investigates the relationship between foreign CEO and bank performance in rising countries, such as Pakistan. Additionally, examine the investigation builds up the accompanying hypothesis.

In exploring the effect of CEOs on bank performance, this paper explores the UET, as devised by Hambrick and Mason (1984). This theory describes that managerial background attributes can reflect tactical decisions. The basic principles of UET are derived from leap rationality theory, which expresses that the decision-making of an individual is not entirely based on rational motives because these cannot capture all relevant information about the subject (March, 1978). Consequently, people generally rely upon their emotional and behavioral factors in decision-making (Hambrick and Mason, 1984), which means that the study of executive decision-making must consider the difficulties in measuring the psychological components of CEOs. Researchers work on UET and additionally contended that directors' apparent demographic indicators (for example, education level, business degree, age, gender, ethnicity, political background, and tenure length, etc.) could suggest to their characters; these are objective and measurable (Marcel, 2009; Abdullah and Said, 2018). By drawing on UET, a few studies have revealed that how different attributes of executives, for example, gender (Kassinis et al., 2016), age (Lee et al., 2018), education (Manner, 2010), ethnicity (Louis and Osemeke, 2017), and political belief system (Chen, 2015), can affect the financial performance of a firm. Generally, our UET-based investigation concentrated on noticeable attribute indicators of CEOs, which comprises education, experience, nationality, political association, and MTB, to clarify the level of effect they have on corporate financial performance.

This research uses a quantitative approach because this manuscript aims to investigate the relationship of independent variables, such as military, political connection, education, experience, and nationality of a CEO, with the dependent variable as ROA. The sample of this research includes 20 private commercial banks of Pakistan. Secondary data have been derived from 2011 to 2020. This period covers the post-crisis and the latest data, and the latest data improve bank governance policies. Next, this is a large sample size and a broader era of the time difference to explore the effect of a CEO on bank profitability in different financial conditions. The total number of observations is 200. Data associated with CEO attributes and financial performance (ROA) are collected from the audited published annual reports and balance sheets of the chosen banks, just as from SBP and LinkedIn publications. Political data were collected from the certified webpage Election Commission of Pakistan (ECP), which holds elections for the National and Provincial Assemblies. A few banks were excluded because the data were not extensively accessible.

ROA is a dependent variable that shows bank performance. ROA is known as the income produced by the bank according to its resource base. It is calculated to control the operating performance of a bank (Yim, 2013; Serfling, 2014). This methodology is also implied by Mishra and Nielsen (2000) and Peng et al. (2007). ROA is measured as net profit after tax to total assets by 100.

Chief Executive Officer with an MTB: In this research, an MTB of CEO is represented by MTB, a dummy variable that takes one if a CEO is an MTB in a given year and zero if otherwise. Some past investigation has also used these variables (Benmelech and Frydman, 2015; You et al., 2020). CEO PC: A CEO is considered politically connected when he stands in the national or provincial election, held from 2011 to 2020. These measures are used in a few different investigations (Asquer and Calderoni, 2011; Saeed et al., 2014). CEO education: Education level of a CEO, for example, holding a higher degree, considers a vital asset for banks and gives a combination of capabilities, expertise that help in applying the governance rules (Ujunwa, 2012). Therefore, in this investigation, the CEO education level is calculated by the percentage of CEOs holding an advanced degree, such as a Master's or Ph.D. degree. This indicated variable has shown up in some past studies (Bhuiyan et al., 2010; Ujunwa, 2012). CEO experience: CEO experience is a significant independent variable. We can measure it in several years. Some past studies also used the experience as an independent variable (Stimpert et al., 2010; Phan, 2016). Foreign CEO: This variable is measured by the proportion of foreign CEOs on the bank boards. In various past investigations, nationality was used as an independent variable (Darmadi, 2011; Ibrahim and Hanefah, 2016).

To control the effect of economic conditions, this paper used the following control variables. Bank size (BSZ) is an important control variable because large banks are ideal for increasing their performance by guaranteeing proficient usage of their considerable assets (Cheung et al., 2007). Subsequently, financial investors favor large banks that expand their market worth. Bank size is determined by taking the common logarithm of total assets some past investigations also used (Abdullah, 2005; Murtaza and Azam, 2019). Leverage measure as total obligations divided by total assets (Ilmas et al., 2018; Khan and Wang, 2021). Measurement of each variables are exist in Table 1.

The statistical technique of Eviews was applied in the process of data analysis. To test the study hypothesis, we followed the FE model (Jensen and Murphy, 1990; Hermalin and Weisbach, 2006; Buck et al., 2008; Ausat, 2018; Khan and Wang, 2021). The following is the regression equation of this study:

Where i refers to the cross-section, t is time, and μ_it is an error term. β₀ is the Constant coefficient of regression. B₁-β₅ refer to the regression coefficient of independent variables. β₆ and β₇ are control variables.

We know that simple Ordinary Least Squares (OLS) model cannot be a perfect estimator. Every board data model is not viewed as the equivalent of time series and cross-section data since it labels to have pairs of subscripts-i and t—where i shows the unit of investigation and t demonstrates time measurement (Baltagi, 2008). The error term of the Pool model comprises of three mechanisms—individual (μi), time-specific effect (λt), and other disorder (vit).

Where Uit = μi + λt + vit μi = cross-section property λt = time-specific property vit = other random errors

Based on these kinds of impacts of pool data, two distinct models are presented to manage cross-sectional effects:

The FE model investigates the relationship between predictor and response variables inside an entity (country, individual, and organization). Every organization has its separate attributes that may affect the independent variables. For example, being a male or female could affect the thinking toward a particular problem, or the political structure of a specific state could have a few impacts on business or GDP, or the trade approaches of a corporation may affect its stock price. When using FE, we accept that something within the individual may affect or bias the independent or dependent variables, and we have to control for this. The justification behind the presumption of the relationship between the error term and independent variables of the organization. The FE model eliminates those time-invariant characteristics to assess the net effect of the independent variable on the result variable.

Table 2 shows the descriptive statistic of all variables implied in this research. These results are beneficial to explain the individual variables and their contribution. The secondary data are collected from different banks working in Pakistan, which contain 200 sample observations. In this study, the financial performance of the bank measured by ROA is considered a dependent variable, and all other variables are considered independent variables. These measurements describe the shape, location, and variation of each variable included in this study. Most of the regression models need some basic assumptions, and this information provides essential information about these assumptions. In this table, skewness and kurtosis provide information about the distribution of each variable. The majority of the variables do not precisely follow the normal distribution but using the theory of central limit theorem, as we increase the sample size, these variables approach a normal distribution. It can be perceived that the average value of the response variable is 0.01, its minimum and maximum values are 0.04 and −0.07, respectively.

Table 3 shows a correlation between predictor and response variables, and a robust measurement implied in assessing the relationship between pair of variables. It has values between −1 and +1, where the positive figure shows that both variables proceed in a similar direction while the negative value shows their opposite direction. It can be observed that MTB, leverage and ROA, CEO education and MTB, leverage and CEO experience, political connection, CEO education, CEO nationality, and CEO experience have a negative correlation. In contrast, return on assets, political connection and Bsize, CEO education, experience, nationality are positively correlated. CEO education and ROA have the highest correlation, and their value of correlation coefficient is 0.01. The weakest correlation is between CEO experience and CEO political connection (r = 0.09).

Table 4 shows a complete estimation of formulated hypotheses, their significant values, and all measures used for model specifications. Our findings show that military and non-MTB CEOs differ from one another. Regression analyses show a significant negative relationship between CEO with an MTB and bank financial performance. The beta coefficient and its P-values are β = −0.0121, P = 0.0012, respectively. When the CEO of a bank has an MTB, the financial performance of a bank is low. The beta coefficient and its P-values are significant at the 01% level. Hence, the financial performance of banks led by CEOs with MTB is lower contrasted with those managed by the CEO with a non-MTB. The results also show that a CEO with military experience conditionally has a negative effect on the performance of a bank. The effect of politically connected CEOs on bank performance is shown in Table 4. The coefficient of PC is negatively significant at the 01% level, which shows that banks with politically connected CEOs have lower financial performance than those non-political relations. The results also show that the banks led by politically connected CEOs are underperforming instead of non-political banks. These suggest that PC intensifies agency problems by influencing the management to be busy in selfish activities that secure the interests of PC CEOs, decreasing bank performance. Our results are consistent with the results of the studies by Faccio (2006, 2010), Fan et al. (2008), and Khawaja and Mian (2009), each of whom shows a low performance of PC banks. The results show that the correlation between the CEO level of education and bank performance is significantly positive under the first hypothesis. The beta coefficient and P-values are, respectively, β = 0.0060, P = 0.0044. This result is related to the required results. These results confirm and accept the first hypothesis H3 of the study. A significant positive impact suggests that a CEO with a higher education level (Ph. D. or Master Degree) in a board gives more significant bank profitability levels. This study is consistent with past research studies (Mohamed Yunos, 2011; Pulungan and Sadat, 2014). Besides, when a CEO has a higher education level, they acquire good tactical thinking and decision-making skills. Finally, this study guesses that all Pakistani listed private commercial banks must have a qualified CEO and have an advanced education degree. Furthermore, Table 4 shows that the beta coefficient and P-values of experience and education are, respectively, β = 0.0001, P = 0.00372 and β = 0.0060, P = 0.0044); experience values is less than education level. Even though both descriptive variables are positively and statistically significant, but education is stronger in deciding ROA. Therefore, we can accept the fourth hypothesis, H4. However, it cannot support the recommendation of Phan (2016), and the reason is that experience is a strong variable than education for management capabilities. Formal education is required for the development of information, aptitude, and attitude. Regression analysis in Table 4 shows that foreign CEOs (CEON) have an insignificant relationship (β = 0.0141, P = 0.3405) with bank performance. However, according to the past literature, the foreign experience of a CEO contributes to the management abilities and other specialized talents, which guarantees the efficiency and expertise lead of the organizations, which improves financial performance (Black, 1999; Carpenter et al., 2000; Daily et al., 2000; Sanda, 2011). Nevertheless, our findings cannot support the fifth hypothesis, H5. Therefore, we can reject the fifth hypothesis, H5; it means that the nationality of the CEO has no relationship with the financial performance of a bank. However, our findings are consistent with the findings of Arioglu and Borak (2015) and Vania and Supatmi (2014). Nevertheless, the rejection might be because of contextual dissimilarities as the investigations are led in Turkey and Indonesia. The control variable Bsize is significantly positive at the 0.1% level; it means that the large Bsize of a bank has more assets, board members, and market strength to show higher performance. Next, bank leverage shows a significant negative impact on bank performance, representing that positive leverage exercise negatively affects bank performance.

The negative effect could be attributed to the high borrowing costs of the Pakistani markets. In regression analysis, R² is used for the model's goodness, but statisticians preferred adjusted R², which is 43.73% in multiple regressions. To measure the overall significance of the model, the F-statistic is used. The F-statistic P-value is 0.000, demonstrating that the overall model is highly significant and well-fitted. The adjusted R² value of 0.43 in the regression model shows that 43% of the reported variability in ROA can be described by the differences in variables to be specific CEO attributes. The remaining 57% is not explaining, implying that the remaining 57% of the ROA shift is associated with variables not represented in this model. The results of this paper are not free from a small number of limitations. We derived (secondary) data of private commercial banks for this research, from 2011 to 2020, that restricted our model size to ensure data accessibility. We used 20 private commercial banks in Pakistan. If other banks have chosen, then the consequences should be more attractive. The data relating to just a single country and banking sector of the economy, the relationship between CG and financial performance, can be investigated by taking multiple countries' data and including other significant variables of the CEO's attributes and ownership structures.

When the size of the error term contrasts across values of independent variables, then heteroscedasticity is present in the data.

The normality test results in Figure 1 show that all variables for the study were consistently distributed as exposed by the all probability in the histogram are smaller than the significance values of 0.05.

Table 5 exhibits that the F-statistic value is 35, and Prob (F-stat) is 60%, which is more than 10%, so there is no heteroscedasticity.

Table 6 shows that the probability value (F-statistic) is 69%, which is more than 10%; therefore, we can say no heteroscedasticity is present.

Table 7 shows that the values of Variance inflation factor (VIF) are <10%, so we can say that there is no multi-collinearity among the variables.

Augmented Dickey-Fuller test (test for data Stationery)

The fact that Prob(F-statistic) is <0.05 indicates that these study variables are stationary, as shown in Table 8. These results demonstrate that the data in this analysis are stationary and cannot deviate from data stationery assumptions.

This research has updated regulators and various policymakers with essential bits of knowledge to improve future strategies and performance. This study is more important for Pakistan and for all those countries where there are decisive political and military interventions in the corporate policies; in Pakistan, there is a strong presence of political and military approaches. Next, this study recommends that policymakers further endorse CEO characteristics in Pakistan and other developing countries and limit the political and military interventions in corporate policies, such as India, Bangladesh, other Asian countries, etc.