“Does executive compensation matter to bank performance? Experimental evidence from Jordan”

The high pays received by executives has gained global attention. This study examines the impact of executive compensation on the performance of Jordanian banks, an area that has not been explored much. The study uses empirical methods for data collection and analysis. Dependent variables include Return on Equity (ROE) and Tobin’s Q performance, while total compensation incentives is the main independent variable. Control variables include bank size, bank age, leverage, and female executives. Through balanced panel data analysis comprising 196 bank-year observations, this quantitative research paper applies Ordinary Least Squares (OLS), fixed-effect, and Generalized Method of Moment (GMM) methods. These methods accurately establish the compensation-performance relationship in the banking sector from 2009 to 2022. The coefficient of determination (R2) for the ROE model: 51.63%, Tobin-Q model: 39.33%. These robust models support the main finding that executive compensation is significantly and positively correlated with operating and market-based performance indicators. Results validate the agency hypothesis, indicating that executives are rewarded for bank performance indicators. Consequently, a one-unit increase in executive compensation leads to a rise of 22.8 cents in ROE and 29.51 cents in Tobin-Q. Additionally, bank size, age, leverage, and female executives positively impact bank performance indicators. A modification of BSIZE, BAGE, LEV, and FEMALE by one-unit results in a proportional adjustment of 26.1 cents, 16.6 cents, 2.07 cents, and 48.6 cents, respectively, in ROE. Additionally, a one-unit alteration in BSIZE, BAGE, LEV, and FEMALE corresponds to variations of 77.6 cents, 56.42 cents, 34.39 cents, and 48.8 cents, in Tobin-Q, all in the same direction.


INTRODUCTION
Due to the separation of ownership and control in modern organizations (Aslam et al., 2019), various internal governance mechanisms have emerged to address the agency issues between agents and principals (Mansour et al., 2022a), including executive compensation systems (Morri et al., 2023).Effective executive compensation systems can aid in attracting competent chief executive officers (CEOs) committed to advancing a firm's goals and protecting shareholders' investments (Olaniyi, 2019).However, a much-debated question is about the effect of CEO compensation on the performance of businesses (Basu et al., 2007).From the principal-agent theory perspective, CEO compensation (Dai et al., 2023), specifically linked with operational performance, could diminish agency costs by aligning the interests of CEOs with owners/shareholders, by incentive contracts (Farooq et al., 2023).The underlying argument is that tying CEO compensation to exceptional firm performance incentivizes them (Alves et al., 2016).
The supporters of agency theory have highlighted the pivotal role of CEO's compensation in achieving desired results for sustained business success (Abdalkrim, 2019;Kayani & Gan, 2022).On the other hand, academics asserted that excessive executive compensation wastes business resources (Wang et al., 2021), erodes investor trust, obstructs corporate growth (Zoghlami, 2021), and renders firms vulnerable, causing dissatisfaction among shareholders and the general public as a result of the notable salary gap between CEOs and regular employees (Khaled, 2020).Additionally, Stanford University surveys from the US demonstrate that CEO compensation was excessive (Harymawan et al., 2020).Resultantly, due to an ongoing debate, prior studies are widely heterogeneous (Dias et al., 2020).Therefore, the escalating issue of CEO compensation has gained global significance in contemporary society (Rasoava, 2019; Sajnóg & Rogozińska-Pawełczyk, 2022), particularly following the 2008 financial crisis (Aslam et al., 2019;Bhuyan et al., 2022).Given the foregoing, this study is significant because it aims to provide an answer to an essential question in the Jordanian context as a developing country: What association exists between the CEOs' compensation and the performance of banks?

LITERATURE REVIEW AND HYPOTHESIS
The success or failure of the business is primarily affected by the efficiency of the methods employed by businesses to reward and encourage their staff, including CEOs (Sajnóg & Rogozińska-Pawełczyk, 2022).CEO compensation is crucial for attracting and keeping talented staff in a fiercely competitive global economy (Deysel & Kruger, 2015;Sheikh et al., 2018).Furthermore, CEO compensation is a key signal of a firm's dedication to effective leadership and strategic decision-making (Morri et al., 2023), impacting investors' and stakeholders' assessments of the firm's overall success (Dai et al., 2023).Although the compensation-performance connection has been well studied (Raithatha & Komera, 2016), mostly in Anglo-Saxon nations, it is still a contested topic in other situations.(Chen & Hassan, 2022;Demirer & Yuan, 2013).Executive compensation is a pillar of good governance practices (Zoghlami, 2021).Consequently, some theories support the role of executive compensation in enhancing firm performance, while others reject it (Ozkan, 2011 ).The existing status of the literature, which has a restricted focus in this field of study, served as the inspiration for this study.Therefore, the aim of this analysis is to empirically explore the effect of CEO compensation on the performanceof the Jordanian banking sector during periods from 2009 to 2022; a topic that has received insufficient attention.As per the above review, the following hypothesis is expected to be proposed: H1: There is a positive and significant correlation between the compensation of CEOs and the performance of the banking sector in Jordan.

Study sample
The whole relevant data for the study, which cover the entire population of listed banks at ASE for the period 2009-2022, were taken from the annual reports of fourteen banks that were deposited in ASE and the Securities Depository Center (SDC).There are 196 firm-year observations in all.By using content analysis, the panel data were manually gathered (Mansour et al., 2020).The data were winsorized at the 1st and 99th percentiles to account for the impact of outliers (Raithatha & Komera, 2016).STATA 14 software will be used to process panel data.
The mentioned Jordanian banks serve as a noteworthy example for a variety of reasons (Al-Dhaimesh, 2019

Definitions and measurement of variables
The following sections outline the elements of the suggested framework, as well as the definitions and methods for measuring the research variables:

Bank performance (dependent variable)
To

Total executive compensation (independent variable)
According to the relevant literature (Khaled, 2020; Wang et al., 2021; Zoghlami, 2021), this study assessed CEO compensation in terms of the entirety of compensation and remuneration received by all CEOs of listed banks, encompassing wages, salaries, remunerations, as well as travel and transportation expenses, both within and outside of Jordan.

Control variables
According to the related literature (

Model specification
The OLS method for panel data is employed in this study, which is one of the most popular statistical tools for testing hypotheses in the social sciences (Alves et al., 2016).Hence, panel data's key advantage is the substantial increase in sample size despite having few firms.Employing panel sets augments the available observations, thereby offering the study significant statistical analysis.This study objective involved a thorough analysis of the correlation between CEO compensation in Jordanian banks and the performance of their respective institutions.Employing the baseline regression model, valuable insights can be obtained that will provide a comprehensive understanding of this relationship: , , where i denotes listed bank (1)(2)(3)(4)(5)(6)(7)(8)(9)(10)(11)(12)(13)(14), t period of study (2009)(2010)(2011)(2012)(2013)(2014)(2015)(2016)(2017)(2018)(2019)(2020)(2021)(2022).All definitions of the variables in Table 1 are provided.

Diagnostic tests
The

Descriptive analysis and data description
Table 2 illustrates the descriptive analysis summary of the variables engaged in the multivariate regression model.Thus, Table 2 primarily offers the descriptive statistics of research variables for the unit of study which are 14 listed banks in ASE during the period 2009-2022, through the total number of observations, mean (average) value, standard deviation, minimum, as well as maximum value of these variables.
As Table 2 shows, the ROE ranged from 0.566% to 15.991%, with a mean value (Standard deviation) of 8.06% (3.395).This highlights the variation in banks' ROE during the study period.Also, the Tobin-Q as dependent variable ranged from 0.436% to 2.441%, with the mean value (Standard deviation) of 0.933% (0.409).These statistics infer that there is little difference in market performance across the listed banks on the ASE.
Likewise, Table 2 delivers summary statistics for CEO-PAY, which is the main explanatory variable.CEO-PAY ranged between 13.997% and 15.701%, with a standard deviation of 0.333, and the mean CEO-PAY value was 14.734%.It seems that all control variables also have a clear positive correlation with the performance measures of banks.This indicates that these variables are important factors to consider when assessing the success of a bank.

Univariate analysis
Based on Table 3, it is evident that the variance inflation factors' (VIF) ratios are significantly lower than the benchmark value of 10.This finding implies that there is no severe multicollinearity concern among the explanatory variables (Khaled, 2020).As a result, the regression analysis outcomes can be interpreted with a higher level of certainty and reliance (Abdalkrim, 2019).

Multivariate regression analysis
An analysis of traditional OLS regression was conducted to evaluate the correlation between performance indicators and additional pertinent factors.Based on the R² values, it can be concluded that the independent variables explain for 51.63% and 39.33% of the variation in ROE and Tobin-Q, respectively.The regression analysis (Table 4) shows that the regression coefficients between the CEO-PAY and the ROE and Tobin-Q that measure banks performance are all positive, and the two are significant at the level of 1%.It has been confirmed that Hypothesis 1, which suggests that there is a positive and significant correlation between the compensation of CEOs and the performance of the banking sector in Jordan, is accurate.This result is in complete alignment with the principles of agency theory and is further corroborated by numerous other studies.Better qualified CEOs are therefore more valued by the market and have superior operating results.In fact, every oneunit increase or decrease in CEO-PAY results in a 22.8 cents and 29.51cents change in the ROE and Tobin-Q in the same direction (other things that remain constant).According to this result, CEO compensation plays a significant role in determining performance for Jordanian banks.
Upon analyzing the control variables, it is evident that all control variables have a noteworthy positive impact on both ROE and Tobin-Q.This is clearly reflected by the coefficients outlined in

Additional analysis
To make sure that the results are solid, the study runs additional analysis in this part.Tables 5 and  6 show further analysis.Table 5 shows that the linear regression outcomes also disclose a strong relationship between CEO pay and bank performance.The ROE model has an R 2 (within) of 71.32%% and an F-Test value of 765.48 at the 1% level or better.The Tobin-Q model has an R 2 (within) of 44.98% and an F-Test value of 387.67 at the 1% level or better.In addition, the fixed-effects regression analysis shows in Table 5 that the regression coefficients between the natural logarithm of total compensation incentives (CEO-PAY) and the two performance indicators (ROE, Tobin-Q) that measure banks' performance are all positive, and the two are significant at the level of 1%.This means that for bank performance, every one-unit change (increase or decrease) in the CEO-PAY has a corresponding change in the ROE and Tobin-Q of 17.1 cents and 13.7 cents, re-spectively, in the same direction.The findings show that the findings about H1 are still relevant.Table 5 shows that when the control variables are examined, it is clear that BSIZE, BAGE, LEV, and FEMALE have a significant positive impact on both ROE and Tobin-Q.

Re-examining the study model using a fixed effects model
Upon analyzing the control variables, it is evident that BSIZE, BAGE, LEV, and FEMALE have a noteworthy positive impact on both ROE and Tobin-Q.This is clearly reflected by the coefficients outlined in Table 5.Based on correlation coefficients for ROE model in Table 5, it is clear that any increase or decrease of one unit in BSIZE, BAGE, LEV and FEMALE while keeping all other factors constant, results in a corresponding change of 52.6 cents, 22.5 cents, 66.8 cents and 3094 cents, respectively, in the ROE indicator, and in the same direction.According to the correlation coefficients also shown in Table 5 for the Tobin-Q model, it is evident that any change in BSIZE, BAGE, LEV, and FEMALE by one unit, while keeping all other variables constant, results in a corresponding change of 70.1 cents, 9.33 cents, 11.77 cents, and 9.741 cents, respectively, in the Tobin-Q, and in the same direction.These results help to validate the previous findings in Table 4.
Table 6 summarized the results of the Hansen and Arellano-Bond tests.These tests indicated the validity of the dynamic GMM models (Ibrahim & Maitala, 2023).The results of the dynamic model in Table 6 indicate that past bank performance measures (ROE (t-1), Tobin-Q (t-1)) have a significant and positive impact on the current bank performance measures (ROE, Tobin-Q).Table 6 also shows that the two-step GMM estimation yields the same results as OLS and fixed-effects regression in Tables 4 and 5.As a result, endogeneity issues were unlikely to skew the results.Thus, in the Jordanian setting, the association between executive compensation and banking performance remains valid, and endogeneity is not a concern.
Tobin-Q model, it is evident that any change in BSIZE, BAGE, LEV, and FEMALE by one unit, while keeping all other variables constant, results in a corresponding change of 37.1 cents, 5.04 cents, 48.08 cents, and 45.9 cents, respectively, in the Tobin-Q, and in the same direction.These results help to validate the previous findings in Table 4 and Table 5.

DISCUSSION
This paper explain the findings based on the accomplished analysis to discuss the results.
According to the primary analysis, which used the OLS regression model, there is a positive and substantial association between total compensation incentives and the operating (ROE) and market performance (Tobin-Q) of Jordanian listed banks, indicating the efficacy of CEO compensation incentives in enhancing banks' performance.
Mutual benefit in executive pays and bank performance supports agency theory (Harymawan et al., 2020), reflecting aligned interests of managers and shareholders, potentially cutting down on agency costs (Olaniyi, 2019).Thus, raising CEO compensation in banks may better inspire executives to raise their efforts for the sake of the business rather than increase bank expenditures (  5 reveals favorable and statistically significant correlations between total CEO compensation and both metrics of the performance of listed Jordanian banks.These findings are compatible with the findings of Zoghlami (2021), who employed a fixed effects model.Table 5 also includes fixed-effect regression findings for all control variables that have a substantial positive influence on the performance of Jordanian banks.In general, the results presented in Table 5 are consistent with the primary findings presented in Table 4.This study likewise used a two-step GMM method and generated largely identical results achieved by fixed effects regression.These findings agree with those of Aslam et al. (2019), who employed the GMM model.

CONCLUSION
The aim of this paper was to study the link between CEO pays and bank performance in Jordan.The key finding suggests that CEO compensation incentives exhibit a robust and positive correlation with the operational and market-based performance of the Jordanian banking sector from 2009 to 2022.As a result, a one-unit increase in CEO compensation increases ROE and Tobin-Q by up to 22.8 cents and 29.51 cents, respectively.According to the study's results, CEO compensation positively affects the executive team's motivation, significantly influences bank strategy, decision-making, value generation, directly impacts bank performance.This confirms agency theory predictions and aligns with most research conducted worldwide.Hence, increasing CEO compensation in banks could incentivize CEOs to prioritize shareholder interests.These distinctive findings remain fitting to various forms of endogeneity.All control variables -BSIZE, BAGE, LEV, and FEMALE -have a significant positive influence on both ROE and Tobin-Q.In different areas, this study adds to the literature on CG.Firstly, the conclusions of the study have significant policy implications as Jordanian banks should increasingly adopt good governance procedures to reduce agency issues and improve performance.Secondly, it enhances the understanding of CEO compensation incentives in Jordan's banking sector from the perspective of CG principles embraced in Jordan, an aspect that has received less attention compared to non-financial sectors elsewhere.Thirdly, unlike other research, this study specifically addresses the endogeneity issues arising from the relationship between CEO compensation and business performance to obtain accurate inferences, thereby contributing to methodological improvement in the developing country.Unquestionably, the current study's findings hold substantial importance for regulators, listed bank management, academic researchers, and corporate owners in the ASE market.To ensure continuous improvement in bank performance, it is recommended, among other things, that banks enhance CEO compensation and prioritize performance as the basis for higher remuneration.In brief, this study possesses notable limitations, such as a limited number of observations due to the small count of listed banks, which should be considered in subsequent investigations, including expanding this empirical work to other MENA countries.Furthermore, employing this paradigm in cross-country contexts, especially in underdeveloped nations, could yield novel insights.
).According to Marei et al. (2022), the banking industry aids in the Kingdom's social and economic growth.For instance, the banking industry's assets, loans, and deposits in 2018 accounted for 173%, 112%, and 117% of GDP, re- spectively.The banking industry is also heavily regulated.Mandatory corporate governance (CG) requirements have been enforced by the Central Bank of Jordan since 2008 (Mansour et al., 2023b), which helps Jordan draw in foreign investment (Mansour et al., 2022b).

Table 1 .
Definitions and measurement of study variables enhance the quality of research results, this study has employed two different indicators to estimate bank performance (Al Farooque et al., 2019; Alodat et al., 2022).While the ROE is utilized as it is considered a popular ratio as an alternative measure of businesses performance which is widely used in the accounting and financing literature (Alodat et al., 2023a; Mansour et al., 2022a; Mansour et al., 2022b), Tobin's Q is employed as an alternative market-based measure of bank performance (Barde et al., 2023; Khaled, 2020).Studies have used operating: ROE, and market-based measures: Tobin's Q as an integra- tive way to get a better picture of businesses performance (Ahamed, 2022; Al-Dhaimesh, 2019; Saleh et al., 2021).

Table 2 and
(Shatnawi et al., 2022)of the normality and tests.As one can see in Table2, the results of normality tests outline the value of skewness and kurtosis, exhibiting that the panel data were normally distributed.In addition, Table3shows the results of multicollinearity tests by variance inflation factors (VIFs) ratios for relevant independent variables(Shatnawi et al., 2022).Consequently, it was revealed that all VIF ratios were below of threshold value which is 10 (Khaled, 2020).Thus, there is no concern for multicollinearity problems in these variables(Morri et al., 2023).

Table 2
(Alodat et al., 2023b)nal control variable is FEMALE, which ranged between 0 and 6, with a standard deviation of 1.577, and the average number of female executives working in banks was almost two.These statistics show that female executives are nowadays taking more on the globe business.Thus, the participation of female executives in bank performance is valuable to advance the grounds of gender diversity in the Jordanian context(Alodat et al., 2023b).
considerable disparity in the banks' age.The LEV ranged between 80.2% and 93.4%, with a standard deviation of 2.864, and the average bank gearing value was 86.9%.These statistics show a high ratio Banks and Bank Systems, Volume 18, Issue 3, 2023 http://dx.doi.org/10.21511/bbs.18(3).2023.14 of financial leverage

Table 2 .
Descriptive statistics and normality tests

Table 3 .
Correlation matrix and multicollinearity test

Table 4 .
Based on correlation coefficients for ROE model in

Table 6 .
Results of GMM model for the CEO compensation-bank performance nexus