Journal of Applied Economics and Business Studies, Volume. 5, Issue 2 (2021) 69-86 https://doi.org/10.34260/jaebs.525 69 Journal of Applied Economics and Business Studies (JAEBS) Journal homepage: https://pepri.edu.pk/jaebs ISSN (Print): 2523-2614 ISSN (Online) 2663-693X Impact of Bank-Specific, Corporate Governance and Environmental Factors on Bank Efficiency and Profitability in Pakistan Hafiz Muhammad Athar1* and Sumayya Chughtai2 1Ph.D. Scholar, Faculty of Management Sciences, International Islamic University (IIU), Islamabad, Pakistan 2Assistant Professor, Faculty of Management Sciences, International Islamic University (IIU), Islamabad, Pakistan. ABSTRACT The study aims to investigate the impact of bank-specific, board structure, gender diversity, and environmental factors on bank efficiency and profitability in Pakistan by taking a sample of seventeen commercial banks for the period 2013-2018. Data envelopment analysis (DEA) and return on assets (ROA) are used as a proxy to measure bank efficiency and profitability. Panel estimation techniques and Generalized Method of Moments (GMM) are used to conceptualize the research framework and to test the hypotheses. The findings indicate a negative relationship of non-performing loans, advances, level of involvement of women into other committees, and CSR index with ROA; while more presence of women on board reveals a positive and significant impact on ROA that is consistent with critical mass theory. However, CEO duality shed a positive impact on technical efficiency; while bank size signifies an inverse relationship with ROA and technical efficiency. Moreover, deposit influences ROA positively; while board size finds a positive and significant relationship with ROA and technical efficiency. The findings are important for various stakeholders as they can efficiently take their decision-making to better understand the factors influence bank performance. This study recommends future researchers do the same research by inculcating a larger sample size. Keywords Too big to fail paradigm, Critical mass/Tokenism theory, CSR index, Data envelopment analysis (DEA), Return on Assets (ROA) JEL Classification C23, E58, G21, G28, G34, K32 * athar23_abbas@yahoo.com mailto:athar23_abbas@yahoo.com Hafiz Muhammad Athar and Sumayya Chughtai 70 1. Introduction The financial sector plays a significant role in the economic progress, growth, and development of the country (Chortareas, Girardone & Ventouri, 2013). History depicts that most of the economic crisis in the banking sector arises due to excessive borrowings, risky over-lending, heavy existence of non-performing loans, illiquidity of assets, weak regulatory and supervisory frameworks/regulations, inadequate credit analysis, bad governance, corruption, political interferences to write-off loans, fraudulent practices such as accounting manipulations, weak internal control systems, etc (Bank for International Settlement, 2015). Thus, a strong and effective financial system is necessary to handle the issues related to poverty elimination, social injustice, unequal distribution of wealth, community’s development, unemployment, inflation, market openness, bad governance, corruption, money laundering, earning manipulations, political interference, global warming, environmental degradation, etc. Broadly, there are two factors affecting bank performance i.e. internal and external. Internal are those that are under the control of management such as bank- specific, governance, corporate social responsibility disclosure, reporting manipulations, etc. However, external are those that are beyond the control of management such as political, social, regulators (central bank and government), macro- economic, technological, competition, demographics/contextual, etc. All these factors may positively or negatively influence the bank's performance, hence ultimately affects stakeholders’ interests. For example, poor asset quality, measured by the non- performing loans to total loans (NPLs/TL) is an indicator of bank failures. This implies that the higher value of this ratio increases the probability of default risk that undermines performance (Batir, Volkman & Gungor, 2017). Liquidity, as measured by deposits to total assets and loans to total asset ratio, may also positively or negatively influence bank performance. They argue that more lending implies more generation of interest revenue if manage risk in a better way, but on the other side, more lending faces a higher risk of bankruptcy that deteriorates performance (Goddard, Liu, Molyneux, & Wilson, 2013). Similarly, the higher level of deposits is a signal of illiquidity of assets which means bank provides lesser lending that not only hampers economic activities but also affects performance (Batir et al., 2017). Capital Adequacy indicates the financial strength of banks or their ability to absorb captivated shocks or losses or unanticipated losses. Larger banks positively influence bank performance as they enjoy the benefits of economies of scale, more market power, ease of obtaining equity, raise debt at a lower cost. etc (Molyneux & Thornton, 1992). Conversely, larger banks may negatively influence bank performance due to the too big to fail paradigm (Batir et al, 2017). Similarly, the dimensions of corporate governance may also positively or negatively influence bank performance. The theories that better explain the relationship between dimensions of corporate governance and performance include information Journal of Applied Economics and Business Studies, Volume. 5, Issue 2 (2021) 69-86 https://doi.org/10.34260/jaebs.525 71 asymmetry theory, agency theory, stewardship theory, tokenism theory, critical mass theory. Agency theory criticizes the duality concept and argues that duality makes managerial monitoring ineffective (Krause, Semadeni, & Cannella, 2014). However, stewardship theory promotes the concept of duality as it states that duality will facilitate strong and unified leadership and makes managerial monitoring effective as insiders have more information and knowledge about the organization than outsiders. Gender diversity is one of the most burning issues regarding the corporate sector because males and females are traditionally, culturally, and socially different (Conyon & He, 2017; Varnita, Niladri, & Jamini, 2018). Regarding gender diversity, few questions arise such as why it is necessary to enter women into the corporate board? What and how they will bring change into a corporate culture that enhances organization performance? Whether her role in the corporate board as a showpiece (token) or key influencers? Likewise, the environmental factors may also positively or negatively influence bank performance. Environmental factors are those that are linked with corporate social responsibility (CSR) disclosure. Disclosure protocols are monitoring devices that reduce information and agency costs by sharing useful and quality information among stakeholders. The banking industry has both a direct and an indirect influence on the environment. Direct impact belongs to the operations within banks such as the use of natural resources (energy management, water management, waste management), digitalization (paperless banking, using ATMs, installation of solar systems, dual utilization of papers), and reduction in greenhouse gas effect (conducting meeting through video conferencing). Indirect impacts belong to the external environmental system i.e. consists of two factors such as the incorporation of environmental risk before lending/ advances and sustainable/green financing. The indirect impact seems to be more important than the direct one as it shed overall influence on society. As in the case of external environmental impacts, the banks provide loans to those projects that lessen pollution and involve the plantation of trees, safe water, renewable energy, etc. The bank should involve in green banking/financing. Green Financing includes 3Ps i.e. Planet, People, and Profit. It refers to engage in such activity that helps to reduce the external carbon emission and provide loans to borrowers compliant with health, safety, and environmental rules. Green banking initiatives include in-house environment management, introducing green finance, incorporation of environmental risk, creation of climate risk fund, introducing green marketing, employee training, customer awareness, and green events. We have also seen that well-established banks also play their role in philanthropic activities (education, health, disaster management, etc), the well-being of employees/citizens, and community development (Szegedi, Khan, & Lentner, 2020; Oyewumi, Ogunmeru, & Oboh, 2018). Primarily, the study examines the impact of multiple factors on bank performance in the context of Pakistan in a single model. A major contribution of this study is the inclusion of the bank-specific, board structure, gender diversity, and Hafiz Muhammad Athar and Sumayya Chughtai 72 environmental factors in a single model to evaluate the bank efficiency and profitability. This study used DEA (two-stage model) and return on assets (ROA) as a proxy to measure bank efficiency and profitability. 2. Literature Review Literature supports a positive relationship of non-performing loans (NPLs) with bank performance (Iveta, 2015; Syafri, 2012). Contrary to the above, Batir et al. (2017) find a negative impact of NPLs on bank performance. This is because of adverse borrower selection i.e. one of the implications of information asymmetry theory. The scholars argue that this happens due to an increase of defaulters that restrict the bank earning resultantly decreases the bank performance. Bourke (1989), Batir et al. (2017), and Gitau, Anyango, and Rotich (2017) find a positive relation of liquidity with bank performance. Contrary to the above, the scholar finds a negative relationship of liquidity with bank performance (Athanasoglou, Brissimis, & Delis, 2008; Havrylchyk, 2006). Bourke (1989), Syafri (2012), Hasanul et al. (2017); and Aziz and Knutsen (2019) find a positive relationship of capitalization with bank performance. This is consistent with the signaling theory and the expected bankruptcy cost hypothesis. Contrary to the above, the scholars find a negative relationship of capital adequacy with a bank’s performance (Bitar, Pukthuanthong & Walker, 2018). This is consistent with the risk-return hypothesis. The supporters of finding a positive relationship of bank size (BS) with performance argue that larger banks enjoy the benefits of economies of scale (Aziz and Knutsen, 2019; Hasanul et al., 2017; Gitau et al., 2017). Conversely, the supporters of finding negative relation of BS with performance posits that larger banks face too big to fail paradigm which implies that it may lessen profits as a result of diseconomies of scale (Bourke, 1989, Batir et al., 2017). The scholars find a relationship between bank size and performance looks like Kuznets inverted U-curve theory, business cycle and product life cycle includes Eichengreen & Gibson (2001). The supporters of findings a positive relationship of board size (BoDS) with performance postulates that the larger board possesses versatile knowledge, skill to make better decisions, and difficult for CEO to dominate (Riyadh et al., 2019; Kiel & Nicholson, 2003). On the other side, the scholar finds a negative relationship of BoDS with performance argues that larger boards are ineffective as it becomes difficult to coordinate, encourages free-riding and each member has their interest that may conflict with the interest of the firm. So, they are in favor of a smaller board size that makes every member more accountable (Adnan, Htay, Rashid & Meera, 2011). Stewardship theory finds positive relation of CEO duality with performance and argues that duality will facilitate strong and unified leadership and makes managerial monitoring effective as insiders have more information and knowledge about the organization than outsiders (Kiel & Nicholson, 2003). Contrary to the above, agency theory finds negative relation of CEO duality with performance and argues that duality makes managerial monitoring Journal of Applied Economics and Business Studies, Volume. 5, Issue 2 (2021) 69-86 https://doi.org/10.34260/jaebs.525 73 ineffective (Krause, Semadeni & Cannella, 2014). Andersson and Wallgren (2018); Beate and Gro (2010), Riyadh et al. (2019) find a positive relation of women’s participation in the board with bank performance. This is consistent with critical mass theory. Torchia, Calabrò, and Huse (2011) argue that most firms still have only one woman on the board that is still considered as a token and finds a negative relationship with bank performance. Token women may face three types of fear such as visibility, polarization, and assimilation. These issues may be resolve when thirty percent of directors are female on the board of directors. Berger, DeYoung, Genay, and Udell (2000) present two hypotheses namely home field advantage and global advantage regarding ownership structure. The supporters of the home-field advantage hypothesis argue that domestic banks are more familiar with local culture, economic, social norms, government policy and regulations, institutional framework, and political factors (Sufian & Habibullah, 2010). Conversely, the global advantage hypothesis advocates that foreign banks may have a comparative advantage of product differentiation, knowledge transfer, modern technology, better risk exposure, and reduction in the cost of capital (Havrylchyk,2006). Two opposite views are prevailing in the literature regarding CSR index and firm performance i.e. (i) Friedman (1970) argues that the manager’s main responsibility to increase firm profit and stakeholder’s wealth and doing anything else will be the misuse of the authority and brings additional expenses. Hence reduces the profit of the company. Fahad and Busru (2021) investigate a sample of 386 companies for the period of 2007- 2016 in India and find that CSR disclosure negatively influences the return on assets. Oyewumi, Ogunmeru, and Oboh (2018) elucidate a negative relationship between CSR and performance. This is consistent with agency theory. (b) Conversely, Freeman (1984) posits that a firm task not only to meet the expectations of shareholders but also to protect the interests of various stakeholders as well. This is in line with the stakeholder theory. The scholars that find the positive relationship of CSR with firm performance claim that CSR disclosure would lead to improving its image/reputation, retention and loyalty of the customer, service delivery, attracting investors and prospective employees, employee productivity (Mravlja, 2017, Maqbool & Zameer, 2018). However, Riyadh et al. (2019) did not find any relation between CSR and profitability. This study used the CSRD index as a proxy to measure environmental factors. Based on the literature and objective, the following hypothesis was tested in this study; H 0: There is a significant relationship of bank-specific factors (asset quality, liquidity, capitalization, bank size) with ROA and technical efficiency. H 0: There is a significant relationship of governance (board size, CEO duality, gender diversity, ownership structure) with ROA and technical efficiency. H 0: There is a significant relationship of environmental factors (CSR Index) with ROA and technical efficiency Hafiz Muhammad Athar and Sumayya Chughtai 74 3. Data and Methodology All banks that fall under the ambit of Pakistan are the population of this study. However, this study concentrates only on commercial banks as the services provided by banks are reasonably homogenous and comparable across countries. Convenience sampling is used and taken data of commercial banks from 2013 to 2018. 3.1. Variables Description and Measurement The variable choice in the study is based on the literature review to align with the past researches. S. No. Variable Name Symbols Formulas Literature Support 1 Asset quality NPL Non-performing loans/total loans Batir et al., (2017); Syafri (2012) 2 Relative Liability Size DEP deposits/total assets Batir et al. (2017); Gitau et al., (2017) 3 Relative Lending Size LOAN loans(advances)/total assets Batir et al., (2017); Syafri(2012) 4 Capitalization CAR shareholders equity/total assets Hasanul et al., (2017); Bitar et al., (2018); Aziz and Knutsen (2019) 5 Bank size BS LN(Total assets) Gitau et al., (2017); Batir et al., (2017); Aziz and Knutsen (2019) 6 Board of director size BoDS LN (total # of board members) Riyadh et al., 2019; Adnan et al. (2011) 7 CEO Duality CEOD 1 if CEO and board chairperson are different persons and 0 otherwise Krause et al., (2014); Kiel and Nicholson (2003) 8 Presence of women on board WTI 1 if the women are more than 30% of board and 0 otherwise Andersson and Wallgren (2018); Riyadh et al. (2019) 9 Level of involvement of women into other committees LIW 1 if women involved in more than one committee and 0 otherwise Andersson and Wallgren (2018); Riyadh et al. (2019) 10 Ownership structure OS %age of shares owned by foreign shareholders to the total number of shares issued Sufian & Habibullah (2010); Herdjiono and Sari (2017) 12 CSR Index CSR CSR score = sum of CSR items/ Total number of CSR items. “1” disclose CSR items and “0” otherwise Fahad and Busru (2021); Mravlja, 2017, Riyadh et al. (2019) 13. Stock market development SMD Stock market capitalization/GDP Sufian, Kamarudin and Nassir (2016) Journal of Applied Economics and Business Studies, Volume. 5, Issue 2 (2021) 69-86 https://doi.org/10.34260/jaebs.525 75 14. Demand Density DD Total deposits of banking sector/Total Area Dietsch and Lozano-Vivas (2000) 14 Return on Assets ROA Net Income/Total Assets Mravlja (2017); Batir et al. (2017); Aziz and Knutsen (2019) 15 Data Envelopment Analysis TE Inputs: deposits, interest on deposits, fixed assets, share capital Outputs: total loans, interest on loans, net income Sealey & Lindley (1977); Batir et al. (2017); Hasanul et al. (2017); Majeed and Zainab (2016) Note: The table exhibits the description, measurements, and the previous scholars used these variables in their studies. Corporate Social Responsibility Disclosure Index (CSR) Literature exhibits the methods to measure CSRD Index such as (1) use of reputation indices such as Dow Jones Sustainability Index, ESG, Asset4, EIRIS, etc (2) content analysis (3) Surveys. This study uses content analysis for the collection of information related to the CSRD Index checklist from annual/sustainability reports, websites, etc. of individual banks. This study uses the dichotomous and unweighted disclosure index method as used by Maqbool and Zameer (2018), and Riyadh et al. (2019). If banks disclose dimensions of CSR in their annual reports or websites, it will be scored one and otherwise assigned zero. The formula used to calculate CSR Index is as under: CSR Index = Sum of CSR items /Total number of CSR items Summary of CSR Disclosure items based on GRI and on Previous Literature “1” if CSR reported in annual/sustainability report and “0” otherwise CRG CSR- Reporting- CCG CSR-Sustainable/CSR Committee-G CSEG CSR-Stakeholders Engagement CNPG CSR-National Policies-G CGA CSR-Accreditation of an international organization CCEG CSR-S_Ethics_Code of Conduct and Ethics CCAMLG CSR-S_Ethics_AML/KYC Policy- S CCGG CSR-S_Ethics_Grievance Redressal Policy CSDG Common forum for dialogue CSIG Other Information disclosure CEWS CSR_Employees Well Being CCPS CSR_Customer Privacy Policy CPAS CSR-Philanthropic Activities Description CDonS CSR_Donations CNRE CSR-E_Natural Resources(N) CDE CSR-E_Digitialization CEDDE CSR-E_Incorporation of Environmental Risk before lending (Environmental Due Diligence) CGFE CSR-EAC_Green Financing CSR CSR Index Hafiz Muhammad Athar and Sumayya Chughtai 76 3.2 Methodology The study for the evaluation of the banking sector's performance is significant as it can influence the interests/decision-making of various stakeholders. The scholars use various proxies to measure bank performance such as accounting or profitability measures (RoA, RoE, etc.), market measures (EPS, market to book value ratio, Tobin Q, etc.), economic measures (economic value added), and efficiency measures (parametric and non-parametric approaches). This study uses return on assets (RoA) and technical efficiency (TE) as a measure of bank performance. ROA indicates that how the firm utilizes its assets to generate income and is used by Athanasoglou, Delis, & Staikouras (2006); Zheng, Rahman, Begum, & Ashraf (2017); Mravlja (2017); Riyadh, Sukoharsono & Alfaiza (2019). The other approach used in this study to measure bank performance is efficiency. Efficiency implies that how the firm utilizes its assets to get the maximum results/output. This study employs data envelopment analysis (DEA) that deals with many inputs and outputs in a single model as used by Farrell (1957); Charnes, Cooper, and Rhodes (1978); Banker, Charnes, and Cooper (1984). The main challenge in DEA is deciding the appropriate selection of input and output variables. This study uses the intermediation approach for the selection of inputs and outputs (Sealey & Lindley, 1977). Literature reveals that the intermediation approach is better than other approaches as it inculcates interest expenses that contribute more than fifty percent of the total costs. Two motives for the application of this approach i.e. (i) banks act as an intermediator to accept deposits and lend them for investments (ii) it is used to assess the efficiency of the entire bank. This study takes deposits, interest on deposits, fixed assets, and share capital as input and produces output in terms of total loans, interest on loans, and net income. Previous scholars that widely used this approach includes Sherman and Gold (1985); Sufian et al. (2016); Alharthi (2016); Majeed and Zanib (2016); Batir et al. (2017); Hasanul, Rubi, and Eric (2017); and Yonnedi and Panjaitan (2019). 3.3. Model Specification To test the hypothesis of this study, the following model is used. 𝑅𝑂𝐴𝑖,𝑡=𝛼0 + 𝛽1𝑁𝑃𝐿𝑖,𝑡 + 𝛽2𝐷𝑒𝑝𝑜𝑠𝑖𝑡𝑖,𝑡 + 𝛽3𝐿𝑜𝑎𝑛𝑖,𝑡 +𝛽4𝐶𝐴𝑅𝑖,𝑡 + 𝛽5𝑆𝑖𝑧𝑒𝑖,𝑡 + 𝛽6𝐵𝑜𝐷𝑆𝑖,𝑡 + 𝛽7𝐶𝐸𝑂𝐷𝑖,𝑡 + 𝛽8𝑊𝑡𝑖𝑖,𝑡 + 𝛽9𝐿𝐼𝑊𝑖,𝑡 + 𝛽10𝑂𝑆𝑖,𝑡 + 𝛽11𝐶𝑆𝑅𝑖,𝑡 + 𝛽12𝑆𝑀𝐷𝑡+𝛽13𝐷𝐷𝑡 + 𝜇𝑡 ………………………………A 𝑇𝐸𝑖,𝑡=𝛼0 + 𝛽1𝑁𝑃𝐿𝑖,𝑡 + 𝛽2𝐷𝑒𝑝𝑜𝑠𝑖𝑡𝑖,𝑡 + 𝛽3𝐿𝑜𝑎𝑛𝑖,𝑡 +𝛽4𝐶𝐴𝑅𝑖,𝑡 + 𝛽5𝑆𝑖𝑧𝑒𝑖,𝑡 + 𝛽6𝐵𝑜𝐷𝑆𝑖,𝑡 + 𝛽7𝐶𝐸𝑂𝐷𝑖,𝑡 + 𝛽8𝑊𝑡𝑖𝑖,𝑡 + 𝛽9𝐿𝐼𝑊𝑖,𝑡 + 𝛽10𝑂𝑆𝑖,𝑡 + 𝛽11𝐶𝑆𝑅𝑖,𝑡 + 𝛽12𝑆𝑀𝐷𝑡+𝛽13𝐷𝐷𝑡 + 𝜇𝑡 ……………………………….B Where, ROAi,t or TEi,t = Technical efficiency or Return on Asset at the ith bank and t time period Journal of Applied Economics and Business Studies, Volume. 5, Issue 2 (2021) 69-86 https://doi.org/10.34260/jaebs.525 77 NPLi,t= Non-Performing Loans to total loans at ith bank and t time period Loani,t= Loans to total asset ratio at the ith bank and t time period Depositsi,t= Deposits to total asset ratio at ith bank and t time period CARi,t = Capital Adequacy at the ith bank and t time period BSi,t = Bank Size at the ith bank and t time period BoDsi,t = Board of Directors Size at the ith bank and t time period CEODi,t = CEO Duality at ith bank and t time period CEOWi,t = CEO Women at ith bank and t time period Wtii,t = Percentage of women on the board at the ith bank and t time period OSi,t = Ownership Structure at the ith bank and t time period CSRIi,t = Corporate Social Responsibility Disclosure Index at the ith bank and t time period Control Variables DDt = demand density at t time period SMDt = Stock Market Development at the t time period μt = Error Term 3.4. Estimation Techniques This study uses panel estimation techniques (common effect, fixed effect, and random effect) to examine the impact of multiple factors on bank performance. Panel data prefer over cross-sectional or time-series data as it addresses individual heterogeneity. However, the diagnostic test such as Breush-Pagan Lagrange Multiplier, Hausman test, Likelihood ratio executes to select which model is best among panel estimation techniques. This study further employs the system GMM estimation technique developed by Arellano-Bover/Blundell-Bond to handle the issues of normality, multi-collinearity, auto-correlation, heteroscedasticity, etc. The scholars who used the same technique in the literature include Aziz and Knutsen (2019), Varnita, Niladri, and Kanta (2018), Maqbool and Zameer (2018), Zheng et al., (2017). Hafiz Muhammad Athar and Sumayya Chughtai 78 4. Results and Discussion 4.1. Descriptive statistics It tells us about the mean, median, maximum, minimum, and standard deviations of different variables used in the study. Table-1 presents descriptive statistics of all variables included in this study. Table-1. Descriptive Statistics Variables Observations Mean Median Maximum Minimum Std. Dev. ROA 102 0.775 0.870 2.718 -2.660 0.900 TE 102 0.758 0.777 1.000 0.224 0.163 NPL 102 11.598 10.244 39.418 1.522 7.298 DEP 102 74.837 75.128 130.951 46.439 11.193 LOAN 102 38.587 37.703 61.098 15.331 7.974 CAR 102 8.723 7.960 25.327 2.240 4.066 BS 102 15.059 15.298 17.011 12.082 1.244 BoDS 102 2.157 2.079 2.565 1.609 0.198 CEOD 102 0.882 1.000 1.000 0.000 0.324 WTI 102 4.933 0.000 66.667 0.000 13.425 LIW 102 0.275 0.000 1.000 0.000 0.448 OS 102 12.172 2.313 78.878 0.000 20.692 CSR 102 0.427 0.392 0.952 0.056 0.226 SMD 102 28.077 28.080 33.000 24.400 2.975 D 102 100.746 100.023 122.366 80.144 15.664 Note: Table-1 shows the statistic summary (mean, median, minimum, maximum, and standard deviation) of dependent variables and explanatory variables used in this study. ROA of Pakistani banks is 0.775. This implies that banks earn 0.775% of total assets with a maximum value of 2.718 and a minimum value of -2.660, whereas standard deviation depicts the variation from means. Furthermore, the median profit of Pakistani banks is 0.870 which is higher than the return on assets. Similarly, TE indicates the efficiency of Pakistani banks with average efficiency and median values as denoted in the table. The explanatory variables are also explained/interpreted in the same analogy. 4.2. Correlation Analysis Correlation analysis exhibits the strength and direction of relationships among variables. Its value lies between -1 to +1. This table indicates that if values are less than 0.70, it means no multicollinearity issue. The variance inflation factor (VIF) also confirmed our previous opinion as all the variables’ values are less than 10. Table-2 indicates correlation analysis and VIF. Journal of Applied Economics and Business Studies, Volume. 5, Issue 2 (2021) 69-86 https://doi.org/10.34260/jaebs.525 79 Table-2. Correlation Analysis and Variance Inflation Factors (VIFs) Probability VIF ROA TE NPL DEP LOAN CAR BS BODS CEOD WTI LIW OS CSR SMD DD ROA 1.000 TE -0.228 1.000 NPL 1.444 -0.420 0.070 1.000 DEP 1.634 -0.039 -0.023 0.222 1.000 LOAN 1.630 -0.456 0.148 0.256 0.097 1.000 CAR 1.629 0.195 -0.028 -0.151 -0.312 -0.184 1.000 BS 4.097 0.605 -0.325 -0.238 0.192 -0.376 -0.290 1.000 BODS 1.214 0.229 -0.139 -0.088 0.041 0.069 0.014 0.178 1.000 CEOD 2.316 0.283 -0.048 -0.053 -0.003 0.097 -0.202 0.438 0.252 1.000 WTI 3.797 -0.338 0.191 0.027 -0.062 0.063 0.281 -0.529 -0.146 -0.637 1.000 LIW 2.144 -0.084 0.031 -0.225 -0.076 -0.155 0.182 -0.033 -0.131 -0.185 0.551 1.000 OS 1.707 0.430 -0.125 -0.225 0.033 -0.162 0.046 0.486 0.017 0.162 -0.096 0.307 1.000 CSR 2.487 0.417 -0.192 -0.423 -0.106 -0.206 -0.062 0.510 0.266 0.079 -0.182 -0.056 0.140 1.000 SMD 1.081 0.000 0.124 -0.070 -0.148 -0.089 -0.013 0.045 0.067 -0.051 0.063 0.085 0.019 0.165 1.000 DD 2.214 0.030 -0.023 -0.288 -0.375 -0.133 -0.114 0.155 0.145 0.037 0.160 0.200 -0.004 0.516 0.229 1.000 Note: This table indicates the correlation analysis. If the values of correlation are less than 0.75 and VIF values are less than 10, this implies no issue of multi-collinearity among variables. Hafiz Muhammad Athar and Sumayya Chughtai 80 4.3 Regression Analysis After the regression, the next step is to test three hypotheses i.e. BP Lagrange Multiplier H0: Pooled is better than random effect model (REM); (ii) Hausman H0: Random effect model (REM) is better than Fixed effect model (FEM); (iii) Likelihood Ratio H0: Pooled is better than fixed-effect model (FEM). The null hypothesis is accepted or rejected based on the p-value. If the p-value is less than .05, reject H0 and vice-versa. After exercising this practice, the Fixed Effect Model is selected as in the case of ROA and Random Effect Model when the dependent variable is technical efficiency. Table 3 presents panel estimation techniques such as pooled, fixed, and random effects. Table-3. Panel Estimation Techniques Variable ROA TE CEM FEM REM CEM FEM REM NPL -0.027 (0.009)* -0.046 (0.016)* -0.030 (0.009)* -0.001 (0.003) -0.006 (0.006) -0.001 (0.003) DEP -0.003 (0.006) 0.006 (0.006) -0.003 (0.005) 0.001 (0.002) 0.001 (0.002) 0.001 (0.002) LOAN -0.028 (0.009)* -0.025 (0.013)*** -0.028 (0.009)* -0.001 (0.003) -0.005 (0.004) -0.002 (0.003) CAR 0.052 (0.018)* -0.003 (0.024) 0.047 (0.016)* -0.006 (0.005) 0.001 (0.008) -0.005 (0.005) BS 0.287 (0.092)* -0.636 (0.315)** 0.313 (0.094)* -0.060 (0.026)** -0.015 (0.110) -0.059 (0.029)** BoDS 0.427 (0.314) -0.871 (0.558) 0.355 (0.324) -0.099 (0.087) 0.065 (0.195) -0.073 (0.100) CEOD 0.431 (0.265) 0.082 (0.307) 0.191 (0.230) 0.132 (0.074)*** 0.240 (0.107) 0.150 (0.076)*** WTI 0.010 (0.008) 0.053 (0.016)* 0.008 (0.008) 0.002 (0.002) 0.004 (0.006) 0.003 (0.002) LIW -0.527 (0.184)* -1.083 (0.225)* -0.600 (0.168)* -0.029 (0.051) -0.052 (0.079) -0.046 (0.055) OS 0.008 (0.004)** 0.021 (0.010)** 0.009 (0.004)** 0.001 (0.001) 0.006 (0.004) 0.001 (0.001) CSR 0.466 (0.394) -3.159 (0.730)* -0.112 (0.400) 0.019 (0.109) -0.454 (0.256) -0.024 (0.123) SMD -0.009 (0.020) 0.009 (0.016) -0.008 (0.015) 0.009 (0.005) 0.009 (0.006) 0.009 (0.005)*** DD -0.009 (0.005)*** 0.027 (0.008)* -0.005 (0.005) 0.000 (0.001) 0.002 (0.003) 0.000 (0.002) C -2.683 (1.725) 11.418 (4.723)** -2.823 (1.659)*** 1.573 (0.479)* 0.460 (1.654) 1.479 (0.523)* R-squared 0.654 0.836 0.488 0.185 0.388 0.149 Adj. R- squared 0.603 0.771 0.412 0.065 0.142 0.023 F-statistic 12.801 12.699 6.447 1.538 1.574 1.181 Journal of Applied Economics and Business Studies, Volume. 5, Issue 2 (2021) 69-86 https://doi.org/10.34260/jaebs.525 81 Prob (F- statistic) 0.000 0.000 0.000 0.120 0.062 0.306 Durbin- Watson stat 1.706 2.477 1.873 2.222 2.586 2.300 Note: *,**,*** at 1%, 5%, 10% respectively. The figure in parenthesis shows standard error. Return on Assets (ROA) and Technical Efficiency (TE) are used as dependent variables. The models represent the impact of the non-performing loans, deposits, loans, capitalization, bank size, the board size, CEO duality, women participation on the board, level of involvement of women on board in other committees, ownership structure, corporate social responsibility index on bank performance. The control variables used in the analysis include demand density (DD) and stock market development (SMD). R-squared is 84% and 15% in the case of ROA and TE. This implies that 84% and 15% variation in our models arises due to explanatory variables. The value of Durbin-Watson lies between 1.5 to 2.5 in the models implies no sign of auto-correlation among the predictors. Bank-Specific Factors and Performance This study reports a significant negative relationship of non-performing loans, advances with ROA, however, insignificant positive relationships exist with technical efficiency. They argue that more lending may cause borrowers to default. This implies that banks may face the risk of bankruptcy that increases the financing cost and reduce profitability. The banks may select adverse borrowers due to imperfect information or high competition exists in the market. Previous literature that supports the findings of this study includes Aziz and Knutsen (2019); Batir et al. (2017) and Sufian and Habibullah (2010). The study elaborates a significant negative relationship of bank size with ROA and technical efficiency. The scholars posit that banks' performance may decline due to diseconomies of scale, mismanagement, bureaucratic issues, and engagement in more risky investments. The scholars that are in support of this argument include Bourke (1989); Syafri (2012); Sufian and Habibullah (2010) and Batir et al. (2017). Corporate Governance and Performance The findings indicate that CEO duality influences technical efficiency significantly and positively. This is in line with stewardship theory arguing that duality creates managerial monitoring effectively. Previous literature that supports this argument includes Kiel and Nicholson (2003). However, an insignificant and positive relationship exists between CEO duality and ROA. Furthermore, the result exhibits a significant positive relationship of the percentage of women on the board with ROA. This aligns with critical mass theory implies that more than thirty percent of women on board enhance the profitability of banks. The supporters of this argument include Andersson and Wallgren (2018); Beate and Gro (2010) and Riyadh et al. (2019). However, the level of involvement of women in other committees finds a significant negative relationship with ROA. This implies that the more presence of a woman in other committees can adversely affect bank profitability. Furthermore, a significant and positive relationship of ownership structure with ROA. The supporters advocate that Hafiz Muhammad Athar and Sumayya Chughtai 82 foreign banks may have a comparative advantage of product differentiation, knowledge transfer, modern technology, better risk exposure, and reduction in the cost of capital (Havrylchyk, 2006). Previous scholars that support this argument include Jayati and Subrata (2018). Corporate Social Responsibility Disclosure Index and Performance The findings indicate a negative and significant relationship of the CSR disclosure index with ROA. They argue that the manager’s main responsibility to increase firm profit and stakeholder’s wealth and doing anything else will be the misuse of the authority and brings additional expenses (Friedman, 1970). This is in line with agency theory (Fahad & Busru, 2021; Mravlja, 2017; Maqbool & Zameer, 2018). System GMM Technique This study also employs the system GMM estimation technique developed by Arellano-Bover/Blundell-Bond as the above analysis includes the problem of endogeneity, heteroscedasticity, autocorrelation. The condition for the application of GMM is (i) when the time period is shorter than the number of groups (T