Nexus between corporate governance and earnings management in family and non-family firms

42 2023, volume 26, issue 2, pp. 42–57, DOI: 10.15240/tul/001/2023-2-003 Business Administration and Management Nexus between corporate governance and earnings management in family and non-family firms Aziza Naz1, Monika Naďová Krošláková2, Iqra Farheen3, Marián Čvirik4, Anna Michálková5 1 Bahauddin Zakariya University, Faculty of Commerce, Law and Business, Institute of Management Sciences, Pakistan, ORCID: 0000-0002-6834-1501, [email protected]; 2 University of Economics in Bratislava, Faculty of Commerce, Department of Tourism, Slovakia, ORCID: 0000-00030518-9538, [email protected] (corresponding author); 3 The Islamia University of Bahawalpur, Faculty of Management Sciences and Commerce, Institute of Business, Management & Administrative Sciences, Pakistan, ORCID: 0000-0002-1907-2157, [email protected]; 4 University of Economics in Bratislava, Faculty of Commerce, Department of Marketing, Slovakia, ORCID: 0000-00034701-1543, [email protected]; 5 University of Economics in Bratislava, Faculty of Commerce, Department of Tourism, Slovakia, ORCID: 0000-00029084-4930, [email protected]. Abstract: This study explores the relationship between internal corporate governance mechanism and accruals and real earnings management (EM), following Pakistan’s 2012 amended Code of Corporate Governance. It also examines the moderating role of family firms on the relationship among corporate governance mechanisms and accruals and real earnings management. For this purpose, we used a sample of 172 firms listed on the Pakistan Stock Exchange (PSX) for 2012–2019. The dependent variables were estimated by utilizing the generalized method of moments (GMM), fixed random effects and moderation analysis were estimated through Stata. Hausman specification test was used to choose among random and fixed effects. Results report that board size showed a significant positive impact on abnormal discretionary expenses and overall real earnings management. Board independence led to a substantial and negative impact with accruals EM. Managers of familyowned firms are also opportunistically altering the reported earnings. Results illustrate to the users of financial statements that the reliability of financial information depends on the effectiveness of corporate governance. Also, highlight the efficiency of monitory mechanisms and suggest that board independence is one of the practical approaches for an emerging market to restrain the EM practices. Findings are helpful for regulators, policymakers, and investors to regulate and make policies to improve the financial reporting quality in Pakistan. As per the authors’ knowledge, this study adds the novelty by focusing on two critical internal monitoring mechanisms (board independence and board size) in mitigating earnings management, following the 2012 amended Code of Corporate Governance in Pakistan. It will add uniqueness by studying the moderating role of family ownership, as the literature is scarce. The findings of this study highlighted the impact of the internal monitoring mechanisms on earning management practices in Pakistan. Keywords: Accruals earnings management, board independence, corporate governance code, dynamic panel data, family ownership, real earnings management, Pakistan. JEL Classification: G32, L60, M20. APA Style Citation: Naz, A., Naďová Krošláková, M., Farheen, I., Čvirik, M., & Michálková, A. (2023). Nexus between corporate governance and earnings management in family and nonfamily firms. E&M Economics and Management, 26(2), 42–57. https://doi.org/10.15240/ tul/001/2023-2-003 E+M_2_2023_kniha.indb 42 24.5.2023 12:57:09 43 2023, volume 26, issue 2, pp. 42–57, DOI: 10.15240/tul/001/2023-2-003 Business Administration and Management Introduction Earnings management has raised global concern due to its detrimental influence on financial reporting and, consequently, on financial markets after financial scandals like Enron, and WorldCom. Earnings are a critical measure of a business’s performance and are thus crucial to financial statement users (Jiang & Kim, 2020; Li et al., 2022, Rankin et al., 2017). Under this perspective, it is frequently in the best interests of business owners and managers to manipulate reported earnings for personal gain (Chan et al., 2020; Habib et al., 2017). Manipulation of reported earnings either within GAAP principles or outside leads to inappropriate information about the firm (Rahman & Mohamed Ali, 2006). The financial reporting transparency may be mitigated by incorporating monitoring mechanisms like Corporate Governance (CG; Abu Afifa et al., 2022; Feng & Huang, 2020; Shleifer & Vishny, 1997). Parfet (2000) demonstrated that managers alter reported earnings by utilizing accounting flexibility of practices or policies. Therefore, these financial scandals have established unethical behaviour and realized a need for accountability and transparency in reported earnings (Alareeni, 2018; Jennings, 2012). Connected to this, the flaws observed in such financial scandals resulted in worldwide enhanced recognition and significance regarding accountability, consistency and transparency (Ahmad-Zaluki & Wan-Hussin, 2010; Shehata, 2015). This phenomenon is also true for Pakistan’s market scandals like Islamic Investment Bank, various Housing Cooperative Societies, Taj Company, Mehran bank, KASB bank, and EXACT (Kiernan, 2005). Most of the firms belong to the family business of Pakistan (Khan, 2018). Country investor protection level also determines the level of reporting quality. Pakistan has an underdeveloped capital market, weak investor protection, low transparency, and a lack of necessary voluntary disclosure standards (Arshad & Javid, 2014; Jadoon et al., 2021). Therefore, underlines a dire need to study the financial reporting quality of Pakistani firms. An extensive literature documents the monitoring role of internal CG mechanisms in improving financial reporting quality. Hence, based on the agency theory, this article investigates the relationship between CG and earnings management by focusing on the monitoring function of two integral internal mechanisms, particularly, board independence and the board of directors (Al-Haddad & Whittington, 2019; Jensen & Meckling, 1976). Also, studies based on US and UK data show good CG can minimize earnings manipulation (Klein, 2002; Peasnell et al., 2005). Following this, we investigate whether United States and United Kingdom findings hold for Pakistani firms, extending research by examining whether board independence also assists in mitigating EM practices. Hence, the objectives of this study are manyfold. Firstly, to explore the relationship between accruals and real earnings management and the independence of corporate boards in Pakistan-listed firms. Secondly, to study the nexus between accruals and real earnings management and nonexecutive directors. Additionally, to investigate the interaction effect of family firms in EM and board independence relationship. Following prior studies, this study uses two types of earnings management, i.e., accruals (AEM) and real (REM). We estimate the AEM from the Kothari et al. (2005) model and REM from the Roychowdhury (2006) model following the Al-Haddad and Whittington (2019), as well as Shah et al. (2020). The system and the difference generalized methods of moments have estimated AEM and REM residuals. In summary, board size showed a significant positive relation with AEM and REM. Secondly, a significant positive association was found by the presence of non-executive directors on the board. Thirdly, one of the efficient monitoring mechanisms to mitigate earnings management was board independence. Finally, when we use overall REM, Pakistani firms were positively associated with one or the other to achieve target earnings. The findings of this study extend the existing literature in EM and CG in different ways. The previous literature has no clear relationship between board independence and EM. Therefore, this study will explore the relation by analysing more comprehensive proxies of board independence and earnings management to advance this subject area’s findings because the past studies examine this relationship either through AEM or REM. To the best of our knowledge, this is the first study to consider both proxies with an additional one, i.e., REM-all, by aggregating the three types of REM in one study. The present study will also contribute to the literature by filling the gap identified by (Alhebri et al., 2021; Massis & Foss, 2018; Kumar & Zattoni, 2018) to consider the moderation E+M_2_2023_kniha.indb 43 24.5.2023 12:57:10 44 2023, volume 26, issue 2, pp. 42–57, DOI: 10.15240/tul/001/2023-2-003 Business Administration and Management factors affecting boards of directors’ functions. Identically, the unique feature of family-owned firms is “heterogeneity,” as family firms govern most of the world’s businesses. The same is true for Pakistani firms. Fa mily business strategic decisions are altogether different from non-family firms. When family and business interface with each other, it brings “familiness” to the businesses, bringing unique resources and focus on financial and nonfinancial performance outcomes (Ferramosca & Allegrini, 2018; Gómez-Mejía et al., 2007; Martin et al., 2016). In a situation of low financial performance, the family firm’s consistency indicates that they care not only about monetary profits but also about non-financial profits, including the satisfaction of the need for belongings, the preservation of family dynasty across generations, and the maintenance of good reputation, among others (El-Chaarani & El-Abiad, 2017, 2019). Family business research has shifted from agency to a socioemotional wealth model (El-Chaarani, 2013; Gomez-Mejia et al., 2014). So, this study will also add literature in the context of a family firm’s decision-making. Furthermore, in the context of an emerging nation, i.e., Pakistan, this is the first study to explore the effect of board independence on restricting the REM and AEM after introducing the 2012 code of CG in Pakistan. The remainder of the paper is section 2, literature review, section 3, research design, section 4 results and discussions, and section 5 conclusion. 1. Theoretical background Previous literature notably classifies earnings management into two. One is accruals EM used to fade actual economic performance by making Generally Accepted Accounting Principles (GAAP; Dechow & Skinner, 2000). According to Healy and Palepu (2001), these include the change in accounting policy, i.e., depreciation methods, deferred taxation, fair value measurement, goodwill impairments, etc. In contrast, real earnings management is the operational manipulations by managers instead of a change in accounting policy. Operational manipulations include the shift in timing or structuring of transactions, resource allocation, and investment decisions to change ongoing year earnings (Dechow & Skinner, 2000; Roychowdhury, 2006). Graham et al. (2005) survey illustrates that managers were tempted to manage REM after the Sarbanes-Oxley Act in 2002. REM provides less chance to identify price discounts or change sales strategy by auditors or regulators as AEM does. At the same time, AEM alone does not fulfil the earnings target gap (Roychowdhury, 2006; Shah et al., 2020). Corporate governance is one of the monitoring mechanisms that help in providing financial reporting quality. Along with the alignment effect, this mechanism lowers the agency cost and possible agency problems among insiders and shareholders (Demsetz & Lehn, 1985). The board of directors is the vital internal mechanism of CG that ensures the smooth process of financial reporting quality. Corporate boards monitor all the business activities, aligning shareholders’ interests with internal management, resulting in less managerial opportunism (Liu & Fong, 2010). Literature also finds corporate boards’ effective monitoring machinery which helps in reducing agency problems between managers and shareholders (Fama & Jensen, 1983; Jensen & Meckling, 1976). This study also applied bibliographic ana lysis of keywords. The analysis relies on the 998 articles. The data was extracted from the Web of Sciences (WOS) due to its ability to the data from a broader perceptive of information this platform offers (Zhao et al., 2019) according to their particular disciplines (Hosseini et al., 2018). Following the Web of Sciences is the platform widely recognized by several authors of existing review scient metric data techniques; we utilize data extraction based on the keywords co-occurrence and co-citation analysis (Chen, 2017). We employed a VOS viewer to map the literature matrix based on the keywords due to interest (Yahaya et al., 2020). Previously, this software has also been utilized to perform its powerful ability to construct a visualization map based on the links and the threshold for the object of bibliometric analysis in many studies (Khatib et al., 2021; Zheng & Kouwenberg, 2019). This study performs science mapping by focusing on the central topic themes comprising of 998 papers based on the following keywords: “earning management,” “accrual earnings management,” “real earnings management,” “Corporate Governance,” and “family firms.” These keywords either detected in title or abstract within the search bar query were selected from the Web of Science database, ranging from E+M_2_2023_kniha.indb 44 24.5.2023 12:57:11 45 2023, volume 26, issue 2, pp. 42–57, DOI: 10.15240/tul/001/2023-2-003 Business Administration and Management 2016–2021 of the 67 articles were dropped due to not meeting the co-occurrence threshold. 1.1 Board size It is the conviction in the literature that board size also impacts the financial reporting quality. Consistent with agency theory, larger boards have more expertise, social resources, diversity, and time to monitor management decisions (Dalton et al., 1998). Hence, there is a negative relation between board size and EM. Previous literature findings are inconclusive regarding EM. For instance, Abu Afifa et al. (2022) and Ghosh et al. (2010) report that larger boards mitigate EM practices. Githaiga et al. (2022) as well as Sarkar et al. (2008) conclude the opposite. Whereas, board size was significantly negative when using the REM proxy (Kang & Kim, 2012). Talbi et al. (2015) report a positive relationship with board size. Abdel et al. (2016) and Jamaludin et al. (2015) report no association between board size and EM. This study focuses on board independence’s monitoring role, as large boards have more independent directors with added monitoring capacity. So, embedded in agency theory, we hypothesized that: H1a: There is a significant negative relation between board size and AEM. H1b: There is a significant negative relation between board size and REM. 1.2 Non-executive directors According to agency theory, non-executive directors (NED) are the outside directors on the board who are required to monitor and control the opportunistic behaviour of executive directors (Jensen & Meckling, 1976). In enhancing board effectiveness, NED acts as a check and balance mechanism. As rooted in their equity slice and stemming straight from the directors’ chair, they have more opportunities to control and go through a multifaceted web of incentives (Mangel & Singh, 1993). According to Fama and Jensen (1983), NED plays a monitoring role in management performance and strategic decisions. Being outsiders, NED is an additional source to the firm as they act as decision experts. Not only because of their expertise but also their professional reputation (Pearce & Zahra, 1992; Weisbach, 1988) argue that as NED are independent of management and CEO influence, they act as a positive influencer over directors’ decisions. Also, assist in mitigating managerial consumption of perquisites (Githaiga et al., 2022). Hence, we hypothesized that: H2a: There is a significantly negative relation between non-executive directors and AEM. H2b: There is a significant negative relation between non-executive directors and REM. 1.3 Board independence In literature, board independence is an effective GC mechanism for monitoring and restraining EM practices. Fama (1980) as well as Fama and Jensen (1983) document that the presence of independent directors (IDs) acts as a control mechanism and helps in improving the effectiveness of the board. IDs have strong incentives to act as competent monitors. Independent directors have motivations to cultivate reputations as specialists in decision-making and monitoring, gaining expertise by working on a managerial position outside the firms (Fama & Jensen, 1983) affirm by Githagia et al. (2022) and Goel and Kapoor (2022) study. To the best of the authors’ knowledge, only one study Nazir (2015) studied board independence with AEM and REM proxies in the Pakistani context. Therefore, this study will fill this gap by examining the relationship between two proxies for board independence with AEM and REM. We postulate the hypotheses as: H3a: There is a significant relation between Board independence and AEM. H3b: There is a significant relation between Board independence and REM. 2. Research methodology 2.1 Sample and data selection The sample is Pakistani firms listed on the Pakistan Stock Exchange (PSX) for 2010 to 2019. Data on board independence, REM, AEM, and family/non-family variables are personally administrated from the firm’s annual reports publicly available on PSX (www.psx. com.pk), indicating the date of acquisition of the data and respective firms’ websites. Data on the share price has been collected from the Stock Indices website (www.investing. com). The Securities and Exchange Commission issues the CG code to regulate the listed firms in Pakistan. The first code of CG was enforced in 2002, amended in 2012, and then in 2017. In April 2012, SECP incorporated three introductory clauses and included nine revisions in the previous code. In addition, listed firms must appoint an independent director, E+M_2_2023_kniha.indb 45 24.5.2023 12:57:12 46 2023, volume 26, issue 2, pp. 42–57, DOI: 10.15240/tul/001/2023-2-003 Business Administration and Management which was optional in the last code. This study chooses 2012 as the base year. As REM models require two lag year data, we also included two years of data before 2012. 2.2 Research approach and demographics In this study, following Bisogno and Donatella (2022), we use the deductive approach to find the answer of our research question and quantitative method to analysis the data. The sample is drawn from a population of 530 listed firms on the Stock Exchange of Pakistan (www. psx.com.pk). Dropping missing and unavailable data, we left with with 32% of the sample. Initially, the study attempted to include all the listed manufacturing firms. In next step, this study excludes financial firms because of their regulatory nature, unique capital structure, and accounting methods (Asghar et al., 2020). Data missing for a year from the study period or discontinued operations have been eliminated, leaving a 32% sample of the total population with 1,376 observations over eight years. 2.3 Variable description and methodology, dependent variables A list of variables with abbreviations is presented in Tab. 1. This study uses the Kothari et al. (2005) model to proxy for AEM and Roychowdhury (2006) model as a REM proxy. Accruals earnings management This study uses abnormal operating accruals as a proxy of AEM. The normal accruals are the actual firms’ condition and sales growth which managers cannot manipulate. In comparison, abnormal (discretionary) results from managerial discretion into the financial reporting process. Hence, we use the Kothari et al. (2005) model of performance-matched discretionary accruals developed initially by Jones (1991). According to Kothari et al. (2005), poor performance firms engaged more in earnings management activities, so they incorporate one performance measure, i.e., return on asset (ROA), in the accruals model and relate the accruals with ROA in poorly performing firms. Kothari et al. (2005) model is: TAi,t /Ai,t–1 = β0 + β1(1 /Ai,t –1) + + β2 (∆REVi,t – RECi,t /Ai,t–1) + + β3(PPEi,t/Ai,t–1 ) + β4 (ROAi,t ) + εti,t (1) Variable Symbol Definition Corporate governance variables Total directors TD Total number of directors on the board Non-executive directors NED Ratio of non-executive directors to the total number of directors Board independence BI Independent directors to the total number of directors Control variables Long term debt ratio LTDR Ratio of long-term debt to total assets ROA ROA Profit after tax/total assets Tobin’s Q TQ (Market capitalization + book value of total liabilities)/book value of total assets Market capitalization – Average stock’s upper and lower price/2, multiplied by the total number of outstanding shares Firm size FS Natural log of total assets Interaction term Family ownership FO Dummy variable (one if managers hold 5% of common stock, zero otherwise) Source: own Tab. 1: Variable description E+M_2_2023_kniha.indb 46 24.5.2023 12:57:13 47 2023, volume 26, issue 2, pp. 42–57, DOI: 10.15240/tul/001/2023-2-003 Business Administration and Management where: TAi,t – total accruals defined as earnings before extra-ordinary items minus cash flow from operations of firm i at year t; Ai,t–1 – total assets of firm i at year t; ∆REVi,t – change in sales revenue of firm i at year t, ∆RECi,t – change in firm’s accounts receivables i at time t; PPEi,t – gross value of plant, property, and equipment of firm i at year t; ROAi,t – profit after tax divided by total assets; εi,t – unspecified random errors. We estimated the residuals of the equation and called them abnormal discretionary accruals. This study uses standardized residuals as a proxy of discretionary accruals. Real earnings management Following prior studies, we use three models for REM (based on managerial manipulations in operational activities), developed by Dechow et al. (1998) and applied by Roychowdhury (2006). These activities include:  Sales-based manipulation – by accelerating the timings of sales, giving price discounts, or offering more lenient credit terms, all at the expense of cash flow from operations;  Discretionary expenses manipulation – reduc tion in discretionary expenditures, such as cutting the research and development (R&D), advertising, selling, and administrating (SG&A) funds; lowering discretionary expenses in a particular year will raise reported earnings respective year;  Production cost manipulations – producing more units to lower the total cost of goods sold, which leads to an increase in the profit margin (over-production to report a drop in COGS). According to Roychowdhury (2006), lenient credit terms and discounts are used to uplift sales in the current period, which results in a drop in sales in the subsequent year, accompanied by reduced profit margins and higher production costs. Hence, we assume an abnormally low operating cash flow (AbCFO) in the current period because of sales manipulation. A negative value signals manipulation in cash flow from operations. Contrary to this, managers manage earnings upward by producing more units to report a high operational margin, resulting in a higher value of abnormal production cost. AbProd with a positive sign interprets as a signal of REM. Finally, to meet earnings targets, managers may reduce R&D, selling, general and administrative expenses as these expenses are expensed in the same period as they incurred. So, we expect a lower abnormal discretionary expenses value. A value with a negative sign is interpreted as evidence of abnormal discretionary earnings management. The routine business operational activities are normal cash flow from the operation, normal discretionary expenses, and normal production costs. In contrast, abnormal cash flow from the operation, abnormal discretionary expenses, and abnormal production costs are the estimated values from models. Residuals, the difference between normal and abnormal, are the proxies used for earnings management in further analysis. All these models are estimated for each firm and each year. REM (abnormal production cost, abnormal operating cash flow, abnormal discretionary expense) models are as below. PRODt ⁄At–1 = α0 + α1 ( 1⁄At–1 ) + + α2 ( St ⁄At–1 ) + α3 ( △St⁄At–1 ) + + α4 ( △St–1⁄At–1 ) + εti,t (2) CFOt ⁄At–1 = α0 + α1 ( 1⁄At–1 ) + + α2 ( St⁄At–1 ) + α3 ( △St⁄At–1 ) + εti,t (3) DisExpt ⁄At–1 = α0 + α1 ( 1⁄At–1 ) + + β1 ( St–1⁄At–1 ) + εti,t (4) △INV⁄At–1 = α0 + α1 ( 1⁄At–1 ) + + α2 ( △St⁄At–1 ) + α3 ( △St–1⁄At–1 ) + εti,t (5) COGSt⁄At–1 = α0 + α1 ( 1⁄At–1 ) + + β1 ( St⁄At–1 ) + εti,t (6) where: Prod – residuals from Equation (1–2); St – current year sales; △St – change in sales from the current – previous year (1st lag of sales variable); △St–1 – change in [previous year sales – previous year (2nd lag of sales variable)]; CFO – cash flow from operations; DisExp – research & development + advertising + selling, general and administrative expense; COGSt – cost of goods sold. E+M_2_2023_kniha.indb 47 24.5.2023 12:57:14 48 2023, volume 26, issue 2, pp. 42–57, DOI: 10.15240/tul/001/2023-2-003 Business Administration and Management This study predicts all four earnings management models using “Arellano-Bover/ Blundell-Bond linear dynamic panel data estimation.” Greene (2000) advocates that estimation models comprising multiple lags, especially in dependent variables, may lead to autocorrelation between lagged endogenous variables and residuals when estimated through simple regression. Hence, the results obtained might be biased and inconsistent. Boujelben and Fedhila (2011) maintained using the Arellano-Bover/ Blundell-Bond linear dynamic panel data estimation, an estimation procedure with system GMM (generalized moments of methods). Therefore, this study uses dynamic panel estimates. The differenced GMM estimator proposed by Arellano and Bond (1991) and System GMM model is given by Blundell and Bond (1998). Generalized moments of methods (GMM) use an instrumental variable to resolve endogeneity, measurement errors, and heterogeneity issues. Following Nazir and Afza (2018), this study uses the lagged difference of the dependent variable as an instrumental variable. Residuals obtained from system and difference GMM used as discretionary, abnormal production, abnormal cash flow from operations, abnormal discretionary earnings management proxies in further analysis. Following Al-Haddad and Whittington (2019) and Shah et al. (2020) abnormal cash flows from operations and abnormal discretionary expenses are multiplied by −1 to interpret the same as abnormal production cost and discretionary accruals. Therefore, a higher value indicates a higher probability of changing operational decisions to increase revenue. All earnings management variables are standardized and winsorized at the 1st and 99th levels. Furthermore, we construct an aggregate real management proxy by combining residuals of all three models. Following Alhadab and Nguyen (2018), we interpret REM_all as the higher the REM_all value, the larger the implied use of REM. REM_all is as follows: REM_all = (−abnormal cash flows from operations + abnormal production cost − abnormal discretionary expenses). 2.4 Moderator Family ownership is taken as a dummy variable, one if it is family owned and zero otherwise. Consistent with Porta et al. (1999), we define family ownership as firms or managers having 5% common stock. In our sample, family ownership varies from a minimum of zero to a maximum of 72%, indicating that most of the firms listed on PSX are under family control. These results are consistent with the findings of previous studies in developing countries (Hiranrithikorn & Joemsittiprasert, 2019; Tomaskova et al., 2021), where a large number of family-owned firms exist. 3. Research results and discussion 3.1 Summary statistics Tab. 2 shows the summary statistics of dependent, independent, and control variables. Average mean of total directors was 8 with a standard deviation of 1.1842 with board independence of 13.6259 means that corporate boards in Pakistan have a lower proportion of independence for external monitoring. The mean value for non-executive and independent directors is 4% and 1%, which Variable Observation Mean Std. dev. Minimum Maximum Corporate governance variables TD 1,376 7.7798 1.1842 6.0000 14.0000 NED 4.5603 1.3357 1.0000 9.0000 BI 0.1683 0.0795 0.0700 0.7100 Earnings management variables AEM 1,376 0.9157 0.9894 0.0019 3.1408 AbCFO 2.8207 0.8906 1.5931 3.3101 AbProd −0.6980 1.3216 −3.6500 0.5061 AbDisc −0.0920 0.0464 −0.1730 −0.0270 Tab. 2: Summary statistics – Part 1 E+M_2_2023_kniha.indb 48 24.5.2023 12:57:15 49 2023, volume 26, issue 2, pp. 42–57, DOI: 10.15240/tul/001/2023-2-003 Business Administration and Management means that there are four non-executive directors on average, and one independent director is present on each firm’s board. AbCFO showed a mean value of 2.8207 while AbProd −0.6985 and AbDisc −0.0924. Return on assets showed a value of 0.0357. ROA and TQ led to a value of 0.0357 and 12.5703, revealing the fundamental position of the profitable firms in accounting measures. In terms of the market-based performance measures, i.e., TQ indicates that investors are thinking positively and placing a higher value on the book assets of firms. The average size of sample firms is 3.5924 and varies between 0.81 to 5.66. Long-term debt is an essential source of debt in capital structure. Pakistani firms use on average 0.1457 long-term debt in their structure, which ranges to 2.26 at maximum. As the Pakistani firms have negative equity, therefore the ratio is greater than 100%. 3.2 Multivariate analysis Tab. 3 shows the Pearson correlation matrix of the independent variable. All the independent variables are free from the problem of multicollinearity. The correlation among variables is below the threshold level of 0.70, a limit drawn by Kervin (1992) after which we have TD NED BI ROA TQ FS LtDR TD 1 NED 0.5941 1 BI 0.1335 −0.3839 1 ROA 0.1732 0.1232 0.0749 1 TQ 0.1501 0.0584 0.1273 0.3393 1 FS 0.2139 0.0268 0.1433 0.1052 0.0670 1 LtDR −0.0972 −0.0207 −0.0265 −0.0568 0.0013 −0.3321 1 Source: own (processing in Stata 14) Note: TD – total directors; NED – non-executive directors; BI – board independence; ROA – return on assets; TQ – Tobin’s Q; FS – firm size; LtDR – long term debt ratio. Tab. 3: Correlation matrix of the independent variable Variable Observation Mean Std. dev. Minimum Maximum Control variables ROA 1,376 0.0357 0.1695 −2.6480 3.1261 TQ 12.570 1.6705 4.8477 16.9737 FS 3.5924 0.7757 0.8152 5.6611 LTDR 0.1457 0.1925 0.0001 2.2625 Interaction term FO 1,376 0.5174 0.4998 0.0000 1.0000 Robustness variable ID 1,376 1.3117 0.6797 1.0000 5.0000 Source: own (processing in Stata 14) Note: TD – total directors; NED – non-executive directors; BI – board independence; AEM – abnormal earnings management; AbCFO – abnormal cash flow from operations; AbProd – abnormal production cost; AbDisc – abnormal discretionary expense; ROA – return on assets; TQ – Tobin’s Q; FS – firm size; LTDR – long term debt ratio; FO – family ownership; ID – independent directors. Tab. 2: Summary statistics – Part 2 E+M_2_2023_kniha.indb 49 24.5.2023 12:57:16 50 2023, volume 26, issue 2, pp. 42–57, DOI: 10.15240/tul/001/2023-2-003 Business Administration and Management a multicollinearity problem. Correlation results are coherent to Asghar et al. (2020) and AlHaddad and Whittington (2019). 3.3 Results (GMM) for estimating AEM and REM Tab. 4 shows the results for estimating dependent variables through GMM. Two specifications tests, i.e., the Hansen (1982)/Sargan (1985) tests and second-order serial correlation, specify the reliability and consistency of GMM estimates. As results showed, an insignificant p-value of 0.319, 0.062, 0.288, 0.081, and 0.131 (respectively) validates the validity of the instrumental variable and that the model is well specified. The second test for the overall validity of GMM estimates is the Arellano-Bover/Blundell-Bond (AR2) test for serial correlation of error terms. An insignificant p-value implies that the original error term is serially uncorrelated, and the moment conditions are correctly specified. Hence the results showed insignificant p-value of AR2 i.e., 0.399, 0.314, 0.085, 0.306 and 0.07, respectively. Panel data regression analysis for dependent and independent variables This research uses panel data regression analysis to study the impact of board independence on earnings management. Tab. 5 presents the diagnostics tests for all models. Another test for detecting multicollinearity is the variance inflation factor (VIF). According to Myers (1990), multicollinearity is pronounced in sample data if its value crosses 10. The VIF value for all four models is 1.43, 1.48, 1.53, respectively, concluding that no variable is a linear transformation of the other. Hence, this study is free from the issue of multicollinearity. These findings are consistent with Al-Haddad and Whittington (2019), and Bataineh et al. (2018). Results are reported in Tab. 5. This study uses a balanced panel of 172 firms with 1,376 observations around eight years. The Hausman specification test was estimated to identify the existence of fixed or random effects. For the accruals model, Hausman test value (Prob > Chi2) is below 0.05; hence the fixed-effect model is preferred. AbCFO, AbProd, and AbDisc values for (Prob > Chi2) are higher than 0.05; we choose the random effect model for both models. Breusch-Pagan/ Cook-Weisberg’s test with the null hypothesis of no heteroskedasticity is estimated. All three models report a value less than 0.05 (Prob > Chi2 = 0.0000); hence, our sample suffers from heteroskedasticity. We estimate the Wooldridge test for the absence of autocorrelation of the error term to fulfil another regression assumption. Results indicate the presence of autocorrelation. This study uses linear panel models with robust standard error suggested by Petersen (2009) to correct for he teroskedasticity and auto-correlation. Board size (represented by the total number of directors) significantly impacted AbDISC and REM_ALL. These results corroborate with findings of Al-Haddad and Whittington (2019), and Daghsni et al. (2016). Board independence is another factor that mitigates the manager’s opportunistic behaviour. When we use AEM as an EM proxy, it has a significant and negative impact. The results are consistent with Alden et al. (2019), Al-Haddad and Whittington (2019), and Zang (2012). Firm size is significant and negative when using the AEM proxy. Results are consistent with Capalbo et al. (2014) and Sáenz González and García-Meca (2014). While with AbProd and Tests Disc AEM AbCFO COGS INV AbDISC GMM 1st step difference 2nd step system 1st step difference 1st step difference 1st step system AR2 0.399 0.314 0.085 0.306 0.070 Srgn 0.319 0.062 0.288 0.081 0.131 Source: own (processing in Stata14) Note: Disc AEM – discretionary accruals; AbCFO – abnormal cash flow from operations; COGS – cost of goods sold; INV – inventory, AbDISC – abnormal discretionary accruals; GMM – generalized moments of the method; AR2 – Arellano-Bover/Blundell-Bond; Srgn – Sargan test. Tab. 4: GMM results for creating a dependent variable (AEM and REM) E+M_2_2023_kniha.indb 50 24.5.2023 12:57:18 57 2023, volume 26, issue 2, pp. 42–57, DOI: 10.15240/tul/001/2023-2-003 Business Administration and Management Business Review, 16(1), 21–34. https://doi. org/10.1177/0972150914553505 Nazir, M. S., & Afza, T. (2018). Does managerial behavior of managing earnings mitigate the relationship between corporate governance and firm value? Evidence from an emerging market. Future Business Journal, 4(1), 139–156. https://doi.org/10.1016/j.fbj.2018.03.001 Parfet, W. (2000). Accounting subjectivity and earnings management: A preparer perspective. American Accounting Association. Accounting Horizons, 14(4), 481–488. https:// doi.org/10.2308/acch.2000.14.4.481 Pearce, J. A., & Zahra, S. A. (1992). Board composition from a strategic contingency perspective. Journal of Management Studies, 29(4), 411–438. https://doi.org/10.1111/j.14676486.1992.tb00672.x Peasnell, K. V., Pope, P. F., & Young, S. (2005). Board monitoring and earnings management: Do outside directors influence abnormal accruals? Journal of Business Finance and Accounting, 32(7–8), 1311–1346. https://doi. org/10.1111/j.0306-686X.2005.00630.x Petersen, M. A. (2009). Estimating standard errors in finance panel data sets: Comparing approaches. The Review of Financial Studies, 22(1), 435–480. https://doi.org/10.1093/rfs/ hhn053 Porta, R., Lopez-de-Silanes, F., & Shleifer, A. (1999). Corporate ownership around the world. The Journal of Finance, 54(2), 471–517. https://doi.org/10.1111/0022-1082.00115 Rahman, R. A., & Mohamed Ali, F. H. (2006). Board, audit committee, culture and earnings management: Malaysian evidence. Managerial Auditing Journal, 21(7), 783–804. https://doi.org/10.1108/02686900610680549 Rankin, M., Ferlauto, K., McGowan, S., & McGowan, P. (2017). Contemporary issues in accounting (2nd ed.). Wiley & Sons. Roychowdhury, S. (2006). Earnings management through real activities manipulation. Journal of Accounting and Economics, 42(3), 335–370. https://doi.org/10.1016/j.jacceco.2006.01.002 Sáenz González, J., & García-Meca, E. (2014). Does corporate governan ce influence earnings management in Latin American markets? Journal of Business Ethics, 121(3), 419–440. https://doi.org/10.1007/s10551-013-1700-8 Sargan, J. D. (1985). Professor J. D. Sargan. Econometric Theory, 1(1), 119–139. https://doi. org/10.1017/S0266466600011038 Sarkar, J., Sarkar, S., & Sen, K. (2008). Board of directors and opportunistic earnings management: Evidence from India. Journal of Accounting, Auditing, and Finance, 23(4), 517–551. https://doi.org/10.1177/0148558X0802300405 Shah, S. F., Rashid, A., & Malik, W. S. (2020). Potential substitution between accrual earnings management and real earnings management among Pakistani listed firms. Global Business Review, 097215092095761. https:// doi.org/10.1177/0972150920957617 Shehata, N. F. (2015). Development of corporate governance codes in the GCC: An overview. Corporate Governance, 15(3), 315–338. https://doi.org/10.1108/CG-11-2013-0124 Shleifer, A., & Vishny, R. W. (1997). A survey of corporate governance. The Journal of Finance, 52(2), 737–783. https://doi. org/10.1111/j.1540-6261.1997.tb04820.x Talbi, D., Omri, M. A., Guesmi, K., & Ftiti, Z. (2015). The role of board characteristics in mitigating management opportunism: The case of real earnings management. Journal of Applied Business Research, 31(2), 661–674. https:// doi.org/10.19030/jabr.v31i2.9147 Tomaskova, A., Smietanski, R., & Halouzka, M. (2021). The interdependency between a family company’s name and a family name. Marketing and Management of Innovations, 2, 115–132. https://doi.org/10.21272/mmi.2021.2-10 Weisbach, M. S. (1988). Outside directors and CEO turnover. Journal of Financial Economics, 20(C), 431–460. https://doi.org/ 10.1016/0304-405X(88)90053-0 Yahaya, I. S., Amat, A., Maryam, S., Khatib, S. F. A., & Sabo, A. U. (2020). Bibliometric analysis trend on business model innovation. Journal of Critical Reviews, 7(09), 2391–2407. Zang, A. Y. (2012). Evidence on the tradeoff between real activities manipulation and accrual-based earnings management. The Accounting Review, 87(2), 675–703. https://doi. org/10.2308/accr-10196 Zhao, X., Zuo, J., Wu, G., & Huang, C. (2019). A bibliometric review of green building research 2000–2016. Architectural Science Review, 62(1), 74–88. https://doi.org/10.1080/ 00038628.2018.1485548 Zheng, C., & Kouwenberg, R. (2019). A bibliometric review of global research on corporate governance and board attributes. Sustainability, 11(12), 3428. https://doi.org/10.3390/ su11123428 E+M_2_2023_kniha.indb 57 24.5.2023 12:57:27