Corporate Governance and Obfuscation in Chairmen’s Letters: The Case of MENA Banks

Abstract

The readability (RDB) of annual reports (ARs) plays a crucial role in determining the effectiveness of disclosure of information to interested parties, particularly investors. Given that investors rely on the financial information provided in ARs, the chairman’s letter serves as a key communication tool and is the most extensively read section of the report. Consequently, companies are under pressure to provide understandable ARs that can be easily interpreted by investors. Nevertheless, managers sometimes obscure such disclosures in an attempt to bury negative information and hide their own behavior. Drawing from the “managerial obfuscation hypothesis”, this study investigated how the corporate governance (CG) structures affect the RDB of ARs for a sample of 95 banks across seven countries in the MENA region from 2018 to 2022. The findings revealed that board size, frequency of board meetings, and ownership concentration significantly affected the RDB of ARs. Additionally, board independence and gender diversity had a significant negative effect on ARs’ RDB. Conversely, the study found that the presence of role duality within the board had an insignificant effect on ARs’ RDB. As a result, this study recommends enhancing CG structures to enhance the clarity of banks’ reports and boost investor trust.

1. Introduction

Recent scandals and corporate failures, such as Enron, Parmalat, WorldCom, and others, have significantly affected investor confidence in financial reporting (Mahboub et al., 2016). Consequently, the prominence of CG and the need for enhanced transparency have emerged as crucial factors in effectively managing and controlling firms’ activities, including disclosure practices (Mohammadi & Naghshbandi, 2019; Mahboub et al., 2017). While various types of disclosures, both narrative and non-narrative, serve as means of communication with corporate stakeholders, the utilization of narrative disclosure (ND) has become a prevalent trend in contemporary annual reporting, employed by managers to convey information (Leite, 2021). In light of this, the present study focuses on a specific type of disclosure found in ARs, namely the chairman’s letter (CL).

The CL, which has gained increasing attention in recent years, holds a prominent position within the AR (Mahboub et al., 2016). It provides a concise overview of the events that took place during the previous financial year and concludes by offering a brief analysis of the future prospects of the company (Wills, 2009; Yuthas et al., 2002). Importantly, the inclusion of content in the CL is voluntary and not regulated, with the auditor’s role limited to ensuring that the information contained in CL is consistent with the information disclosed in the financial statements (Clatworthy & Jones, 2003). There are no specific requirements regarding the content that should be included in this section of the AR (Leite, 2021). Consequently, it offers management an excellent opportunity to shape external perceptions of the company without being overly concerned about regulatory consequences (Hooghiemstra, 2010). Notably, the CL is one of the most widely read sections of the AR (Mahboub et al., 2017; Cen & Ca, 2015; Clatworthy & Jones, 2006). It is widely regarded as one of the most impactful sources of information that captivate the attention of financial analysts and institutional investors alike (Leite, 2021).

Regarding the utilization of the CL, the available evidence suggests that financial statement users make use of the information enclosed within these letters and that the letters themselves provide valuable information. Previous research has indicated that the CL is widely employed in the decision-making processes of investors (Hooghiemstra, 2010). For example, Kaplan et al. (1990) demonstrated the significant impact of CL on subjects’ perceptions of a company’s future prospects, as well as their likelihood of engaging in actions such as tendering a proxy to management, purchasing additional stock, and retaining existing stock. Tennyson et al. (1990) discovered that the qualitative information found in the CL, along with the “management discussion and analysis” (MD&A), proved to be useful in predicting bankruptcy, surpassing the information provided solely by financial data. Fulkerson (1996) highlighted the high level of interest exhibited by portfolio managers and individual investors in the CL as a source of investment information. Abrahamson and Amir (1996) found that market participants rely on CL information to evaluate the quality of a company’s earnings. Bartlett and Chandler (1997) revealed that investors incorporate the CL into their decision-making processes. Baird and Zelin (2000) discovered that the sequence of positive and negative information within the CL has a substantial impact on the decision-making process of investors. Breton and Taffler (2001) revealed that analysts depend on non-financial, qualitative, soft, and compelling information found in the CL when formulating stock recommendations. Beynon-Davies et al. (2004) demonstrated that the CL contain information that can predict both a company’s future financial performance and the likelihood of corporate failure. Rimmel and Jonäll (2010) acknowledged that managers utilize the CL as a means of establishing legitimacy and influencing readers’ perceptions of the company’s excellence and future viability. Amernic et al. (2010) illustrated the existence of impression management in the CL of a mortgage firm that declared bankruptcy in 2007. Patelli and Pedrini (2014) found that companies with strong historical and projected financial results typically exhibit higher levels of optimism in their corporate disclosures, while also maintaining transparency during challenging macroeconomic environments.

Over the past few years, the significance of ND in the CL has increased (Liao et al., 2023; Jones & Smith, 2014). Previous research suggests that ND should serve as an essential means of communication to ensure that users of ARs can fully grasp the intended message provided by the preparers of these ARs (Smith & Taffler, 1992). However, despite this notion, there has been considerable debate regarding the efficacy of disclosing pertinent information to interested parties in the CL. This is primarily owing to the growing difficulty of accounting regulations and the technical language used in financial information (Guay et al., 2016). Consequently, these factors may contribute to the presence of intricate and incomprehensible language within this section of the AR. As a result, the inability to effectively convey the desired information has become an interesting issue (Ezat, 2019). Moreover, the RDB of these reports, which refers to the ease with which a text can be quickly comprehend, understood, and memorized, has become a challenging issue. This is predominantly significant considering the current surge in information volume and superficial network reading (Luo et al., 2018; Courtis, 1995).

Therefore, ensuring the RDB of the CL is crucial for maintaining good CG. Numerous studies have been conducted to address this issue. Several studies have shown a significant association between the low RDB of ARs and the low performance of companies (Subramanian et al., 1993; Smith & Taffler, 1992). It has been revealed that management consciously or subconsciously attempts to obscure bad news by using obfuscation techniques or by making the text in ARs difficult to read. Eugene Baker & Kare (1992) discovered a significant association between the Flesch score (a measure of RDB) and return on equity (ROE). Smith and Taffler (1992) concluded that low RDB is linked to poor performance while high RDB is associated with financial success. Subramanian et al. (1993) found that the ARs of well-performing companies are significantly easier to read in comparison to those of underperforming companies. Given that many researchers have observed that poor performance tends to make ARs more challenging for users to read. This has implications for CG.

As a result, the RDB of ND documents has garnered attention from both scholars and regulators (Luo et al., 2018). However, most RDB studies have been carried out in Western contexts, with few examining the RDB of disclosure reports in developing countries (Ezat, 2019; Luo et al., 2018). Moreover, the majority of these previous studies have concentrated on the economic consequences of AR’s RDB. This raises the question of whether and how CG influences ND RDB, but very limited studies, if any, have investigated this issue (Luo et al., 2018). Therefore, the main objective of this study is to examine whether CG structures, such as “board size, role duality, board independence, board meetings, board gender diversity, and ownership concentration”, have an impact on the RDB of the CL in banks operating in the MENA region. The banking sector in the MENA region holds significant importance as it plays a dominant role in the financial system of the region, contributing to its economic growth (Mahboub & Fawaz, 2022). Moreover, the narrative sections of major banks in the MENA region have been observed to be complex and challenging to the decipher (Mahboub et al., 2016). Hence, since there are no existing studies conducted that have explored the impact of CG on the RDB of ARs in MENA banks, this study aims to promote horizontal and longitudinal comparisons with other international studies.

The study makes a noteworthy contribution to the current body of knowledge in two main ways. While previous studies on RDB have primarily focused on documents in English from Western countries, this study will focus on seven Arab countries, including “Egypt, Jordan, Kuwait, Lebanon, Qatar, Saudi Arabia, and the United Arab Emirates”. This study stands out as one of the few that takes several countries in the MENA region to examining CG structures and RDB. By doing so, it complements earlier comparative studies that have looked at measurement and disclosure practices. Another key contribution of this study is its exploration of the impact of CG structures on RDB levels, a topic that has received less attention in previous RDB studies. As the majority of RDB studies examine the relationship between company characteristic variables and the RDB level of ARs. Therefore, this study seeks to address this gap by looking at this relationship within the context of the MENA region.

This study is organized as follows: the second section outlines the theoretical framework and provides a concise literature review on the relationship between CG structures and RDB, along with the hypotheses to be tested. The third section details the research methodology employed in this study. The fourth section delves into the interpretation and analysis of the findings. The study’s conclusions are presented in the fifth and final section.

2. Theoretical Framework and Hypotheses Development

Financial statements are crucial sources of information for decision-making in the capital market (Etuk & Akpan, 2023). It is essential that this financial information is easily comprehensible (Hasan et al., 2017). However, some studies have shown that not all investors are able to quickly grasp ARs owing to the technical terminology and intricacy of accounting rules (Guay et al., 2016; Kumar, 2014). Previous research has suggested that managers may intentionally make ARs difficult to understand, especially when the company is facing financial difficulties, in order to convey a biased positive image (Gosselin et al., 2021). Effective CG has been shown to limit such opportunistic behavior by managers (DeBoskey et al., 2019). Therefore, studying the impact of CG structures on the RDB of ARs is a crucial topic to examine, as it affects the effectiveness of communicating information to users of ARs.

Consequently, the complexity of the ND of the AR has garnered significant attention from scholars, policymakers, and practitioners. This is primarily due to its crucial role in conveying the financial and non-financial performance of a firm to its shareholders (Bushee et al., 2018). In light of this, the literature has put forth various theories to elucidate the motivations behind managers’ decisions to make ARs more or less readable. One such theory is “signaling theory”, which posits that RDB serves as a signal employed by well-performing companies with substantial profits. Conversely, underperforming companies tend to obfuscate unfavorable information and make their ARs more challenging to comprehend and interpret (Aldoseri & Melegy, 2023). Specifically, companies may utilize the RDB of their ND to signal a particular situation by concealing undesirable occurrences that could potentially affect their competitive position (Ezat, 2019).

Another theory, “agency theory”, suggests that increased “agency problems” prompt managers to enhance the disclosure of non-financial information, making it more understandable and readable (Aldoseri & Melegy, 2023). However, the textual structure of companies’ reports, according to agency theory, may exhibit complex language necessary to disclose intricate transactions and events. This can create information asymmetry issues as readers struggle to comprehend the content (Ezat, 2019). In contrast, “legitimacy theory” proposes that the RDB of the AR serves to legitimize a company’s activities and cultivate a perception of social responsibility within the community (Aldoseri & Melegy, 2023). Additionally, legitimacy theory suggests that managers utilize the content of the AR to secure stakeholders’ approval. By producing readable ARs, firms can uphold or restore their legitimacy, thereby mitigating information asymmetry and narrowing the legitimacy gap (N’Guessan Pierre, 2022).

The current body of the literature indicates an increasing fascination with the association between CG and the RDB of ARs. Most previous studies exploring the link between CG and RDB have found that improved CG practices tend to result in enhanced RDB (Prusty & Kumar, 2018). This study provides a more detailed analysis of this relationship by considering various CG structures such as “board size, role duality, board independence, board meetings, board gender diversity, and ownership concentration”. These structures are deemed crucial in comprehending the association between CG and RDB.

2.1. Board Size

It is widely acknowledged that the size of the board (BS) is a critical internal component of CG and plays a substantial role in the management of the firm (Isik & Ince, 2016). Boards are responsible for making final decisions on what information should be communicated to external stakeholders through ARs (Dalwai et al., 2023). However, there are different viewpoints in the literature regarding the optimal size of the board. According to “resource dependence theory”, a larger board is expected to enhance monitoring, reduce managerial opportunism and promote greater disclosure in AR (Kolsi, 2017). However, “agency theory” suggests that larger boards may lead to less efficient control over company activities, hindering improvements in the quality of voluntary information disclosure (Navarro & Urquiza, 2015). As a result, studies investigating the association between BS and the RDB of ARs have yielded mixed results. For example, Ginesti et al. (2017) did not find evidence supporting the impact of BS on the RDB of MD&A in Italy. Similarly, Ezat (2019) did not find evidence approving the impact of BS on the RDB of board reports in the Egyptian context. Additionally, Dalwai et al. (2023) discovered no significant association between BS and RDB in the financial sector, including finance, insurance, and investment subsectors. However, BS was found to have a negative association with Flesch RDB in the banking sector in Oman. Ginesti et al. (2018) further show that BS is significantly and negatively correlated with the Fog index, indicating that increased participation of board members enhances information transparency. This is because a greater number of directors can bring more experience, expertise, and skills to the board, potentially strengthening control mechanisms to safeguard shareholders’ interests and promoting transparent AR disclosure to stakeholders (Chithambo & Tauringana, 2017). Nonetheless, Etuk and Akpan (2023) provided evidence that a significant increase in the BS can enhance the RDB of ARs for oil and gas companies listed in Nigeria. Based on these inconclusive results, it is hypothesized that the influence of the BS on the RDB of ARs can be summarized as follows:

H1. 

There is a significant impact of board size on the readability of annual report.

2.2. Board Meetings

Board meetings (BMs) serve as a platform for discussing various aspects and concerns related to the organization with the board of director (BOD), leading to informed decision-making (Kakanda et al., 2017). While some argue that a high frequency of meetings may indicate inefficiency among directors (Vafeas, 1999), others suggest that a greater number of meetings can enhance oversight of business operations and increase the likelihood of voluntary disclosure (Lipton & Lorsch, 1992). Despite this, research on the relationship between BMs and CG has yielded mixed results. For example, Prusty & Kumar (2018) found a significant association between the RDB of business responsibility reports and BM in India. Conversely, Ezat (2019) reported a negative correlation between the frequency of BM and RDB scores, indicating that companies listed on EGX100 with more frequent BM tend to have more readable ARs. This negative relationship can be explained by “agency theory”, as increased BMs can enhance the board’s monitoring of management and facilitate clearer disclosure of information. However, Astari et al. (2020) found that the frequency of BMs does not significantly impact the RDB of ARs, as these meetings cover various aspects beyond just the ARs. Additionally, Dalwai et al. (2021) discovered a weakly positive correlation between Flesch−Kincaid RDB and BMs in Italy. Therefore, the influence of BMs on the RDB of ARs remains a subject of hypothesis and further investigation. Based on these mixed results, it is postulated that the influence of BMs on the RDB of ARs can be postulated as follows:

H2. 

There is a significant impact of board meetings on the readability of annual reports.

2.3. Board Independence

Having independent directors who are not affiliated with the company is of utmost importance (Dalwai et al., 2023). Advocates of “agency theory” argue that increasing the number of “non-executive directors” (NEs) on the board is advantageous in protecting the interests of stakeholders (Forker, 1992). Previous research in CG suggests that an independent board effectively monitors and ensures more disclosures and greater transparency through increased monitoring (improvement of disclosure hypothesis) (Goh et al., 2016). However, some theoretical studies raise concerns about the effectiveness of monitoring by an independent board in improving disclosure quality (Adams & Ferreira, 2007). These studies propose that independent boards may not necessarily encourage greater disclosure quality, as managers may obscure information to avoid scrutiny from the board (avoidance of monitoring hypothesis) (Rahman & Kabir, 2023). The relationship between board independence (BI) and RDB of ARs remains unclear based on the prior literature. For example, Ezat (2019) did not find empirical evidence for the impact of NE members on the RDB level of board reports in the Egyptian context. However, Gosselin et al. (2021) demonstrated a positive and significant association between BI and the FOG index. Rahman and Kabir (2023) found that BI decreases the RDB of ARs, aligning with the notion of managers avoiding costly board monitoring. This result was confirmed by Dalwai et al. (2023), who showed a statistically significant negative correlation between BI and the Flesch RDB index for the financial sector and bank subsector. This finding contradicts the previous study by Harjoto et al. (2021), which reported that BI enhances the RDB of corporate social responsibility (CSR) reports. Based on these contradictory results, the impact of BI on the RDB of ARs is hypothesized as follows:

H3. 

There is a significant impact of board independence on the readability of annual reports.

2.4. Board Gender Diversity

Gender diversity (GD) is an integral part of CG and cannot be separated from it. It is a well-established fact that every company has both male and female leaders in various positions (Shauki & Oktavini, 2022). Scholars have employed different theories to shed light on how the presence of female directors influences the quality of financial reporting (Kılıç & Kuzey, 2016). From the perspective of “agency theory”, female directors play a beneficial role in the CG system. They are known for their diligence, active participation in board meetings, and a tendency to be less assertive and more risk-averse (Ho et al., 2015; Adams & Ferreira, 2009; Huse & Grethe Solberg, 2006). On the other hand, “resource dependency theory” suggests that having female representation on the board can bring valuable advice, enhance the legitimacy of organizational practices and open up new channels for communication and organizational improvement (Hillman et al., 2007). However, “social psychological theory” presents a contrasting viewpoint. According to this theory, directors who hold a strong position are more likely to exert influence over the decision-making process of the entire board, thereby potentially limiting the actions of a more gender-diverse board (Mateos de Cabo et al., 2012). Despite the growing interest in the relationship between board GD and financial reporting quality, there is limited knowledge regarding the impact of GD on the RDB of corporate reporting. For example, Ginesti et al. (2018) discovered that female board participation has a positive effect on disclosure RDB in companies with small boardroom connections, but the opposite effect in companies with large boardroom connections. Nadeem (2022) validated this result and identified a significant positive influence of board GD on the RDB of 10-K reports. This suggests that having female directors is linked to enhance RDB. The findings align with theoretical expectations that female directors advance the interests of various stakeholders by enhancing firms’ communication with the capital market. Given that the clarity of information in 10-K reports is crucial for stakeholder decision-making, especially for individual investors, female directors, who prioritize the concerns of others and exhibit higher ethical standards, are connected with improved RDB of 10-K reports. Moreover, Ben-Amar et al. (2024) discovered that GD in the board improves the RDB of ND and is correlated with a less optimistic, litigious, and ambiguous tone in ARs. However, Shauki and Oktavini (2022) demonstrated that the presence of female directors in the organization did not affect the RDB of ARs in Indonesia. Based on these opposing results, the impact of board GD on the RDB of ARs is hypothesized as follows:

H4. 

There is a significant impact of board gender diversity on the readability of annual reports.

2.5. Role Duality

Another characteristic of strong CG is the separation of the roles of the chief executive officer (CEO) and board chairperson (Boateng et al., 2022). CEO duality occurs when one individual simultaneously holds both positions within the same company (Barnawi & Abdullah, 2023). According to “agency theory”, organizations with CEO duality tend to concentrate power in a single individual, which can limit the board’s ability to effectively monitor and govern the company functions (Fama & Jensen, 1983). Consequently, CEO duality has been examined as a factor influencing the RDB of ARs. However, existing empirical studies on CEO duality have produced inconclusive results. For instance, Ginesti et al. (2017) found that firms where the chairman and CEO roles are combined in one person tend to have lower levels of ARs’ RDB. This finding aligns with previous research suggesting that CEO duality indicates weaker board oversight, leading to reduced transparency in corporate disclosures (Gul & Leung, 2004). Furthermore, Roiston and Harymawan (2022) discovered a significant positive relationship between CEOs holding multiple positions in subsidiaries and the RDB of the company’s financial statements. In contrast, Ezat (2019) indicated that the division of responsibilities between the chairman and CEO has a significant correlation with the RDB of board reports in EGX100 companies, suggesting that companies with a separate chairman and CEO tend to produce more comprehensible reports. This outcome aligns with “agency theory”, which advocates for the segregation of these roles to uphold the independence of the board. Sun et al. (2023) corroborated this finding and demonstrated that dual roles enhance the RDB of ARs. Additionally, Rahman and Kabir (2023) contended that organizations where the CEO also serves as the board chair experience an enhancement in ARs’ RDB. This observation aligns with the efficiency theory of CEO duality. Based on these inconsistent results, the impact of role duality on ARs’ RDB is hypothesized as follows:

H5. 

There is a significant impact of role duality on the readability of annual reports.

2.6. Ownership Concentration

Ownership concentration (OC) pertains to the percentage of shares owned by the principal shareholders of a corporation (Ahmed et al., 2012). It plays a critical role in determining the distribution of power within the firm (Thomsen & Pedersen, 2000) and influences the composition of the BOD (Dalwai et al., 2023). Consequently, the ownership structure (OS), as a part of the CG mechanism, has garnered significant attention in relation to its impact on voluntary disclosure. Drawing on “agency theory”, Morck (2000) posits that when a company has broadly dispersed ownership, a “free rider problem” arises due to the lack of means and motives for dispersed owners to address managerial agency problems. This, in turn, results in an increase in agency costs (Fama & Jensen, 1983). To mitigate these costs, firms tend to disclose further information (Garcia-Meca & Sanchez-Ballesta, 2010). Conversely, a high degree of OC may give rise to agency conflicts between majority and minority shareholders (Fama & Jensen, 1983). Consequently, firms with controlling shareholders may perceive a reduced need for superior—quality public disclosures and financial reporting to effectively monitor management. This scenario gives rise to an increased possibility of minority shareholders being expropriated and important information being withheld thereafter. Considering the impact of OS on the RDB of ARs, empirical findings are varied and contradictory. For example, Kumar (2014) demonstrated that companies with a more diverse OS produce ARs that are easier to read for Asian companies listed in the United States. This finding is supported by Dalwai et al. (2023), who found that dispersed ownership leads to more readable reports specifically for banks. However, Khan et al. (2017) presented evidence showing that the concentration of ownership is positively correlated with RDB, aligning with the arguments of “agency theory”. Velte (2018) also discovered supporting evidence for the notion that OC can contribute to improved RDB of integrated reports. Similarly, Melón-Izco et al. (2021) obtained the same result and confirmed that there is a positive and statistically significant association between OS and RDB. On the other hand, Etuk and Akpan (2023) provided evidence suggesting that OC has an insignificant effect on the RDB of ARs for listed oil and gas firms in Nigeria. Their results indicated that OC does not significantly decrease the page length of ARs for these firms. Based on these varying findings, it is hypothesized that OC has an impact on the RDB of ARs in the following manner:

H6. 

There is a significant impact of ownership concentration on the readability of annual reports.

3. Research Methodology

3.1. Research Sample and Data Collection

The research sample comprised 95 banks from seven distinct countries in the MENA region, namely “Egypt, Jordan, Kuwait, Lebanon, Qatar, Saudi Arabia, and the United Arab Emirates”, over a five-year period from 2018 to 2022 (

Table 1. Sample size.

	Country	Number of Banks	Number of ARs
1	Egypt	18	90
2	Jordan	11	55
3	Kuwait	10	50
4	Lebanon	15	75
5	Qatar	11	55
6	Saudi Arabia	14	70
7	United Arab Emirates	16	80
Total		95	475

Source: developed by the researcher.

). This time period was chosen on the basis of data availability. More specifically, the reason for selecting this time frame is that the ARs used in previous studies were limited to those ending in 2018. To the knowledge of the researcher, it represents the first attempt to examine RDB in the narrative section of the ARs in the MENA region around the year 2018. The selection of sample banks was based on the criterion that the bank had published ARs in the English language on its websites consistently for five years and the banks selected are country-specific. The data related to CG variables were obtained from the CG reports and websites of each bank. Additionally, the data concerning RDB were extracted from the CL in the AR of each bank. Furthermore, financial data were gathered from the financial statements section of the AR of each bank.

3.2. Measurement of Variables

Six variables related to CG were utilized to investigate the potential influence of the CG mechanism on the level of RDB. Furthermore, this research incorporated three control variables that are widely acknowledged in the literature and are commonly examined due to their significant association with RDB.

3.2.1. Measurement of the Dependent Variable

In general, as a common rule, ARs should be readable and understandable to be considered a key means of communication between organizations and their external users (Alzarouni et al., 2011). Therefore, all of the authors seek to measure the RDB of their materials (Dawson, 2008). However, several authors face the problem of how to judge the RDB of their text (Jabbari & Saghari, 2011). Over time, various techniques and methods have been developed to accurately anticipate the reading difficulty of written materials (Jabbari & Saghari, 2011). RDB formulas are one of these techniques that provide the solution (Ghahramani & Jabbari, 2014). RDB formula is a “predictive device intended to provide quantitative objective estimates of reading difficulty” (Klare, 1984, p. 684). Thus, RDB formula aims to provide an objective analysis about the RDB of a specific text (Jabbari & Saghari, 2011). RDB formulas were primary emerged in the 1920s and 1930s (Redish & Selzer, 1985; Loughran & McDonald, 2014). Since 1920, hundreds of formulas have emerged (Freiman, 2010), such as the Dale−Chall readability, Flesh Reading Ease, Flesch−Kincaid, FOG, Forcast, Fry, Lexile, PSK, SMOG, and Spache (Begeny & Greene, 2014). Each formula is considered as significant attributes, such as “average number of words per sentence, average number of syllables per word, the number of single-syllable words, the number of polysyllables, the number of words not on some predetermined list of so-called easy words, and the like” (Marnell, 2008, p. 1). Even though several RDB formulas have been embraced as measurement techniques in prior studies of corporate reports’ RDB, the Flesh RDB formula has become the most common and popular approach among researchers (Mohamad & Abdul Rahman, 2006; Abdul Raman et al., 2012; Rahman, 2014; Moreno & Casasola, 2016). Some of the reasons for utilizing the Flesh RDB formula are its reliability, validity, simplicity, practicability, and comparability with other researches (Courtis, 1998; Mohamad & Abdul Rahman, 2006; Rahman, 2014). Hence, it becomes the most commonly used formula as well as the most tested and reliable one (DuBay, 2004; Abdul Raman et al., 2012). For objectivity and greater comparability with prior studies, this research used Flesh RDB formula to measure the RDB of the ARs (

Table 2. Measurement of dependent variables.

Variable	Abb.	Measurement	Reference
Dependent variable
Readability	RDB	“Flesh readability scores: the chairman’s letter of each bank was copied and then pasted it in Microsoft Word 2016 to get the readability index. Once the score has been determined, it is compared with a predetermined table (Appendix A), which provides an explanation of its readability”.	Mahboub et al. (2017)

Source: developed by the researcher.

).

3.2.2. Measurement of Independent and Control Variables

Table 3. Measurement of independent and control variables.

Variable	Abb.	Measurement	Reference
Independent Variables
Board size	BS	“The total number of board members”	Muchemwa et al. (2016)
Board meetings	BM	“The total number of board meetings”	Kyei et al. (2022)
Board independence	BI	“The ratio of non-executive members to total members on the board”	Muchemwa et al. (2016)
Board gender diversity	BGD	“Dummy variable equal to 1 if a corporate board has female directors and 0 otherwise”	Jun et al. (2023)
Role duality	RD	“Dummy variable equal to 1 if the chairman is the same person as the CEO, 0 otherwise”	Aktas et al. (2019)
Ownership concentration	OC	“Ratio of the shares held by the largest shareholder to total shares”	Sousa & Galdi (2016)
Control Variables
Profitability	PROF	“Return on Assets”	Diaz and Pandey (2019)
Size	SIZ	“Log of Total Assets”	Azaro et al. (2020)
Leverage	LEV	“Debt to Equity Ratio”	Ertugrul et al. (2017)

Source: developed by the researcher.

provides a comprehensive overview of the independent and control variables and their corresponding proxies.

3.3. Research Model

In order to test the hypotheses, a model using “ordinary least squares” (OLS) is formulated in the following manner:

RDB = β0 + β1 BS + β2 BM + β3 BI + β4 BGD + β5 RD+ β6 OC + β7 PROF+ β8 SIZ+ β9 LEV + ε

where

RDB is the dependent variable (readability of the annual reports);

BS is the first dependent variable (board size);

BM is the second dependent variable (board meetings);

BI is the third dependent variable (board independence);

BGD is the fourth dependent variable (board gender diversity);

RD is the fifth dependent variable (role duality);

OC is the sixth dependent variable (ownership concentration);

PROF is the first control variable (profitability);

SIZ is the second control variable (bank size);

LEV is the third control variable (leverage);

ε is the error term.

4. Results and Discussion

4.1. Descriptive Statistics

Table 4. Descriptive statistics.

	Min	Max	Mean	SD	Skewness	Kurtosis
RDB	17.90	43.20	32.0547	8.42349	−0.314	−1.140
BS	6.00	12.00	9.4084	1.38006	−0.935	0.822
BM	4.00	9.00	6.5916	1.12768	0.232	−0.025
BI	0.20	0.91	0.6646	0.16861	−0.727	−0.439
OC	0.17	0.70	0.4371	0.16679	0.079	−1.109
BGD	N			%
The board has female directors.	146			0.69%
The board has no female directors.	329			0.31%
RD	N			%
The chairman is the same person as the CEO.	137			0.71%
The chairman is not the same person as the CEO.	338			0.29%

Source: SPSS outputs (25).

shows the descriptive statistics for the quantitative variables (RDB, BS, BM, BI, and OC) and frequency distributions for dummy variables (BGD and RD). Like most previous studies, MENA banks were more likely to have more complex ARs. The average Flesh score was 32, which entailed that the RDB level of the banks was very low. This showed that ARs were difficult to read. Further, the results indicated that, on average, the banks had boards composed of nine members, who were mostly NEs and the largest shareholders, and met around six times per year. About 71% of MENA banks had a board chairman who played a dual role. Additionally, about 31% of the board had no female directors.

4.2. Correlation Analysis

Table 5. Correlation analysis and VIFs.

	BS	BM	BI	BGD	RD	OC	RDB	VIF
BS	1							1.473
BM	−0.124 **	1						1.224
BI	−0.337 **	0.010	1					1.476
BGD	−0.240 **	−0.046	0.315 **	1				1.238
RD	0.229 **	−0.206 **	−0.324 **	−0.072	1			2.193
OC	−0.503 **	−0.075	0.548 **	0.357 **	−0.653 **	1		3.347
RDB	0.512 **	−0.023	−0.356 **	−0.371 **	−0.059	−0.143 **	1

** Significance at a p-value of 0.05. Source: SPSS outputs (25).

exhibits the correlation matrix. RDB was positively correlated with BS, while it was negatively correlated with BI, BGD, and OC. Further, OC was negatively correlated with both BS and RD, while it was positively correlated with BI and BGD. Moreover, RD was negatively correlated with both BMs and BI, while it was positively correlated with BS. Additionally, BGD was negatively correlated with BS, while it was positively correlated with BI. Moreover, BS was negatively correlated with both BI and BMs. The variance inflation factors (VIFs) of the variables ranged from 1.224 to 3.347. This indicated that there was no multi-collinearity issue. This is because when VIF is less than ten, there will be no multi-collinearity in the data (Mahboub, 2021).

4.3. Regression Analysis

Table 6. Regression analysis.

Variables	OLS Model
Constant	0.430 (0.668)
BS	12.999 (0.000 *)
BM	1.809 (0.071 *)
BI	−6.937 (0.000 *)
BGD	−7.206 (0.000 *)
RD	−0.962 (0.337)
OC	5.772 (0.000 *)
PROF	0.884 (0.377)
SIZ	−3.393 (0.001 *)
LEV	−0.628 (0.531)
F-Statistics	43.289 (0.000)
R-Squared	0.456
Adj-R Squared	0.445
N	575

*: Significance at 0.05; Source: SPSS outputs (25).

presents the findings of the regression model that examined the impact of CG structures on the RDB of ARs in banks operating in the MENA region.

The model’s validity for drawing inference was confirmed by the statistically significant F-statistic value of 43.289 (0.000) at a 5% significance level. Regarding the coefficient of determination (R-squared), it was observed that 45.6% of the systematic variations in ARs’ RDB were accounted for by the independent and control variables in the model during the study period. This suggests that the CG mechanisms employed in this study cannot fully explain the variations in ARs’ RDB of MENA banks, as they only account for 45.6% of the total variability. The remaining unexplained changes in ARs’ RDB is attributable to other independent variables that were excluded from the analysis but were recorded as error terms.

The multivariate analysis has shown that five aspects of CG structures greatly influenced the RDB level of ARs for banks in the MENA region. The findings confirm hypotheses H1, H2, H3, H4, and H6. However, the multivariate analysis did not provide evidence to support that the RD had an impact on the RDB level of ARs in the MENA context. Consequently, hypothesis H5 was not supported. Additionally, none of the control variables showed a significant association with RDB, except for the size of the bank.

The findings indicated that BS significantly influenced the RDB of ARs (t = 12.999; p-value = 0.000 < 0.05). This suggests that an increase in the involvement of board members enhances the comprehensibility of the bank’s AR, as the BOD is responsible for ensuring that the information presented in the AR is unbiased, equitable, and easy to understand. Therefore, these results provide support for hypothesis H1. These results align with the research conducted by Etuk and Akpan (2023), who demonstrated that a larger BS leads to improved RDB of ARs for oil and gas companies listed in Nigeria. However, they contradict the findings of Ginesti et al. (2017), Ezat (2019), and Dalwai et al. (2023), who were unable to provide empirical evidence supporting the relationship between BS and ARs’ RDB.

Moreover, the results demonstrated a positive impact of BMs and the RDB of ARs (t = 1.809; p-value = 0.071 < 0.1). This suggests that financial institutions with a higher number of BM tend to produce more comprehensible reports. This outcome can be explained by “agency theory”, which posits that an increased number of BMs enhances the supervision of management by the board and enables directors to fulfill their responsibilities effectively, leading to the disclosure of clearer and more understandable information. Therefore, these findings provided support for H2. This finding aligns with the research conducted by Prusty and Kumar (2018), who found a significant relationship between the ease of reading business responsibility reports and BMs in India. However, it contradicts the findings of Ezat (2019), who identified a negative association between BMs and RDB, and those of Astari et al. (2020), who reported that BMs do not influence the RDB of ARs.

Furthermore, the findings indicated that BI has a significant negative impact on the RDB of AR (t = −6.937; p-value = 0.000 < 0.05). The results revealed that a higher degree of BI leads to a decrease in RDB. This implies that having independent boards may not necessarily promote better disclosure quality, as managers might intentionally obscure information to evade rigorous board oversight (as per the avoidance of monitoring hypothesis). Consequently, these results lend support to Hypothesis 3. These findings align with those of Rahman and Kabir (2023), who demonstrated that BI diminishes the RDB of AR. However, they contradict the findings of Ezat (2019), who failed to provide empirical evidence for the relationship between BI and ARs’ RDB and Gosselin et al. (2021) who observed that BI is positively and significantly linked to the RDB of AR.

Moreover, the findings indicated that BGD had a negative impact on the RDB of ARs (t = −7.206; p-value = 0.000 < 0.05). This result can be attributed that a higher level of GD within the board leads to an increase in divergent opinions and challenging inquiries, which in turn escalates the probability of conflicts, diminishes unity, contentment, and dedication. Additionally, GD introduces a variety of English writing styles, which in turn diminishes the RDB of ARs, as board members with different linguistic backgrounds contribute varied writing inputs to the ARs, making it challenging to harmonize these diverse writing styles during the AR preparation process, consequently impeding the creation of a coherent and easily understandable AR. Therefore, the results provide support for Hypothesis 4. These results are in contrast with the findings of Ginesti et al. (2018), Nadeem (2022), and Ben-Amar et al. (2024) who suggested that GD in the board enhances the RDB of ND in ARs. Similarly, the results contradict the study by Shauki and Oktavini (2022) which indicated that the presence of female directors in a company does not influence the RDB of ARs.

Moreover, the findings indicated that the presence of OC has a significant positive impact on the RDB of ARs (t = 5.772; p-value = 0.000 < 0.05). Drawing upon “active supervision theory”, it can be deduced that OC enhances the oversight of managers and their subordinates, thereby preventing the concealment of unfavorable information. Consequently, this ultimately enhances the clarity of their ARs. Thus, the results lend support to H6. These results align with the research conducted by Kumar (2014), Khan et al. (2017), Velte (2018), Melón-Izco et al. (2021), and Dalwai et al. (2023), who all reported that dispersed ownership leads to more easily understandable reports for banks. However, these findings contradict the evidence provided by Etuk and Akpan (2023), who found that OC has an insignificant impact on the RDB of ARs.

4.4. Robustness Check

To understand the impact of CG structures on RDB of ARs more rigorously and to contribute a robust result, the main result findings were tested further. In so doing, a “robustness test” was applied for the model. In the literature, a significant number of prior studies examine the impact of CG structures on RDB by considering the “Flesch−Kincaid Grade Level” (FKG) (Huong Dau et al., 2024; Worrall et al., 2020; Solnyshkina et al., 2017). Hence, as a “test of robustness”, the impact of CG structures on RDB was assessed by replacing “Flesh Readability Ease score” with the FKG. Thus, the research model was altered into the following model, and detailed results are presented in

Table 7. Robustness check.

Variables	OLS Model
Constant	−0.494 (0.622)
BS	12.997 (0.000 *)
BM	1.784 (0.075 *)
BI	−7.144 (0.000 *)
BGD	−7.482 (0.000 *)
RD	−0.394 (0.694)
OC	6.192 (0.000 *)
PROF	1.680 (0.094)
SIZ	−3.114 (0.002 *)
LEV	−0.120 (0.904)

*: Significance at 0.05; Source: SPSS outputs (25).

.

FKG = β0 + β1 BS + β2 BM + β3 BI + β4 BGD + β5 RD+ β6 OC + β7 PROF+ β8 SIZ+ β9 LEV + ε

Based on the results in Table 7, the impacts of CG structures on RD of ARs were similar, with some exceptions for control variable PROF (t = 1.680; p-value = 0.094 < 0.1). Therefore, CG structures of MENA banks play a significant role in making their ARs readable while managing, monitoring and controlling the bank.

5. Conclusions and Recommendations

This study examined whether CG structures have an impact on the RDB of the CL of banks operating in the MENA region. The findings of this study indicate a positive impact of BS on the RDB level of banks, suggesting a larger BS can improve the RDB of ARs. BMs positively influence the RDB of ARs, thus indicating easy-to-read reports. BI negatively influences the RDB of ARs, indicating a higher level of BI decreases RDB of ARs. Similarly, BGD has a negative association with the RDB of ARs, demonstrating diverse board members hinders the production of readable ARs. Inversely, OC has a positive impact on the RDB of ARs, implying OC improves the RDB of ARs. These findings have significant implications. The study contributes to existing research by examining the linguistic complexity of disclosure reports in the MENA region, from a theoretical standpoint. From a practical standpoint, this study will be beneficial for bank managers as it enables them to effectively communicate necessary information by understanding their report-writing style. It is important for MENA managers to acknowledge potential difficulties that users may face when reading their reports, as this can result in a breakdown of communication with stakeholders. Hence, managers should be mindful of the RDB level of their ARs. The study is subject to certain constraints. Initially, the study was centered on 95 banks across seven countries in the MENA region spanning from 2018 to 2022. Subsequent studies could broaden the scope by expanding the sample size, extending the time frame and including additional countries and additional industries. Secondly, the assessment of RDB levels was based on Flesh RDB scores. Future investigations could incorporate alternative measures of RDB, such as the FOG index. Lastly, the study delved into the RDB levels of AR for MENA banks; future research could explore the RDB levels of other types of reports, such as board reports and CSR reports. The focus of the study was on certain CG structures, while other factors like board duration and board composition diversity could be further explored. The study suggested that banks’ management should contemplate augmenting the frequency of their board meetings to a minimum of six gatherings annually. Additionally, it was proposed that the board’s size should be significantly expanded to enhance the RDB of ARs. Furthermore, it is crucial to have a certain number of independent directors on the company’s board, although an excessive amount should be avoided. Moreover, the appointment of women to boards is advocated to promote gender equality, rather than solely for the purpose of enhancing company performance through GD. Lastly, the study emphasized the necessity for management in the MENA region’s commercial banks to amplify OC.