Spotlight on Corporate Fraud: How Is Takaful Insurance Stability Affected by Its Disclosure?

Abstract

This study examines the influence of fraud disclosure (FR_DISC) in annual reports on the financial stability of Takaful insurance (TKI) in Saudi Arabia over the period of 2014 to 2022. Moreover, the current study aims to explore the mediating impact of Shariah board size in shaping this relationship using agency theory and examines whether the different Islamic governance attributes could affect this stability differently. Using the dynamic generalized method of moments (GMM) approach to address the possibility of endogeneity, it was found that FR_DISC is significantly negatively related to the financial stability of a sample TKI. We also provide evidence that the larger the size of a Shariah board, the less FR_DISC affects TKI stability. Furthermore, significant negative influence of ownership concentration and the proportion of non-executives’ independent board members on the stability of insurance companies was also observed. Overall, our analysis reveals several significant challenges if accounting and whistleblowing are to contribute to financial stability.

1. Introduction

Tackling fraud acts in insurance companies remains a strategic breakthrough. Insurance fraud (hereafter IF) cuts across multiple activities and business lines within insurance and across different entities in an insurance group. Ernst and Young’s 2011 European fraud survey found that the root causes of IF can be split into three distinct types: (i) policyholders (or a third party) fraud, where they encounter an opportunity to invent or exaggerate a claim or to deliberately or recklessly give false or misleading information when applying for insurance; (ii) intermediary fraud, which is committed against the insurer or policyholders by independent brokers or agents who may be fraudsters, diverting premiums and, in addition, could falsify records; and (iii) internal fraud, involving managers and/or other staff members or officers. However, IF remains a taboo topic, and this often leaves many wondering about anti-fraud implementation and enforcement strategies. Apart from the lack of available quantitative data that include required information about fraud victimization, Sánchez-Aguayo et al. (2021) suggested that the complex nature of IF is an important leading cause that hampers the accumulation of literature.

As opposed to other non-financial risk information, crime information not only allows investors to have the most holistic understanding of the theft or misappropriation of a company’s assets, but it may also lead to a sudden and substantial drop in stock prices and/or investor decision making and welfare (AICPA 2002). It seems natural to believe that, in the case of insurance context, fraud losses were embedded into an insurance’s culture. Therefore, the central idea that suggests that informed insurance managers tend to disclose positive information and hide negative information is not treated as valid evidence in the context of fraud. Thus, if fraud information is not disclosed, the insurance manager may be suspect of hiding unfavorable information and incur reputation or litigation costs (Frenkel et al. 2020). Consequently, fraud disclosure (FR_DISC) tends to be rather opaque (Arum and Wisdianti 2020), which inhibits its usefulness for insurance investors, because it could even exacerbate the speculative cycles of bubbles and crashes generated in the insurance sector.

Prior studies have revealed that investors who deem fraud risk assessment to matter in investment decision making make greater use of fraud red flags to be better aware of fraudulent investment schemes (Hoberg and Lewis 2017). In addition, insurance policyholders should avoid worthless policies written by fraudulent companies. Gurun et al. (2018) argued that the trust shock in financial intermediaries could lead depositors, retail investors, and policyholders to update their beliefs about confidence that one’s assets will not be stolen and are “in good hands”, causing them to withdraw their money from delegated managers in favor of safe investments. Importantly, FR_DISC may not only influence financial instability due to the alienation of insurers from investors and policyholders, but it may also be shaped by other issues.

The International Association of Insurance Supervisors reports that FR_DISC may initially cause direct losses for insurers, and this fact can be exploited to raise premiums, leading to high pressure on financial stability. Beyond fraud loss consequences, there has also been evidence that harmful effects might continue even after attempted fraud is discovered and disclosed. Thus, the range of consequences that is colloquially subsumed under risk disclosure is far broader. Furthermore, Haß et al. (2015) suggested that discovering fraud can have a crucial impact on the optimal auditing strategy and vice versa. According to Jung and Kim (2021), IF may initially cause direct damage, which may be ultimately shifted to policyholders by raising premiums. More recently, Kasoga and Tegambwage (2023) found that policyholder fraud, internal fraud, and intermediary fraud hindered the effective performance of insurance.

Exploring the influence of FR_DISC on insurance stability is theoretically and practically significant. In terms of theoretical importance, attribution studies from agency theory (Ho et al. 2023; Mamahit and Urumsah 2020) suggest that if investors do not have access to accurate and reliable fraud information, a company runs the risk of not being able to sustain itself, leading to deficiency. Legitimacy theory, associated with FR_DISC, emphasizes that an organization must seem to consider the rights of the public at large, not merely those of its investors. Failure to signal integrity may lead to sanctions being imposed by society and can be devastating. These sanctions are enforced through various mechanisms, such as legal restrictions imposed on a company’s operations or by investors and policyholders through limited financial resources and contributions collected from policyholders. Furthermore, they can accentuate instability. From a deontic view, the moral justification of reporting fraud refers to a special type of practice that would further liberalize planned self-correction rules vis-à-vis forms of wrongdoing (Ceva and Bocchiola 2019). Thus, this kind of disclosure keeps a company on the right track through the interrelated actions of internal and external stakeholders, and ultimately ensures financial stability. Furthermore, it serves as a shield with which to protect organizations from legal and reputational risks, and investors are more inclined to look to individuals, such as directors and officers, and trust them rather than “faceless” corporations. However, the rising costs associated with this kind of disclosure can negatively affect price informativeness and volume within rational equilibria, and may cause some market makers to exist due to the higher cost of capital, thereby lowering a firm’s value (Cheynel 2013). Furthermore, “unwarranted fraud disclosure” can also have an even more direct undesired effect on financial stability, followed by the higher underpricing of financial funds. From a practical perspective, fraud information in the literal sense of the term will trigger serious public opinion and business reputation crises, impeding the going concern status and ultimately lead to a crash in a company’s stock price.

The heavy focus on conventional corporate finance in the public debate has diverted attention to non-financial information reporting issues in Islamic ones that also matter for financial stability. This may seem a bit controversial, and it seems reasonable to expect that this will be brought to Takaful insurance (TKI) with regard to its legal basis, policy design, and ethical as well as moral standards. The major principles of Takaful insurance, which differ from those of conventional insurance, are a ban on interest (riba), a ban on uncertainty (gharar), adherence to risk sharing and profit sharing, the promotion of ethical investments, and asset backing. TKI is required, in addition to standard audits, to rely on Shariah principles that would review and improve communication about and awareness of the potential for fraud and corruption for compliance with Shariah governance policies (Rahman and Anwar 2014). This exclusive compliance characteristic of Islamic finance should be recognized as a credible prevention method of fraud occurring and being covered up. However, in reality, Shariah principles do not necessarily make TKI avoid and/or disclose fraud cases. As a matter of fact, fraud committed in an Islamic entity provides conditions under which the market, on average, reacts more to this bad information (Hemrit and Belgacem 2024; Grassa et al. 2020). This is due to the potential massive risk of the sudden withdrawal of funds by both policyholders and shareholders who own the equity-based capital in cases of investment in unlawful businesses, a lack of transparency, and acts of irregularity or omission, which threatens financial stability. This is particularly important as TKI is subject to a patchwork of business and religious rules that differ across jurisdictions, reflecting the varied and fragmented nature of the insurance industry. Albeit TKI is often considered more stable than conventional insurance, empirical evidence to support the stability view has the potential to be misleading if TKI does not ensure transparency.

Only a very limited number of studies have been conducted to address the issue of suspected practices of illegal or unethical conduct reporting in insurance and its effect on its financial stability. At this time, the majority of work on fraud in financial institutions has been centered on the fraud–misstatements arising from fraudulent financial reporting and misstatements–financial stability nexus (see, Achmad et al. 2023; Shams et al. 2020). Additionally, the majority of studies have mainly focused on assessments of Islamic banks, with no study having examined the impact of FR_DISC on the financial stability of TKI. This may be due to the inherent difficulty in determining the extent of fraud within the varied business lines and plenty of agents engaging in dubious practices or outright fraud. Accordingly, there is a paucity of conceptual research focusing on theoretical frameworks underpinning the incentives to provide specific fraud-factor disclosures in this industry. In light of previous research, this study addressed the issue of fraud prevention and detection in public and private institutions in Saudi Arabia by examining the influence of fraud disclosure (FR_DISC) in annual reports on the financial stability of Takaful insurance (TKI) in Saudi Arabia.

This study targets a real-world issue and contributes to the literature on corporate fraud disclosure literature in a number of ways. First, the purpose of this research is to find out the extent to which FR_DISC can influence financial stability in TKI operating in Saudi Arabia. The current study is set out in the Saudi Arabian context, driven by the following foundational principles, underpinning its robust academic significance: (i) In 2023, Saudi Arabia was the largest insurance market within the Gulf Cooperation Council region, with a combined profit of USD 894 m, according to the Middle East Insurance Review report (2024). In conjunction with this development, this sector has seen a surge in financial fraud owing to rapid developments in financial and non-oil spaces in the past few years (Hemrit and Benlagha 2020). (ii) In line with Vision 2030, the insurance sector in Saudi Arabia is undergoing significant transformation as it embraces a counter-corruption model (Saudi Central Bank—SAMA 2023) and wide-ranging transparency as well as accountability mechanisms in accordance with Art- 42 under the Civil Transactions Law, 2023. (iii) A new analysis of insurance coverage in Saudi Arabia reveals a complex and troubling picture of insurance instability due to the vibrant environment and depressed oil revenue (Orlando and Bace 2021). This slowdown has recently been compounded by the COVID-19 pandemic. Second, our study is also novel in that it explores financial stability as a function of multi-dimensional insurers’ and Islamic corporate governance attributes that can create an imperative for change in business and risk-related regulations in Saudi Arabia. Third, as Shariah boards play a significant role in internal governance, which aims at transparency in the disclosure of information (Neifar et al. 2020), exploring the mediating role of the fraud disclosure–financial stability linkage is imperative. By examining this research problem, we can provide a new addition to the prior literature by testing the relationship between FR_DISC and financial stability and the interaction effect of Shariah boards. The remainder of the paper is as follows: Section 2 is the literature review. Section 3 presents the methodology and data, while Section 4 describes the empirical findings and Section 5 concludes the study.

2. Literature Review

2.1. Theoretical Overview

Many theoretical approaches have been proposed to justify FR_DISC practices: (i) an economic theory approach and (ii) a social as well as political theory approach. From a pure profit-maximization perspective, the first theoretical framework that underpins the discourse on FR_DISC relies on self-interested agents who seek to maximize personal economic wealth (agency theory, signaling theory, and proprietary costs theory). A central premise of agency theory is that social information is disclosed to increase the welfare of management. When the quality of agents facing a positive probability of internal or external fraud occurring is not easily verified by principals, an agent may engage in fraudulent activities with relative impunity (Ndofor et al. 2013). Because managers bear the costs, they wish to be seen to be acting in shareholders’ interests (Jensen and Meckling 1976). The underlying intuition is that, to deflect attention away from themselves, managers attempt a kind of word “shell game”, through which they exhibit certain aspects of fraud detection and management (Hoberg and Lewis 2017). Firms that refuse to protect shareholders, whether by disclosing or otherwise, would face a higher cost of capital and lower market visibility (Diamond and Verrecchia 1991). Over time, such changes push up financing costs and make it much more difficult for firms to produce attractive risk-adjusted returns, which can have a potentially sizable macroeconomic effect that could adversely affect financial stability. Drawing from signaling theory, disclosure is generally beneficial for companies in reducing severe information asymmetries (Spence 1973). However, fraud risk differs from other risks as these intentional misconducts are specifically intended to evade detection, as a Deloitte report (2021) states. Therefore, FR-DISC can negatively influence core stakeholders’ reception and interpretation of fraud signals (Ren et al. 2022). Recently, Dai et al. (2024) argued that each investor can observe a noisy signal of systemic vulnerability disclosure and then decide whether to run, which adversely disturbs the whole financial sector, and financial prospects are poor. Another relevant strand of research suggests that, even though corporate fraud information is a negative signal, its disclosure before it becomes public can enhance its stability and the legitimization of an organization within society (Handajani et al. 2022; Pittroff 2014). To conclude, the taxonomy can serve as a building block for further applications of signaling theory in FR-DISC research. When applying proprietary costs theory (Verrecchia 1983), researchers have suggested that the communication of proprietary information could compromise a company’s competitive position by providing strategic information to competitors. The latter asserts fraud “as bad news”, and thus disclosing entities could also trigger a successful strike by opponents and support higher proprietary costs (Gago-Rodríguez et al. 2020). Prencipe (2004) suggests that benefits that can accrue from greater risk disclosure can be completely offset by these costs.

Apart from economic theories, some other theories for FR-DISC and its effect have been proposed (such as political costs theory and legitimacy theory). In terms of the social and political theory approach, many researchers focus on the role of information and disclosure in the relationships between organizations, state, and society (Gray et al. 1995). The political costs theory states that firms may be subject to deep scrutiny from politics by way of corporate taxes, regulations, subsidies, etc. (Watts and Zimmerman 1978). An intuitive view suggests that fraud in insurance (whether internal or external) is a common and inevitable problem (Hemrit and Belgacem 2024). This is especially true when it comes to dealing with uncertainty, potential hazards, and opportunities related to the nature and scope of insurance coverage.

To be specific, insurers may disclose fraud information in order to reduce the chance of more detailed and perhaps more costly requirements being introduced by regulation in an effort to limit IF. Clearly, as with some other Islamic organizations, TKI invests heavily in developing political capital and spends significant sums on political contributions and charitable nonprofit organizations. However, despite these investments, policymakers can impose some funding restrictions and place burdens on the underwriting process (Cuny et al. 2022). In terms of the legitimacy view, the implementation of an FR-DISC system should be a result of societal demand (Pittroff 2014). In this context, organizations can be motivated to make FR-DISC for two different arguments: first, compliance with minimum mandatory fraud reporting requirements may result in the revocation of a company’s ability to operate in society, leading to enhanced financial well-being in the long term (Martens and Bui 2023), and second, sharing the same objective as the community will increase perceived authenticity by stakeholders and enhance acceptance and legitimization within society (Lindblom 2007). Accordingly, the following sub-hypothesis was formulated to examine the effect of FR_DISC on the financial stability of TKIs.

H1. 

Fraud disclosure has a significant and positive impact on the financial stability of TKI in Saudi Arabia.

2.2. Hypotheses

2.2.1. Corporate Governance Attributes

Ownership Concentration (OWC)

The significance of ownership in the financial stability of TKIs is not well documented in the corporate governance and insurance literature (Kader et al. 2014). The theoretical framework of ownership structure is constructed based on agency theory, which elucidates the conduct of relevant parties within the dynamic between shareholders and managers (Shleifer and Vishny 1997). As suggested by La Porta et al. (1999), due to the majority shareholders’ practice of the “tunnelling of resources” at the expense of others, creditors and minority shareholders may be reluctant to invest or, if they do so, they will impose higher costs on capital, leading to underinvestment and market disruption. Grassa et al. (2020) contend that sizable stakes haunted by large outside shareholders give them the possibility to exert governance through two fundamental characteristics: (i) the power to set and implement their chosen business strategy, and (ii) the ability to manipulate a firm’s stock price and punish managers for their potential misbehavior. Rubio-Misas (2020) also asserts that, in firms with more concentrated ownership structures, the alignment of owner and agent interests may also be an incentive for risk taking. More recently, Hemrit and Belgacem (2024) argued that companies with dispersed ownership tend to be more prone to conflict, thereby leading to a decline in financial stability. Within this context, one may suppose that TKI with a more concentrated ownership structure tends to be less stable financially. We posit the following:

H2. 

Ownership concentration has a negative impact on the financial stability of TKI.

Board Size (BSZ)

There is evidence in the literature suggesting that when making decisions regarding the size of a board, corporate decisionmakers take into consideration both resource dependency and agency theories (Goyal and Gulati 2023). The existing literature comprehensively discusses the importance of board size as a bank of resources (Akram and Ul Haq 2022) and financial stability. Most studies have reached the conclusion that there is a positive association between them. For instance, Adams and Kastrinaki (2023) found that insurers with more board members can instill stability by bringing business knowledge and technical expertise. According to Ali (2018), board members can also offer valuable advice in establishing various insurance contracts sustaining a higher ratio of premiums to shareholders’ surplus. Other studies suggest that large boards may be less cohesive and might not necessarily contribute to financial stability (Lipton and Lorsch 1992) due to agency problems to reach a consensus (Cheng 2008), procedural issues, and free riding among directors. Jensen (1993) stipulated that CEO power in decision making increases with board size. Consequently, this leads to ineffective monitoring, implying higher instability. Thus, empirical evidence regarding the relationship between board size and insurance financial stability also remains mixed. Accordingly, we state the following testable hypothesis:

H3. 

The board committee size influences TKI financial stability (without perceived sign).

Proportion of Non-Executive Independent Members (PNEXM)

It is widely held that independent members are better suited than members of top management for their monitoring role; hence, they might maximize share value and mitigate agency problems (Fama and Jensen 1983). This vital role in upholding and promoting good corporate governance practice is well supported by modern theories, while empirical evidence has not reached a consensus about its potential to make objective decisions. A common assumption in legitimacy theory is that the more independent a board of directors, the more a company maintains its legitimacy and the financial sustainability of its business. It is ultimately responsible for “meeting the aspirations of various stakeholders…” (Al Amosh and Khatib 2022, p. 54) and for ensuring insurers’ allocative and financial performance (Sallemi and Zouari 2024). Moreover, the presence of non-executive members escalates the share value of TKI (Hemrit 2020) and leads to lower mangement earnings (Zhu et al. 2016). In the banking context, Karkowska and Acedański (2020) argued that banks with more independent directors might be a hidden source of high probability of default. However, some studies found a negative relationship between the presence of non-executive independent members on a board and financial stability (Asare et al. 2023; Nguyen et al. 2022; Marie et al. 2021), while some studies did not find any relationship between them (e.g., Wang and Hsu 2013). To explore this relationship in the TKI sector, a fourth hypothesis has been formulated, after taking into account the legitimacy theory predictions:

H4. 

The extent of TKI stability is positively associated with the proportion of non-executive independent members.

Board Meeting Frequency (BMFR)

BMFR is a major means of enhancing loyalty and moral values among board members (Vafeas 2003). According to resource dependency theory, board meeting attendance makes a company’s essential resources available by reducing environmental interdependence and uncertainties. Hossain and Oon (2022) found that a higher BMFR leads to increased monitoring through closer control over inside directors as well as greater advisory roles and oversight to the board of directors. Vafeas (1999) suggests that high meeting frequencies are a reactive rather than proactive value-enhancing strategy for improved governance. Additionally, meetings also have the benefit of strengthening cohesion among members, resulting in better corporate governance outcomes, which could improve sustainable growth and ensure market integrity as well as financial stability (Ji et al. 2020). An opposing view is that more meetings may also result in the poor quality of enterprise performance information, thus decreasing future corporate value (Kakanda et al. 2017). Furthermore, refreshments, higher traveling expenses, and directors’ meeting fees, which are related to the meetings, increase agency costs and adversely affect firm stability (Alsartawi 2019). We state our hypothesis as follows:

H5. 

There is a significant positive association between the frequency of TKI board meetings and its financial stability.

Audit Committee Size (ACSZ)

Being the first line of defense in safeguarding organizational transparency through financial statements, it is argued that the efficiency of an audit committee is enhanced by a limited number of other boards, as it reduces information asymmetry and safeguards investors, which improve firms’ stability (Kamaludin et al. 2023). More importantly, the literature emphasizes that firms should set up a limited number of audit committee members without considering shareholders’ primacy (Bazhair 2022). As a result, an audit committee becomes more capable of enhancing firm value and top management. In favor of agency theory, if the size of an audit committee is large, a company would realize poor financial performance. This evidence is consistent with arguments put forward by Ghafran and O’Sullivan (2017), who suggested that a large audit committee probably has many experts, which would eventually lead to discrepancies between audit cost dimensions and abnormal fees. Ettredge et al. (2007) assumed that these fees could lead to an auditor permitting management to engage in aggressive earnings management, threatening financial stability. Therefore, this research study hypothesized the following:

H6. 

Audit committee size has a positive effect on financial stability.

Moderating Role of Shariah Board Committee Size

Saudi Arabia has a comprehensive legal and regulatory framework regulating corporate and Shariah governance in TKI, suggesting Saudi’s promise to endorse effective Shariah monitoring in the insurance sector (Mukhibad et al. 2022). The dual board structure in TKIs can be explained as two boards: a board of directors and a Shariah board. Shariah board governance (SBG) is examined in the literature in terms of several characteristics (such as size, qualification, independence, reputation, expertise, and cross-membership). The social norm theory posits that religion impacts individuals’ actions and choices by creating shared beliefs and values (Raphael and Macfie 1976). As a result, managers in religious societies might employ SBG as a key feature of fraud prevention and risk disclosure. As discussed recently, Hemrit and Belgacem (2024) argued that Shariah board size (SBSZ) is the most important governance mechanism in TKI. It serves not only to instill shareholder confidence in the purity of their insurance and investment operations, but also to have the strength and authority to ensure that the senior level of managers behaves in a way that makes their disclosures more effective. Signaling theory predicts that Islamic companies communicate fraud information to outsiders in order to signal to potential investors insurers’ apparent sound fraud risk management practices and performance. From the above theory, TKI fraud disclosure increases when employing more Shariah board members, because they are inherently conservative institutions, when it comes to fraud risk.

Few studies have examined the moderator role of SBSZ in the association between risk disclosure and financial performance and/or stability in Islamic financial institutions (Neifar et al. 2020; Nomran and Haron 2020). The resource dependence theory argues that a large board size can connect a firm’s external environment more effectively and improve access to resources. In the TKI context, Eldaia et al. (2022) found evidence that SBSZ combined with a high level of board effectiveness improves Takaful performance. More recently, Hemrit and Belgacem (2024) confirmed a negative relationship between the number of Shariah scholars on a board and fraud disclosure in TKI, which protects stakeholder interests. Another reason for using SBSZ as a moderating variable is that Shariah boards are among the most consequential Islamic governance mechanisms to promote end-to-end compliance with Shariah requirements. Therefore, their role as a moderator in the FR_DISC and financial stability nexus in Takaful companies could be used to shed light on the role of SBSZ in boosting TKI stability. In fact, although there are few studies investigating the impact of SBSZ on TKI (Sallemi and Zouari 2024; Hemrit 2020; Rubio-Misas 2020; Kader et al. 2014), there is a dearth of literature dealing with the impact of SBSZ on the relationship between FR_DISC and the stability of TKI.

H7. 

Shariah board committee size moderates the relationship between FR_DISC and financial stability.

2.2.2. Control Variables

In recent years, the literature on TKI has paid attention to empirically understanding whether insurance size (INSZ) is relevant to explaining the performance and stability of Islamic insurance. Size is largely used as a proxy for economies of scale. The costs of calculating a premium rate, policy classification, and paying a claim increase at a lower rate than the increase in an insurer’s output. Zinyoro and Aziakpono (2023) found evidence that size is the major firm-level determinant of life insurer financial strength. Lee (2023) argued that size is understood as sign of internal control material strengthening in the financial stability of insurers, and can therefore present as significant. Another control variable is the reinsurance ceded (RCD), which is considered as the extent to which an insurance company relies on ceding risk to reinsurers. The Solvency II framework fully recognizes reinsurance’s risk-reduction effect and allows its use as an alternative to traditional forms of capital when seeking to raise some specific coverage needs. This notion of stabilizing insurers’ loss experience is related to the finding in Mao et al. (2017) that reinsurers hedge a primary insurer’s underwriting risk. Boonen et al. (2021) posited that multiple reinsurers complicate the pricing of reinsurance contracts but provide a higher flexibility of trading for insurers, leading to the stability of the insurance market. However, Lee and Lee (2012) found that insurers with a high retention level outperform those with a low retention level.

Thus, the following hypotheses are proposed:

H8. 

Size has a positive effect on the financial stability of TKI.

H9. 

Reinsurance ceded has a positive effect on the financial stability of TKI.

3. Research Design

3.1. Data and Sample Selection

The study’s initial sample size consisted of 29 TKI organizations, yet 3 companies were excluded due to data unavailability. Thus, the final sample size consisted of 26 TKI organizations operating in Saudi Arabia and licensed by the SAMA. Data were collected over 9 years, from 2014 through 2022, and data sources included TKI organizations’ annual reports and websites (www.saudiexchange.sa, [accessed date 15 May 2023] as well as www.argaam.com [accessed date 29 July 2023]). This leaves us with 234 observations. This study period encompasses many important events that shaped the future of the financial sector in Saudi Arabia, such as international financial as well as pandemic crises (such as the oil price collapses in 2014, 2017, and 2020, the COVID-19 crisis, and the Russo-Ukrainian War) and patterns of structural transformation in the Saudi economy (such as the launch of Vision 2030 in 2016, the introduction of the value-added tax (VAT) in 2018 and its increase from 5 to 15% in 2020, and the enforcement of the Saudi Anti-Corruption Act in 2017).

3.2. Variable Measurement

In our study, we have one independent variable, eight dependent variables, and one moderator. These variables are measured and defined in

Table 1. Summary of the measurements of the dependent and control variables.

Explanatory Variable	Definition	Measurement	Sources	Expected Sign
Dependent Variable
Z-score	Insurance’s distance from insolvency.	${\displaystyle \frac{ROA+(Equity/Assets)}{\sigma \left(ROA\right)}}$	(Moreno et al. 2022).
Independent Variables
FR_DISC	Index of the quantity of Fraud disclosure.	Index measuring the quantity of FR_DISC for TKI.	(Hemrit and Belgacem 2024).	+
OWC	Ownership concentration.	Number of blockholders—shareholders whose ownership is ≥5% of the total number of shares issued.	(Elfeky 2017).	-
PNEXM	Proportion of non-executive independent members.	Ratio of the number of non-executive directors to the total number of directors.	(Karkowska and Acedański 2020; Ji et al. 2020).	+
BMFR	Board meeting frequency.	Natural logarithm of the number of board meetings.	(Bazhair 2022; Hemrit and Belgacem 2024).	+
ACSZ	Audit committee size.	Natural logarithm of the number of members of the audit committee.	(Hemrit 2020; Ghafran and O’Sullivan 2017).	+
SBSZ	Shariah board committee size.	Number of Shariah committee members.	(Eldaia et al. 2022; Grassa et al. 2020).	+
Control Variables
INSZ	Size.	Natural logarithm of total assets.	(Lee 2023; Hemrit 2020; Grassa et al. 2020).	+
RCD	Reinsurance ceded.	The total reinsurance ceded divided by gross premiums written.	(Lee and Lee 2012)	+

Source: authors’ own creation.

.

3.2.1. Dependent Variable

We employed the Z-score as TKI’s distance from insolvency. The popularity of the Z-score stems from the fact that it does not necessitate strong assumptions to overcome the issue of the return on assets (ROA) distribution.

This measure has been used in related studies as a proxy of financial stability in financial intermediaries (Moreno et al. 2022; Karkowska and Acedański 2020). The Z-score reflects the number of standard deviations that a company’s ROA has to fall by to deplete its equity. As suggested by Altuntas and Rauch (2017), to calculate the ROA we used the measure of dispersion (σ(ROA)) over the past 5 years. It is a popular indicator of the probability of insolvency. A greater Z-score means that the likelihood of insurance insolvency risk is low, and hence greater financial stability. After this, the natural logarithm of the Z-score is necessary to control for the skewness exhibited by the original variables (Moreno et al. 2022).

3.2.2. Fraud Disclosure Index

To measure the level of FR_DISC across TKI organizations, a content analysis approach was used to quantify qualitative information by sorting data and comparing different pieces of fraud information from different financial reports. Another aspect of the data was assigned a descriptive label that encompassed significant types of fraud schemes for TKI and focused on six types of fraud, including internal fraud, external fraud, investment fraud, insurance broker fraud, fraud by service providers, and specific fraud information for TKI.

To identify the disclosure contents, we first downloaded the annual report of each company in our sample for the period of 2014 to 2022. Although previous research has investigated the binary categories that were used to code for the existence or absence of the identified items (word counts) in each sub-category (Hemrit and Belgacem 2024). We adopted a more nuanced scoring system to assess disclosure quality, such as a 0–3 scale, to rate disclosures based on depth and detail. Therefore, we assigned a “0” when there was no mention, “1” for a passing mention, “2” for a more detailed discussion, and “3” for an in-depth analysis or a comprehensive description of fraud risk disclosure practices. Hemrit (2020) argued that this method can easily pull out single keywords and insights from textual content, thus allowing us to accurately extract meaningful information and hidden messages. The overall index outlined 44 items from conventional and Islamic perspectives and grouped them into 6 fraud sub-categories (see

Table A1. Fraud Disclosure Index.

Internal Fraud	Real estate investment scams
Embezzlement of cash	Online investment fraud
Fraudulent payments	Promissory Notes
Skimming	Ponzi bubble
Theft	Insurance Broker Fraud
Invoice fraud	Underwriting Discrepancies
Payroll fraud	Settlement check distribution
Forgery of declarations	Unlicensed Agents
Fake policies and certificates	Churning
Dividend check distribution	Forgery
Misappropriation of assets	Fictitious policies
Premium fraud	Lapping
Fraudulent financial reporting	Fraud by Service Providers
Fabricated recipient	Adjusters Fraud
External Fraud	Underwriters Fraud
“Padding” (inflating claims)	Overbilling or double-billing
Eligibility fraud	Unbundling
Application fraud	Up-coding
Rate evasion	Dishonesty and misconduct
Falsified beneficiary claims	Specific Fraud Information for TKI
Cybercrime	Fraudulent philanthropy
Phishing and spoofing	Legitimacy risk
Opportunity Fraud	Shariah non-compliance risk
Credit card fraud	Coercion or deceit
Investment Fraud
Financial investment scams

Source: Hemrit and Belgacem (2024).

in Appendix A). The selection of items included in the index was guided by recommendations provided in some disclosure rules and reports. Thus, we addressed some aspects unique to International Financial Reporting Standards, IFRS 9 and 17, Pillar 3 reporting in Solvency 2, Islamic Financial Services Board (IFSB) guidelines, Accounting and Auditing Organization for Islamic Financial Institutions (AAOIFI) standards, and anti-fraud regulation in Saudi Arabia.

Subsequently, we manually confirmed that the content identified by each keyword search was a valid FR_DISC score. This step was compulsory as the content obtained from a keyword search may represent something that is entirely different from a fraud issue. For example, a keyword search on “skimming” may pick up a “rapid reading” sense instead of “fraud logic”. To enhance the validity and reliability of our content analysis, we followed the content analysis steps suggested by Botosan (1997). First, the disclosure index scores were considered valid if they measured the extent to which the disclosure of fraud represented the variable it was intended to. Here, we expect that, for each year, these scores will be positively correlated with the number of articles on “Tadawul financial group” and “SAMA” with the insurer company’s name in the headline. Our analysis demonstrates that all coefficients of correlation are significantly positive on a 1% level or better. Second, the reliability means that the scale is trustworthy and produces credible outcomes independent of the circumstances of its implementation. In the present study, FR_DISC produced a Cronbach’s coefficient alpha of 0.79, indicating that consistency between the scores of six types of fraud disclosures is high when compared with the generally acceptable measure in social science (70%) (Deumes and Knechel 2008).

3.3. Econometric Specification for Dynamic Panel Data Estimation

3.3.1. Model

Our study uses the dynamic panel data GMM estimators for testing the theoretical hypotheses. The model specification is based on the equation mentioned below, where y is the dependent variable for insurer i in the year t (ln (Z-Score i,t)), Xi,t represents a vector of variables affecting y, μi is the unobservable TKI-specific effect, and εi,t is the error term.

y_i.t = α₁y_i,t−1 + βX_i,t + μ_i + ε_i,t

(1)

y_i,t = α₁ y_i,t−1+ β₁ FR_DISC_i,t + β₂ OWC_i,t + β₃ BSZ_i,t + β₄ PNEXM_i,t + β₅ BMFR_i,t + β₆ACSZ_i,t
+ β₇ SBSZ_i,t + β₈ FR_DISC_i,t * SBSZ_i,t + β₉ INSZ_i,t + β₁₀ RCD_i,t + μ_i,t + ε_i,t

(2)

Existing research on signaling theory and disclosure suggests that corporate financial stability and performance could affect FR_DISC, while our research contends that FR_DISC has the potential to impact the financial stability of TKI. Regarding the bilateral relationship between these two variables, endogeneity may arise, precluding the consisting estimation of the coefficients, βi, i.e., the dependent variable, the financial stability of an insurance company, may depend on its past realizations. Theoretically speaking, there is a build-up of stability or instability of financial institutions from one year to the other (Hemrit 2020; Fiordelisi and Mare 2014). Thus, the use of lagged ln (Z-Score) as an independent variable becomes necessary, which also offers the advantage of controlling for observed and unobserved factors of the current level of TKI soundness. In a similar vein, the assumption of strict exogeneity is necessarily violated and the explanatory variables in our regression model are correlated with past values of the error term (Roberts and Whited 2013), i.e., hidden fraud information, and acts in various annual reports could be a function of its initial level and might persist in a company due to the lag in the actual implementation of new risk reporting reforms (Elfeky 2017). Therefore, the dynamic panel model is applicable when dynamic endogeneity may be present and accounts for unobservable, time-invariant, and TKI effects that may impact the current FR_DISC level. To resolve this problem, a difference generalized method of moments (D-GMM) was proposed by Arellano and Bond (1991) to eliminate the individual fixed effects by using the lagged variable as the instrumental variable (IV). Li et al. (2021) stated that using the lagged variable as an instrument eliminates the need for external instruments and can satisfy the relevance and exogeneity conditions.

In the presence of weak instruments, Blundell and Bond (1998) suggested using the system GMM estimator that uses lagged variables as instruments in both the difference and the level equations involving a system of two equations. This study therefore relies on the system GMM estimation method for the dynamic panel data analysis. Given that our analysis involves data of a small number of TKI organizations in a short period, we preferred the findings of the two-step system GMM over the one-step system GMM.

3.3.2. Dynamic Panel Data Analysis Assumptions

The dynamic panel data analysis requires that certain underlying assumptions be satisfied (Hall and Mairesse 2002). It is important to satisfy certain underlying assumptions (such as the stationarity, serial correlation, and homogeneity of variances (homoscedasticity)). For the stationarity assumption, the Levin–Lin–Chu bias-adjusted t-statistic is −7.1226 (with p < 0.05), which is significant at all of the usual testing levels. Therefore, we rejected the null hypothesis and concluded that the series is stationary. Furthermore, we performed a Wooldridge test for autocorrelation and observed that there is no significant first-order autocorrelation in our model (p > 0.05), thereby making the use of system GMM relevant. Finally, a Wald test and Breusch and Pagan Lagrangian multiplier tests were designed to test whether our data violate the assumption of homoscedasticity. Our results suggest that the fixed/random effects models could be heteroskedastic (p < 0.05), which again demonstrate that the system GMM estimators are much more reliable (Blundell and Bond 1998). To deal with any patterns of heteroskedasticity, we used the robust standard errors along with two-step estimators in this study.

4. Empirical Results

4.1. Descriptive Statistics

Table 2. Descriptive statistics.

	M	Min	Max	SD	SD (BG)	SD (WG)	JB
FR_DISC	15.000	1.000	34.000	8.122	0.548	0.214	8.096 **
OWC	1.974	0.000	6.000	1.361	1.453	0.290	14.743 ***
BSZ	2.064	1.609	2.484	0.210	0.186	0.074	6.230 **
PNEXM	0.483	0.285	0.800	0.104	0.547	0.063	39.093 ***
BMFR	1.672	1.098	2.639	0.294	0.365	0.077	19.636 ***
ACSZ	1.221	0.693	2.079	0.209	0.125	0.080	56.398 ***
SBSZ	1.277	0.000	3.000	1.463	1.444	0.024	38.021 ***
INSZ	20.995	17.541	26.531	0.981	0.684	0.435	251.780 ***
RCD	0.482	0.281	0.710	0.117	1.129	0.019	15.847 ***

Note. M: mean; SD: standard deviation; BG: between-group; WG: within-group; JB: Jarque–Bera test; FR_DISC: fraud disclosure index; OWC: ownership concentration; BSZ: board size; PNEXM: proportion of non-executive independent members; BMFR: board meeting frequency; ACSZ: audit committee size; SBSZ: Shariah board committee size; INSZ: size; and RCD: reinsurance ceded. (**) and (***) indicate significance at the 5 and 1% levels, respectively. Source: authors’ own creation.

summarizes the characteristics of our study variables of interest. We can observe that there was considerable variation over time and across the TKIs in our main sample. Additionally, the between-group (BG) variation in the data was found to be considerably larger than the within-group (WG) variation. In addition, we show that the between-group variance is substantially greater than the within-group variance, thereby highlighting the variations in the extents of FR_DISC between companies and exploring the financial stability at the corporate governance level. Finally, the results of the Jarque–Bera (JB) statistic show that all of the data series obey a Gaussian distribution.

Pearson correlations in

Table 3. Correlation matrix.

	FR_DISC	OWC	BSZ	PNEXM	BMFR	ACSZ	SBSZ	INSZ	RCD
FR_DISC	1
OWC	0.1265	1
BSZ	−0.0413	−0.1645	1
PNEXM	0.0609	−0.0046	−0.4251	1
BMFR	−0.1142	0.0707	0.1330	−0.0706	1
ACSZ	0.2125	0.0360	0.0807	−0.0929	0.2797	1
SBSZ	0.0920	−0.3087	0.2455	0.2997	0.0250	0.1114	1
INSZ	0.1740	−0.0544	0.1886	0.0003	0.2830	0.2134	0.2537	1
RCD	−0.1621	−0.0382	0.1039	−0.0857	0.1349	−0.0351	−0.0895	−0.0346	1
VIF	1.18	1.17	1.56	1.58	1.26	1.20	1.58	1.24	1.07

Source: authors’ own creation.

show that SBSZ is distinctly correlated with some explanatory variables. For instance, it is moderately and inversely correlated with ownership concentration (−0.3087) and positively correlated with the proportion of non-executive independent members (0.2997) as well as the size of the board risk committee (0.2455). Furthermore, BSZ is negatively and significantly correlated with the PNEXM, while Shariah board size is positively and significantly correlated with the size of the company. Taken together, our analyses confirm that the concern of multicollinearity does not primarily impact our findings.

Furthermore, multicollinearity tests were carried out on all explanatory variables by using the variance inflation factor, VIF. We found that all of the values were below the threshold of 4, which was within the traditional acceptable range of less than 10, proving that our estimation is not subject to the multicollinearity problem (Gujarati 1995).

In

Table 4. Results of hypothesis testing.

Dependent Variable: Z-Score
Variable	Dynamic Panel-Data Estimation, One-Step System GMM		Dynamic Panel-Data Estimation, Two-Step System GMM
Z-SCORE (lagged)	−0.1428 ***	−0.1981 ***	−0.2378 ***	−0.1554 ***
FR_DISC	−0.0302 **	−0.001	−0.0509 ***	−0.0367 *
OWC	−0.2142 ***	−0.1847 **	−0.4404 ***	−0.4182 ***
BSZ	−0.2153	−0.0987	2.6333 ***	2.6016 ***
PNEXM	−3.6501 ***	−3.6155 ***	−3.7891 ***	−3.7012 **
BMFR	0.9867 *	−0.5222	−0.6731	−0.5419
ACSZ	0.0010	−0.0259	0.0010	0.0008
SBSZ	0.2318	0.2258	−0.4554 *	−0.4312 *
SBSZ*FR_DISC	-	−0.0561	-	−0.1575 **
INSZ	0.5297 ***	0.5295 ***	0.5528 ***	0.5527 ***
RCD	1.9458	1.9052	2.7309 *	2.1263
Goodness-of-Fit
AR(1) p-value	0.0101	0.0095	0.0105	0.0103
AR(2) p-value	0.0641	0.0649	0.0545	0.0551
Hansen p-value	0.0629	0.0892	0.0627	0.0890
Observations	234	234	234	234
Number of TKIs	26	26	26	26

Note. Unstandardized regression coefficients were reported. FR_DISC: fraud disclosure index; OWC: ownership concentration; BSZ: board size; PNEXM: proportion of non-executive independent members; BMFR: board meeting frequency; ACSZ: audit committee size; SBSZ: Shariah board committee size; INSZ: size; RCD: reinsurance ceded; and SBSZ*FR_DISC referred to the interaction term. (*), (**), and (***) indicate significance at the 10, 5, and 1% levels, respectively. Source: authors’ own creation.

, we report the results of our baseline equation using both one-step and two-step system GMM methods established by Arellano and Bond (1991) and Blundell and Bond (1998) for dynamic panel data models. The values of the goodness-of-fit parameters suggested that the model is assumed to be correctly specified.

Let us start with the first-order auto-correlation in the residuals, using the AR(1) test. The null hypothesis of no first-order serial correlation was rejected in both the one-step and two-step system GMM methods. However, the AR(2) test confirmed that error terms are not serially correlated at levels, so the null hypothesis cannot be rejected (with p > 0.05). Furthermore, the Hansen tests confirmed the validity of the instruments we used to avoid the endogeneity problem (with p > 0.05). Although the one-step system GMM and two-step system GMM methods give very similar results, the two-step system GMM seems to be more appropriate. However, the similarity of the findings of both one-step and two-step system GMM methods ensured the robustness of our results (Windmeijer 2005).

From the two-step system GMM estimation, the higher coefficients of the lagged dependent variable (Z-SCOREt-1) approved the dynamic character of the specification of our model, indicating strong persistence, i.e., the slower the speed will be. As expected, the regression coefficients indicated a negative relationship between FR_DISC and the dependent variable, Z-SCORE, and H1 is then rejected. Thus, we found empirical support for the “Proprietary costs theory”: an insurer might obtain a higher level of financial stability if disclosure is able to provide benefits greater than the costs incurred. If not, FR_DISC can backfire, as it interferes with the ability of the market to anchor the financial soundness. As suggested by Goldstein and Yang (2019), information or concern about possible fraud does not always result in positive investor reactions, as long as there is no precise aggregation from different sources. In addition, FR_DISC might crowd out other risk information, possibly even attenuating the positive direct effect of disclosure on financial stability. This finding corroborates with some previous studies (Dai et al. 2024; Gago-Rodríguez et al. 2020; Prencipe 2004). In addition, we can still conclude that higher ownership concentration leads to financial instability of insurance companies. Our finding does not support the agency theory. This negative effect might support the argument of Rubio-Misas (2020), who stated that in widely held corporations the owner and manager incentives are not aligned. Yet this setting may influence stability negatively. In Islamic entities, this effect becomes more pronounced due to another misalignment of interest between principals (policyholders) and operators who act on behalf of major shareholders (Sallemi and Zouari 2024; Grassa et al. 2020; Altuntas and Rauch 2017). The coefficient of board size is positive and significant. It reveals that a larger board size will positively enhance the stability of TKI. It is associated with the findings of Adams and Kastrinaki (2023), as well as Hemrit (2020). In fact, with regard to the effect of board size on TKI stability, we found that the coefficient of this variable is significantly positive, supporting the resource dependence view. The underlying argument is that when board size increases, the marginal benefits from a wide range of expertise and skills outweigh the marginal costs, leading to lower variability in insurance performance and increased stability. Regarding the PNEXM, we found a result that is in line with Asare et al. 2023 and Nguyen et al. 2022, and it does not support H4. It may seem contrary to intuition; our result is consistent with traditional value maximization. Well-governed financial institutions may have tried to improve their stability by reducing information asymmetry faced by independent directors and the additional costs of monitoring them performing their functions. As noted by Karkowska and Acedański (2020), a high proportion of independent members has constrained rather than encouraged risk-taking to the point where financial stability is inhibited.

As for the topic of the Shariah compliance variable, we found that a larger SBSZ reduces TKI stability, supporting the agency theory, which views a high number of Shariah scholars as increasing agency costs and information asymmetry (Jensen 1993). However, this contradicts the resource dependence theory argument that a large SBSZ is more likely to arrange resources for a corporation. Furthermore, as a matter of principle, unanimous consent in decision making and all matters relating to compliance with Shariah are easier with a small SBSZ. Additionally, it can be expected to contain wisdom and soundness of judgment (Greuning and Iqbal 2008). This result is consistent with Khalil and Taktak (2020). The moderating effect of Shariah board size on the relationship between the disclosure extent of fraud and the financial stability of TKI was significant and statistically negative (β = −0.1575; p < 0.05), thus confirming H7. SBSZ can influence the relationship between FR_DISC and TKI stability through its effective oversight of Shariah principles. The significance of the interaction term between SBSZ and FR-DISC implies that the effectiveness of fraud disclosure in driving financial stability is negatively contingent on the SBSZ. When both SBSZ and fraud disclosure are favorable, their combined effect leads to decreased financial stability. Our results are in good agreement with the work of Ajili and Bouri (2018), who judged that the combination of functions increases to a great extent the performance of a firm because a Shariah board engages in complying with FR_DISC requirements in order to contribute to social justice and protect the interests of all stakeholders. In fact, the pressure from a high number of Shariah board members to disclose more fraud information can dominate the positive direct effect of providing new information so that better FR_DISC can harm stability, thus confirming the legitimacy viewpoint. Given this critical role of Shariah boards in TKI, the disclosure of fraud “leveled the playing field” by reducing informed traders’ information benefit.

4.2. Index of Fraud Disclosure and Insurance Categorisation

We performed another sensitivity analysis to verify whether the alternative indicator for FR_DISC measure and the classification of sample TKI organizations have any effect on financial stability. We then divided our primary sample into two sub-samples. The first includes the TKI years with a score of disclosure more than the median (FR_DISCscore > median). Here, the index of disclosure takes the value 0 and the insurer is considered as disclosing TKI. In contrast, the second consists of the TKI years with a score of disclosure equal to or less than the median, and the insurer is classified as non-disclosing TKI, following Botosan (1997). Our regression is then run for the two sub-samples based on the index of FR_DISC, which is I-FRD.

Table 5. Estimation results.

Dependent Variable: Z-Score
	Disclosing TKI Organizations (FR_DISCscore > Median)		Non-Disclosing TKI Organizations (FR_DISCscore ≤ Median)
Variable	Dynamic Panel Data Estimation, One-Step System GMM		Dynamic Panel Data Estimation, Two-Step System GMM
Z-SCORE (lagged)	−0.4887 ***	−0.5101 ***	−0.5482 ***	−0.5142 ***
I-FRD	−0.1175 ***	−0.1458 ***	−0.0192	−0.0278
OWC	−0.1507 ***	−0.1412 ***	−0.0418 *	−0.1010 *
BSZ	−0.6644	−0.9912	0.0455 **	0.0398 **
PNEXM	−1.5653 ***	−1.4736 ***	−1.5079 ***	−1.3903 ***
BMFR	0.4336	0.1426	−0.3648 *	−0.2977 *
ACSZ	0.5209 *	0.2449	0.3972	0.4071
SBSZ	0.0063	0.01555	0.0075	0.0031
SBSZ*FR_DISC	-	−0.3880 ***	-	−0.0008
INSZ	−1.0333	−1.1141	−1.1452	−1.2576
RCD	2.6918	1.9999	1.9824	2.0417
Goodness-of-Fit
AR(1) p-value	0.0155	0.0153	0.0169	0.0173
AR(2) p-value	0.0989	0.0857	0.0613	0.0689
Hansen p-value	0.0717	0.0711	0.0725	0.0805
Observations	91	91	143	143

Note. Unstandardized regression coefficients were reported. FR_DISC: fraud disclosure index; OWC: ownership concentration; BSZ: board size; PNEXM: proportion of non-executive independent members; BMFR: board meeting frequency; ACSZ: audit committee size; SBSZ: Shariah board committee size; INSZ: size; RCD: reinsurance ceded; and SBSZ*FR_DISC referred to the interaction term. (*), (**), and (***) indicate significance at the 10, 5, and 1% levels, respectively. Source: authors’ own creation.

exhibits the results.

As shown in Table 5, all of the coefficients on the variables have the expected size, show similar significance, and are consistent with our main findings. This suggests that the results are robust to the alternative indicators. Meanwhile, it can be noticed that the effects of the governance mechanisms and FR_DISC index of disclosing TKI organizations are not very different from those of non-disclosing ones. Our efforts reveal that the lagged dependent variable is negative and statistically significant at the 1% significance level, which indicated that the stability of TKI in a particular year largely depended on the level of its stability of the previous year for either disclosing or non-disclosing TKI. When controlling the influence of FR_DISC, it was found that the difference was prominent. The disclosing TKI organizations are more likely to achieve financial stability regarding their strategy for providing less information about fraud. Then, despite the fact that FR_DISC could discipline TKI organizations and provide incentives to take corrective actions early (Hemrit and Belgacem 2024), disclosing TKI organizations might have found themselves caught in conflicting duties of confidentiality and disclosure of fraud. However, we found evidence that FR_DISC level mattered less for non-disclosing TKI organizations.

Regarding corporate governance mechanisms, we found almost no differences related to some attributes when we analyzed disclosing and non-disclosing TKI organizations separately. The ownership concentration and the proportion of non-executive independent members are negative and highly significant at the 1 percent level for disclosing TKI organizations. More precisely, the higher ownership concentration strengthens block shareholding, which has a negative impact on the financial stability of TKI. We also found a significant negative relationship between the proportion of non-executive independent members and financial stability. This negative effect may be justified in view of the contention that a high number of independent members may reduce a CEO’s willingness to share information about fraud with others, causing high levels of uncertainty, which lead to a reduction in TKI financial stability (Berger et al. 2016). The results align with the view that a high proportion is an inefficient allocation of time and resources. This finding is in contrast to the prediction of agency theory. For non-disclosing TKI organizations, the results are in line with the view that board size positively affects financial stability. This is consistent with Adams and Kastrinaki (2023) and Ali (2018). However, it is not significant for the disclosing counterparts. This is, indeed, surprising, because according to Hemrit (2020), better Islamic governance is often associated with larger board size in a risk-disclosing company. In relation to BMFR, we find that this extra governance mechanism tends to significantly reduce the financial stability of non-disclosing companies. The negative coefficient of this variable here is predictable because of the often-expressed concern that FR_DISC could have adverse consequences and trigger panics. Thus, holding regular meetings can equate to low returns on assets and equity and lead to the presence of endogeneity problems in relation to financial stability (Vafeas 1999). Additionally, audit committee board size is also positive, but weakly significant at the 10 percent level for disclosing TKI organizations. However, Shariah board size is positive but insignificant in the financial stability of TKI. As stated earlier, our result indicates a significant moderation effect of Shariah board size on the association between FR_DISC and financial stability in the disclosing TKI organizations. This result further confirms the important role played by the Shariah committee in making a more coherent strategy regarding FR_DISC decisions, i.e., there was a negative association between FR_DISC and financial stability when the number of Shariah board members was high, and the association was somewhat positive when the number was low. The Shariah board function is responsible for reporting Shariah non-compliance risk exposures (such as fraud acts) to stakeholders to demonstrate their adherence to these Islamic principles (Hemrit and Belgacem 2024). Accordingly, compliance with the transparency principle is imperative, and a very natural unintended consequence to consider is that the disclosure of fraud, which is approved by a large Shariah board committee, is relevant, reliable, and not subject to some form of privilege. As a last resort, this bad information may cause TKI organizations to become more fragile, and investors may withdraw their funds (Sallemi and Zouari 2024). Finally, the results of Table 5 display no significant indications that the insurance size and reinsurance ceded variables have any influence on the financial stability of TKI organizations.

4.3. Robustness Tests

To shed some light on the detected weak evidence effect of corporate governance measures (BSZ, BMFR, and ACSZ) and control variables (INSZ and RCD) on the Z-score of TKI, we explore the association between one component of the Z-score measures (e.g., capital asset ratio (CAR)) (Berger and Mester 1997) and the explanatory variables. The CAR is computed by dividing a TKI organization’s capital over its risk-weighted assets. After this, we try to compare the dynamic panel model main results with the models most frequently employed in previous studies, specifically ordinary least squares (OLS) and the fixed effect model (FEM). The results are summarized (see

Table A2. Results of the estimation of the capital asset ratio (CAR).

	CAR
Variable	Dynamic Panel Data Estimation, Two-Step System GMM
Y (lagged)	−0.9400 ***
FR_DISC	−0.3816 ***
OWC	−0.1412 ***
BSZ	0.0236
PNEXM	−1.4736 ***
BMFR	−0.9693 ***
ACSZ	0.0417
SBSZ	−0.4719 *
SBSZ*FR_DISC	−0.4760 ***
INSZ	0.0162
RCD	−0.8985 **
Goodness-of-Fit
AR(1) p-value	0.0270
AR(2) p-value	0.1017
Hansen p-value	0.0898
Observations	234

Note: *, **, and *** indicates significance at the 10%, 5%, and 1% levels. Source: authors’ own creation.

in Appendix B).

The results of our model lead to identical conclusions for the CAR measure. Our findings regarding FR_DISC and the moderating effect of Shariah board size are unchanged. Additionally, with regard to the relationship between corporate governance tools and financial stability, we must indicate that the coefficient of board meeting frequency becomes significant at the 1% level. In addition, the insignificant coefficient of audit committee size has changed from a negative sign (under the proxy of the Z-score) to be positive (under the CAR). Furthermore, TKI organizations characterized by high retention levels are likely to exhibit high financial stability. These differences can be attributed to the difference in the composition of the proxies, given that the Z-score is computed on the basis of some income statement and balance sheet items, whilst the CAR measure focuses only on items extracted from the balance sheet (Marie et al. 2021).

Subsequently, we run the OLS and FEM regressions for the Z-score measure (see

Table A3. Results of the estimation of Z-SCORE by using OLS and FEM methods.

	OLS	FEM
Variable
Constant	7.7807 ***	9.4519 ***
FR_DISC	−0.1669 **	−0.3664 ***
OWC	−0.0235	−0.4256 ***
BSZ	0.0979 ***	0.1781 ***
PNEXM	−1.3283 ***	−0.3009
BMFR	0.0947	0.0926
ACSZ	1.3214 ***	−0.0038
SBSZ	−0.2682 **	−0.3394 ***
SBSZ*FR_DISC	−0.0848 *	−0.2524 ***
INSZ	−0.2805 ***	−0.3365
RCD	−0.0008	−0.3365 ***
Adj R2	0.2172	0.4911
Fstat	-	7.4247 ***
Obs	234	234

Note: Regression carried out using OLS and FEM. ***, **, and * demonstrate statistical significance at the 1%, 5%, and 10% levels, respectively. Source: authors’ own creation.

in Appendix B). The significance of FR_DISC and the mediating effect of the Shariah board size impact have further been established. This confirms what we observed above. For the control variables, under the FEM estimation results, we found that whereas the effect of the reinsurance ceded level on TKI stability was negative, the effect of the insurance size was not significant. Our new findings on the role of corporate governance attributes are also unaltered. Overall, our previous findings with this new specification remain consistent.

5. Discussion

5.1. Theoretical Contributions

Our study adds considerably to the current literature on corporate fraud disclosure in the insurance sector. The potential contributions of this paper are mainly in the following three aspects: First, the outcome broadens our knowledge of the moderator role of the Shariah board committee size on the relationship between FR_DISC and TKI stability. Thus, the significance of the interaction term between SBSZ and FR-DISC shows that the effectiveness of fraud disclosure in promoting financial stability is inversely correlated with the size of a Sharia board committee. In the Islamic governance context, it seems that legitimacy theory may lose some of its predictive ability because investors may see fraud disclosure as a given and not as legitimization too. For this reason, unlike conventional insurance, which is profit-driven, the voluntary disclosure in an Islamic entity is a priority to fulfilling the essentials (transparency), even though the fraud disclosure processing costs may exceed the targeted level of financial stability, which is consistent with the finding of Grassa et al. (2020).

Second, even though there are several studies on corporate fraud and financial stability (e.g., Dai et al. 2024; Gago-Rodríguez et al. 2020; Prencipe 2004), most research was conducted in the setting of developed countries such as the US and UK, and the conclusions of this study may not apply to other economies with different contextual circumstances, like Saudi Arabia. In such a case, the clarity of regulatory requirements related to the disclosure of information in Saudi Arabia is not at a sufficient organizational level to ensure insurance system stability (SAMA 2023). Such regulation negatively shaped stakeholder expectations about information on assumed fraud collateral effects. Therefore, understanding accounting practices in the disclosure process requires more than a technical description. The empirical usefulness of positive accounting theory would be called into question in the context of the insurance sector in Saudi Arabia.

Third, for revealing TKI organizations, the ownership concentration and the proportion of non-executive independent members are negative and significant at the 1 percent level. More specifically, a larger concentration of ownership reinforces block shareholding, which is detrimental to TKI financial stability. The proportion of non-executive independent members and financial stability was also found to be significantly negatively associated. In this regard, this result contradicts the agency theory, which assumes that having high numbers of independent members can ensure the most effective functioning of a firm, leading to high stability. Our results confirm the reported results of previous studies in the field (Karkowska and Acedański 2020; Masulis 2020).

5.2. Practical Contributions

Our results could be of potential benefit to policymakers, investors, and TKI insurers, as well as informing future research on the stability of Takaful operators. Then, investigating FR_DISC and corporate governance determinants that affect the likelihood of TKI stability in this alternative and fast-growing Islamic insurance industry provides strong evidence on the importance of distinguishing the informativeness of FR_DISC when studying the impact on the financial stability. For policymakers, to be effective in promoting financial stability in the Takaful insurance sector, the disclosure of fraud information in TKI must be complemented by strong incentives for counterparties to engage in management and monitoring processes. This is becoming more evident due to the pressures of new initiatives issued by regulators and the need to maintain transparency compliance requirements. Furthermore, our study advises the investors of TKI organizations disclosing fraud to not only make more informed investment decisions, but also assess the identified fraud risks alongside other relevant factors set out to jointly influence other potential investors’ stock valuation judgments. Fraud reporting research (Hemrit and Belgacem 2024; Hoberg and Lewis 2017) suggests that having knowledge of fraud information disclosure plans can improve the risk perception level of investors (Bach et al. 2023). In other words, FR_DISC is more subdued with a negative tone, and thus investors would have to rebuff behavioral biases to follow through and explore some of the best practices for dealing with this kind of information. Finally, this paper’s findings could be useful for TKI. Specifically, due to increased decision-making frictions, free-riding, and coordination issues, TKI organizations need to continually evaluate how many Shariah board members and non-executive independent members they truly need. In addition, upon fraud materialization and in the absence of FR_DISC, TKI organizations should perhaps diffuse fraud information with a more cautionary tone to maintain a strong standing with regulatory bodies and avoid negative impacts being generated from this information. Undoubtedly, this will have a positive effect on the financially stable foundation for Takaful operations.

6. Conclusions

In the corporate world, the empirical evidence on the usefulness of FR_DISC strategies, particularly in relation to financial stability, is scarce. Thus, this study extends the fraud reporting and Islamic corporate governance literature by examining the relationship between fraud disclosure and the financial stability of global TKI organizations, an area that has received limited attention. Furthermore, we examine the extent to which corporate governance attributes influence TKI stability. This study also tested the moderation effects of audit committee members’ multiple directorships on the above association. To test these relationships, a dynamic panel model was performed on selected TKI companies, allowing us to uncover an unexpected result: subjects equalize behavior across repetitions.

6.1. Summary of Research

Using a comprehensive sample of 26 Takaful insurers in Saudi Arabia for the period of 2014 to 2022, our results indicate that TKI organizations with lower fraud disclosure levels are more likely to have higher financial stability, providing new insights about the equilibrium between agency costs and benefits of nonfinancial risk disclosure. This finding is in contrast to the prediction of agency theory and the common belief that the disclosure of any information is an effective means with which to manage conflicts of interest, which in practice often involves ensuring financial stability. Additionally, our findings provide support for our theoretical hypothesis of the study, which posits that Shariah board committee size moderates the relationship between FR_DISC and TKI stability. Moreover, a deeper investigation suggests that ownership concentration, the presence of non-executive independent members, and Shariah Board committee size negatively affect this stability. However, the board size of an insurance company seems to be important in boosting its stability.

6.2. Limitations

Notwithstanding its contribution, this study is not immune to some limitations. Primarily, according to the trend data, the desire to move on should not deny the effect that COVID-19 still has on the financial stability of the world. However, such effects that might have appeared within the time range were excluded. Secondly, the index of FR_DISC is derived exclusively from an inspection of annual reports, potentially yielding disparate conclusions. Hence, some other reports might capture the entirety of fraud information in TKI organizations. Finally, the restricted number of insurance companies within the Saudi field might have an impact on the generalizability of our findings.

6.3. Future Research Directions

We close with a caveat that the empirical results of our study should not be taken too seriously unless confirmed by future research. We suggest several examples here: First, within a Saudi context, an opportunity exists to discover issues regarding investors’ and policyholders’ perceived fraud or the information content of FR_DISC. Second, future research might examine the comparative effects of the mandatory and voluntary disclosure of fraud on TKI stability in the post- and pre-COVID-19 periods. Third, it is quite likely that some of our empirical results will differ in future studies on cross-country data that use different sets of assumptions.