Financial fraud continues to pose a significant threat to organisational sustainability in both emerging and developed economies. In Southeast Asia, particularly in Indonesia, Malaysia and Singapore, financial statement fraud (FSF), asset misappropriation, and corruption remain persistent challenges within the corporate sector. These issues not only erode investor confidence and corporate governance standards but also present systemic risks to national economies (Xu, Wenlong & Siyu 2025). Recent evidence continues to show that fraud remains a significant issue despite improvements in governance and internal control systems. According to the Association of Certified Fraud Examiners (ACFE) global report, occupational fraud persists worldwide, with organisations estimated to lose approximately 5% of their annual revenue to fraud, indicating that fraud risk remains pervasive across sectors (ACFE 2024). However, in terms of financial impact, FSF represents a serious threat to the stability of financial markets, necessitating effective detection mechanisms to prevent substantial investor losses and protect corporate reputations (Li, Yen & Wang 2024). This pattern is mirrored across the region. In Malaysia, financial fraud resulted in cumulative losses exceeding USD 3.6 billion annually between 2018 and 2021 (Dorros 2020). According to PricewaterhouseCoopers’ 2020 survey, 24% of Malaysian companies reported losses greater than USD 1 million, while the overall incidence of fraud increased to 43% in 2020 from 41% in 2018. Similarly, Singapore witnessed a sharp 64.5% increase in reported fraud cases in 2023 compared to the previous year, although the total financial losses declined slightly to SGD 334.5 million (Darisman & Ananda 2023).

Among industrial sectors, the basic materials sector is particularly susceptible to fraudulent activities. As a key supplier of raw materials to various industries, firms in this sector often manage high-value assets and large-scale financial flows (Hakim et al. 2024). These characteristics create greater opportunities and incentives for financial statement manipulation, especially under conditions of economic pressure or weak corporate governance (Rozi & Munari 2024). Several high-profile cases illustrate this vulnerability. PT Waskita Beton Precast, a major Indonesian construction company, was involved in a large-scale corruption scandal between 2016 and 2020, resulting in state losses of approximately IDR 2.5 trillion. PT Krakatau Steel reported substantial financial irregularities during the development of its Blast Furnace Complex, with losses exceeding IDR 6 trillion. Similarly, PT Timah Tbk – one of the world’s largest tin producers – faced legal proceedings for financial misstatements and operational inefficiencies, leading to significant erosion of shareholder value (Soda 2016). These cases underscore the critical need for robust fraud risk management, particularly in sectors characterised by volatile assets and liabilities that are prone to manipulation.

The dynamics underlying FSF have evolved beyond traditional explanatory models. Early frameworks, such as the Fraud Triangle proposed by Cressey (1953), identified pressure, opportunity and rationalisation as the primary drivers of fraudulent behaviour. Subsequently, the Fraud Diamond Model introduced by Wolfe and Hermanson (2004) incorporated capability as a critical fourth element. The Fraud Pentagon, developed by Marks (2012), further expanded the model by adding arrogance, emphasising the influence of executive hubris. Building on these earlier frameworks, the Fraud Hexagon Theory proposed by Vousinas (2019) introduces collusion as a sixth dimension, recognising that coordinated fraudulent schemes involving multiple actors pose greater risks and complexities. Together, the six elements of the Fraud Hexagon – pressure, opportunity, rationalisation, capability, arrogance and collusion – provide a more comprehensive framework for understanding and detecting FSF (Achmad, Ghozali & Pamungkas 2022; Honesty, Honesty & Setiawan 2024).

In addition to internal organisational factors, cultural context, particularly uncertainty avoidance, plays a significant role in shaping fraudulent behaviour. As conceptualised by Hofstede (2011), uncertainty avoidance reflects a society’s tolerance for ambiguity and its tendency to implement formal rules and structures to reduce uncertainty. In cultures characterised by high uncertainty avoidance, managers may adopt risk-averse strategies, including unethical practices such as financial manipulation, to maintain organisational stability and safeguard personal interests (Hasan et al. 2020; Khairani et al. 2024). Empirical evidence suggests that uncertainty avoidance can both suppress and exacerbate fraudulent behaviour, depending on the strength of institutional controls and individual rationalisation processes (Tsalavoutas & Tsoligkas 2021). In high uncertainty-avoidance environments, where mistakes and failures are heavily stigmatised, managers may perceive fraudulent financial reporting as a rational means of avoiding reputational damage, regulatory sanctions, or career setbacks (Hofstade 2020). Accordingly, examining the moderating effect of uncertainty avoidance within the Fraud Hexagon framework is essential to understanding the dynamics of FSF in Southeast Asia, a region marked by strong cultural collectivism and considerable regulatory heterogeneity.

Although numerous studies have analysed the Fraud Hexagon in isolation (Haeronnisa & Isnawati 2024; Nugroho & Diyanty 2022; Suryandari et al. 2023), few have incorporated cultural dimensions as moderating factors. Moreover, prior research has often focused on single-country contexts, thereby limiting the generalizability of findings across regions with differing legal systems, corporate governance structures and cultural environments. This study addresses these gaps by examining the combined effects of Fraud Hexagon elements on FSF, while also exploring the moderating role of uncertainty avoidance. To enhance cross-cultural relevance, the study employs empirical analysis across Indonesia, Malaysia and Singapore, three economically interconnected yet culturally diverse nations. In doing so, this research offers nuanced insights into how organisational and cultural factors jointly shape fraudulent financial reporting, providing practical implications for academics, regulators and corporate policymakers.

This study aims to empirically investigate the influence of the Fraud Hexagon elements (pressure, opportunity, rationalisation, capability, arrogance and collusion) on the incidence of FSF in basic materials companies listed on the Indonesia Stock Exchange (IDX), Bursa Malaysia and Singapore Exchange (SGX) between 2021 and 2023. Furthermore, it examines the moderating effect of cultural factors, specifically uncertainty avoidance, on the relationship between these fraud risk elements and fraudulent reporting practices. By incorporating cultural dimensions into the fraud risk framework, this research contributes to the refinement of fraud theory and the development of more culturally responsive models for detecting and mitigating FSF in the Southeast Asian context.

The study offers both theoretical and practical contributions. Theoretically, it extends existing literature by emphasising the significance of cultural context, particularly uncertainty avoidance, influencing fraudulent behaviour. Practically, the findings provide insights into improving fraud detection mechanisms, strengthening internal audit strategies, and guiding the formulation of regulatory and corporate governance policies. These insights are especially relevant for multinational corporations operating in culturally diverse environments. From a policymaking perspective, the study supports the design of more targeted anti-fraud regulations and governance reforms that address not only organisational vulnerabilities but also broader socio-cultural drivers of FSF.

This study investigates FSF among basic materials companies listed on the IDX, Bursa Malaysia, and the SGX over the period 2021 to 2023. Specifically, it examines the influence of six key factors: pressure, opportunity, rationalisation, capability, arrogance and collusion, on the occurrence of FSF, with uncertainty avoidance introduced as a moderating variable. Employing a causal research design and a quantitative approach, the study utilises secondary data extracted from annual financial reports available on the official websites of the respective stock exchanges. Cultural dimension data, particularly uncertainty avoidance, are sourced from Hofstede Insights. The documentation method is applied for data collection, enabling systematic analysis of financial and governance-related information. The study population comprises all basic materials companies listed on the IDX, Bursa Malaysia and SGX during the observation period. A purposive sampling technique is used, based on the following inclusion criteria: (1) the company must be classified under the basic materials sector; (2) it must have maintained financial solvency (i.e. not reported losses) throughout 2021 to 2023; and (3) complete data must be available for all variables under investigation

The study utilises a sample of 72 companies, selected based on predefined inclusion criteria. The unit of analysis comprises 216 firm-year observations from basic materials companies listed on the IDX, Bursa Malaysia and the SGX during the 2021 to 2023 period. All companies included in the sample meet the eligibility requirements specified for the study. The dependent variable, FSF, is measured using the F-Score method. Independent variables include pressure (measured by leverage ratio), opportunity (percentage of independent commissioners), rationalisation (frequency of accounting policy changes), capability (director changes), arrogance (number of CEO photographs in annual reports) and collusion (market performance measured by Price to Book Value). Uncertainty avoidance, serving as the moderating variable, is measured using national cultural indices from Hofstede Insights.

Data was collected through the documentation of annual reports and relevant secondary sources. The analysis employs panel data regression techniques, using E-Views 12 software for model estimation. The primary model examines the direct relationships between the independent and dependent variables. Additionally, a moderated regression analysis (MRA) is conducted to assess the moderating effect of uncertainty avoidance on the relationships between the predictors and FSF. The study applies three panel data regression models: the Common Effect Model (CEM), Fixed Effect Model (FEM) and Random Effect Model (REM). The selection of the most appropriate model is determined using Chow, Hausman and Lagrange Multiplier tests. To ensure the validity of the models, classical assumption tests are performed, with particular attention to multicollinearity diagnostics. Concerns regarding heteroskedasticity and autocorrelation are addressed by applying the Generalised Least Squares (GLS) method. For hypothesis testing, R² is used to assess model fit, t-tests evaluate the partial effects of independent variables, and the MRA is employed to examine interaction effects. The expected findings of this study are twofold: firstly, they will empirically test the theoretical frameworks surrounding fraud behaviour in the corporate sector; secondly, they will provide insights into how cultural factors, such as uncertainty avoidance, influence fraudulent practices within the corporate financial environment in Southeast Asia.

To determine the most appropriate estimation model for panel data, two diagnostic tests were conducted: the Hausman Test and the Lagrange Multiplier (LM) Test. The Hausman Test was employed to decide between the FEM and the REM. The null hypothesis (H₀) posits that the REM is appropriate, while the alternative hypothesis (H₁) supports the FEM. If the p-value is greater than 0.05, the REM is preferred. The Hausman test yielded a Chi-square probability of 1.0000, which is well above the 0.05 threshold. Consequently, H₀ is accepted, indicating that the REM is more appropriate than the FEM. The LM Test was subsequently applied to compare the REM with the CEM. The null hypothesis (H₀) favours the CEM, while the alternative hypothesis (H₁) supports the REM. A p-value below 0.05 indicates that the REM is more suitable. The test results revealed that the Breusch-Pagan statistic produced a p-value of 0.0000, leading to the rejection of H₀. Therefore, the REM is deemed superior to the CEM. Based on the outcomes of both tests, the REM was selected as the most appropriate model for panel data regression in this study, as it was found to be preferable over both the Fixed Effect and Common Effect models.

The LM test, developed by Breusch and Pagan (1980), is employed to determine whether the REM is more appropriate than the CEM in panel data analysis. This test assesses the significance of random effects by evaluating the residuals obtained from the CEM. The null hypothesis (H₀) posits that the CEM is adequate (i.e., random effects are not significant), while the alternative hypothesis (H₁) supports the use of REM because of the presence of significant random effects. If the LM statistic exceeds the critical value derived from the Chi-square distribution, H₀ is rejected in favour of H₁, indicating that REM provides a better fit. Based on the results presented in Table 1, the LM test yielded a Breusch-Pagan probability value of 0.0000, which is well below the 0.05 significance level. This finding confirms the presence of significant random effects. Accordingly, H₀ is rejected, and the REM is selected as the most appropriate and best-fitting model for the panel data regression employed in this study.

The t-test is commonly employed to assess the partial effect of each independent variable on the dependent variable (Hsiao 2022). A p-value less than 0.05 indicates statistical significance, leading to the acceptance of the hypothesis; conversely, a p-value greater than 0.05 suggests insufficient evidence to support the hypothesis. This study utilises panel data regression analysis, performed using EViews software, to examine the influence of selected independent variables on the dependent variable, profitability, which is proxied by FSF. The REM was identified as the most appropriate estimation technique based on model selection tests, and classical assumption tests confirm that the model does not violate key econometric assumptions (Okoro 2020). The results of the REM regression analysis are presented in Table 2.

Based on the panel data regression results, the constant value is 28.42093, indicating that when all independent variables – pressure, opportunity, rationalisation, capability, arrogance and collusion – are held at zero, FSF would have a baseline level of 28.42093. This suggests a predisposition for FSF to occur even in the absence of the examined predictors.

The regression coefficient for pressure is positive at 0.644842, implying that a 1% increase in pressure is associated with a 0.644842% increase in FSF, assuming all other variables remain constant. Likewise, opportunity shows a positive coefficient of 0.061677, indicating a 1% increase in opportunity leads to a 0.061677% rise in FSF. In contrast, rationalisation displays a negative coefficient of −0.006463, suggesting that a 1% increase in rationalisation would result in a 0.006463% decline in FSF. Capability has a positive coefficient of 0.171355, indicating that a 1% increase in capability is associated with a 0.171355% increase in FSF. Similarly, arrogance exhibits a positive coefficient of 0.019853, suggesting a marginal positive influence on FSF. However, collusion shows a negative coefficient of −0.003868, implying that a 1% increase in collusion is associated with a 0.003868% decrease in FSF, ceteris paribus.

Hypothesis testing reveals that the p-values for pressure (0.0389) and capability (0.0334) are below the conventional significance threshold of 0.05. These results indicate that both variables have a statistically significant influence on FSF, thereby supporting Hypotheses H1 and H4. On the other hand, opportunity (p = 0.8006), rationalisation (p = 0.9370), arrogance (p = 0.2199), and collusion (p = 0.8515) have p-values greater than 0.05, indicating no significant effect on FSF. As a result, Hypotheses H2, H3, H5 and H6 are rejected.

This study employed panel data regression using the MRA approach to examine whether uncertainty avoidance moderates the relationships between six independent variables (pressure, opportunity, rationalisation, capability, arrogance and collusion) and the dependent variable, FSF. This approach allows for the assessment of interaction effects between the cultural dimension of uncertainty avoidance and each fraud risk factor. The estimation results are summarised in Table 3.

Table 3 shows that among the six interaction terms, only two variables pressure (Press*UA) and collusion (Coll*UA) have statistically significant moderating effects, with p-values of 0.0257 and 0.0000, respectively. These findings suggest that uncertainty avoidance strengthens the relationship between these variables and FSF. More specifically, the interaction term between pressure and uncertainty avoidance (Press*UA) has a positive and significant coefficient (β = 0.098, p < 0.05), indicating that in environments with higher levels of uncertainty avoidance, the influence of pressure on FSF is amplified. This supports the notion that when individuals or organisations operate in cultures that are less tolerant of ambiguity, the stress induced by pressure becomes more likely to lead to fraudulent reporting behaviours.

Similarly, the interaction between collusion and uncertainty avoidance (Coll*UA) is positively significant (β = 0.010, p < 0.01), implying that high uncertainty avoidance strengthens the impact of collusion on FSF. This finding may reflect the cultural tendency to rely on trusted networks or informal agreements in uncertain environments, potentially enabling collusive practices to manipulate financial outcomes. Conversely, the interaction terms for opportunity (Press*UA), rationalisation (Ras*UA), capability (Cap*UA), and arrogance (Aro*UA) are not statistically significant (all p > 0.05). These results indicate that uncertainty avoidance does not significantly alter the effect of these variables on FSF. The absence of moderation could suggest that the impact of these factors on fraudulent behaviour operates independently of cultural attitudes toward uncertainty.

These findings contribute to the literature by demonstrating that cultural dimensions, particularly uncertainty avoidance, play a selective role in moderating fraud-related behaviours. The significant moderation effects for pressure and collusion are aligned with theoretical expectations from the Fraud Triangle and Diamond Theory, wherein psychological and social dynamics interact with cultural norms to influence unethical decision-making.

The hypothesis testing results indicate that pressure has a statistically significant positive effect on FSF, with a regression coefficient of 0.644842 and a p-value of 0.0389 (<0.05), thereby supporting H₁. This result reinforces the Fraud Hexagon Theory, which posits that pressure, particularly in the form of financial strain or ambitious performance targets, can drive individuals to manipulate financial reports as a coping mechanism. Theoretically, this finding is also supported by Agency Theory, which emphasises the conflict of interest between managers and shareholders. Under conditions of pressure, managers may engage in opportunistic behaviour to fulfil expectations or to avoid negative consequences. Logically, firms experiencing financial deterioration or heightened scrutiny from external stakeholders (e.g. lenders or investors) may inadvertently incentivise fraudulent behaviour as a means of signalling stability. These findings are consistent with previous research (Honesty et al. 2024; Hudayati et al. 2022), but contrast with the results of Handoko and Natasya (2019), and Khairani et al. (2024), which may reflect contextual differences in pressure intensity or regulatory enforcement environments.

In contrast, the results reveal that opportunity does not have a statistically significant influence on FSF (coefficient = 0.061677; p-value = 0.8006 > 0.05), leading to the rejection of H₂. This finding challenges the core proposition of the Fraud Hexagon Theory, which suggests that weak internal controls or ineffective oversight mechanisms create a conducive environment for fraudulent behaviour. Theoretically, the lack of significance may be attributed to the phenomenon of institutional isomorphism, wherein organisations adopt standardised governance structures (e.g. the presence of independent commissioners) that comply symbolically with regulatory expectations but may lack substantive enforcement. From a practical perspective, the mere presence of opportunity may be insufficient to induce fraud unless it coexists with strong motivational factors (e.g., pressure) and rationalisations. This outcome is consistent with studies by Koharudin and Januarti (2021) and Wicaksono and Suryandari (2021), which argue that opportunity alone does not guarantee fraudulent activity in the absence of complementary enablers. However, this result diverges from findings by Wibowo and Lastanti (2024) potentially because of differing operationalisations of the opportunity construct or variations in governance quality across sampled firms.

Rationalisation was found to have a statistically non-significant negative effect on FSF (coefficient = −0.006463; p-value = 0.9370 > 0.05), leading to the rejection of H₃. Theoretically, rationalisation is a key psychological mechanism that enables individuals to justify unethical actions, as proposed by Cognitive Dissonance Theory. This theory suggests that individuals seek to align their actions with internal beliefs or reduce psychological discomfort resulting from misconduct. However, in organisations characterised by strong ethical cultures and transparent governance structures, such rationalisation may be curtailed by institutional norms and external accountability. Logically, even when individuals attempt to legitimise fraudulent behaviour, effective internal controls and stakeholder oversight may suppress the translation of such justifications into actual misconduct. Moreover, the implementation of Good Corporate Governance (GCG) practices may reduce the cognitive leeway required for rationalisation, thereby minimising its influence on fraud decisions. This finding is in line with studies by Khairani et al. (2024), Yusrianti et al. (2020), and Wibowo and Lastanti (2024), but contrasts with the results of Pradipta and Bernawati (2019), which may stem from variations in the operationalisation of rationalisation or differences in ethical climates across firms.

In contrast, capability was found to have a statistically significant and positive effect on FSF (coefficient = 0.171355; p-value = 0.0334 < 0.05), thereby supporting H₄. This result is consistent with the Fraud Hexagon Theory, which posits that individuals in positions of authority who possess specialised knowledge and experience are more likely to initiate and conceal fraudulent acts. Logically, individuals with extended tenures or control over the financial reporting process not only have the technical means but also the organisational confidence to override internal controls undetected. Theoretically, this finding is aligned with the Fraud Diamond model, in which capability is regarded as a crucial enabler, often associated with traits such as dominance, intelligence, and strategic acumen. These insights underscore the paradox that while managerial competence is essential for performance, it may also facilitate manipulation in environments lacking robust checks and balances. This result supports the findings of Uciati and Mukhibad (2019), but contrasts with those of Wibowo and Lastanti (2024), Lyra et al. (2022), and others, likely as a result of differences in the proxies used for capability (e.g. tenure vs. educational background) or the regulatory and governance environments across industries and countries.

Arrogance was found to have no statistically significant effect on FSF (coefficient = 0.019853; p-value = 0.2199 > 0.05), leading to the rejection of H₅. According to the Fraud Hexagon Theory, arrogance, defined as an executive’s belief in being above rules or oversight, can foster fraudulent tendencies. However, in this study, arrogance was operationalised through the frequency of CEO photographs in annual reports, which may insufficiently reflect deeper behavioural traits such as narcissism or overconfidence. Theoretically, Upper Echelons Theory posits that executive personality traits, including egocentricity, influence organisational outcomes and ethical judgements. Nonetheless, reliance on symbolic or visual cues may limit the construct’s validity in capturing the psychological dimensions of arrogance. Logically, the presence of CEO images may reflect standardised corporate communication practices rather than manifestations of self-aggrandisation or rule-defying attitudes. Thus, the non-significant result aligns with findings by Khairani et al. (2024) and Situngkir and Triyanto (2020), while diverging from Uciati and Mukhibad (2019), who adopted more comprehensive behavioural indicators of executive arrogance.

Collusion, likewise, was found to have a non-significant effect on FSF (coefficient = −0.003868; p-value = 0.8515 > 0.05), resulting in the rejection of H₆. The Fraud Hexagon Theory conceptualises collusion as a critical enabler that reduces the probability of detection and increases fraud concealment. However, its actual influence may be contextually moderated by factors such as organisational transparency, the presence of whistleblower protections or the effectiveness of internal audit functions. From a collective action perspective, engaging in collusion involves shared risk, which may deter participation unless mutual trust, secrecy, and aligned incentives are strongly present. Logically, this implies that while relational networks exist within firms, the potential legal consequences and reputational damage of collusion may outweigh its perceived benefits, thereby limiting its occurrence. This finding is consistent with Lyra et al. (2022), suggesting that collusion may not always be a dominant determinant of fraud in governance-rich environments. However, the result contradicts studies by Khairani et al. (2024) and Wibowo and Lastanti (2024) possibly because of differences in measurement approaches or industry-specific dynamics influencing the feasibility of collusion.

The results of the MRA, presented in Section 4.6, indicate that Uncertainty Avoidance significantly moderates the relationship between pressure and collusion with FSF. Specifically, the interaction between pressure and Uncertainty Avoidance yielded a positive regression coefficient of 0.098177 with a p-value of 0.0257 (< 0.05), while the interaction between collusion and Uncertainty Avoidance produced a coefficient of 0.010470 with a highly significant p-value of 0.0000 (<0.05). These results provide empirical support for hypotheses H₇ a and H₇ f, confirming that Uncertainty Avoidance amplifies the influence of both pressure and collusion on FSF. Theoretically, these findings align with Hofstede’s cultural dimensions theory, which suggests that individuals in low Uncertainty Avoidance cultures tend to tolerate ambiguity and are more inclined toward risk-taking behaviour. In such contexts, when individuals experience high levels of pressure, such as financial performance demands or regulatory scrutiny, they may be more likely to engage in unethical practices, including financial statement manipulation, as a means to regain control or ensure short-term survival. This behaviour is further supported by prospect theory, which posits that under threat or loss, decision-makers prioritise loss aversion, even if it entails unethical or deceptive actions. Logically, individuals in low Uncertainty Avoidance environments may perceive fraud as a rational risk management strategy, particularly under conditions of intense pressure. Moreover, the significant interaction between collusion and Uncertainty Avoidance suggests that cultural tolerance for ambiguity can weaken individual moral resistance when unethical actions are conducted collectively. This is consistent with social facilitation theory, wherein individuals are more prone to deviant behaviour in group settings, particularly in ambiguous or weakly regulated environments where personal accountability is diffused. In this context, collusion becomes easier to justify, enabling collaborative fraud and undermining the integrity of financial reporting. These findings are in line with previous research by Roshanpoor et al. (2024), Yoo and Lee (2019) and Lannai et al. (2020), all of which highlight the amplifying role of Uncertainty Avoidance in facilitating fraud under specific environmental or psychological pressures.

Conversely, Uncertainty Avoidance does not moderate the relationship between opportunity, rationalisation, capability and arrogance with FSF. The interaction terms for each of these variables yielded statistically insignificant results: opportunity (coefficient = –0.001712; p = 0.9150), rationalisation (coefficient = –0.021571; p = 0.0970), capability (coefficient = 0.004049; p = 0.7270) and arrogance (coefficient = –0.001485; p = 0.5249). These findings indicate that Uncertainty Avoidance neither strengthens nor weakens the influence of these four factors on FSF, resulting in the rejection of hypotheses H₇b through H₇e. From a theoretical standpoint, these outcomes can be interpreted through the lens of institutional theory, which posits that actors embedded in high Uncertainty Avoidance environments tend to adhere more strictly to formal rules, procedures and normative expectations in order to minimise ambiguity and risk. Therefore, even when conditions conducive to fraud exist, such as perceived opportunities or internal justifications, individuals may abstain from unethical behaviour as a result of fear of sanctions, social disapproval or institutional consequences. Similarly, individuals with high capability or arrogance may be constrained by strong governance frameworks and regulatory enforcement in such settings, thereby neutralising their ability to act on personal motivations or cognitive biases. Logically, these results suggest that the mere presence of enablers like opportunity or capability is insufficient to trigger fraudulent conduct without a corresponding cultural or organisational context that permits risk-taking or ethical flexibility. In high Uncertainty Avoidance cultures, individuals tend to avoid behaviours that could disrupt established order, preferring compliance over opportunism. As a result, the influence of these factors is muted in environments where ambiguity is actively managed and deviance is institutionally discouraged. These findings are consistent with prior studies by Wang et al. (2023), Khairani et al. (2024), and Chen (2024), which also report that Uncertainty Avoidance weakens or fails to moderate the relationship between opportunity-driven fraud determinants and actual fraudulent outcomes.

In conclusion, this study finds that Uncertainty Avoidance significantly moderates the effect of pressure and collusion on FSF, but not the effect of opportunity, rationalisation, capability or arrogance. This nuanced understanding underscores the importance of incorporating cultural dimensions into fraud risk assessments, as they influence the behavioural pathways through which fraud determinants manifest within organisations.