Financial Risk Governance in the Digital Economy: The Complementary Roles of Whistleblowing Systems and Big Data Analytics in Indonesian State-Owned Enterprises

1. Introduction

The digital transformation of financial and organizational processes is reshaping the economic environment in which firms generate information, allocate resources, and manage financial risk. The increasing availability of large and heterogeneous datasets creates opportunities to improve financial decision-making and risk management, but it also creates new challenges for identifying financial misconduct (Putra et al., 2022). As financial and organizational processes become increasingly interconnected and data-intensive, risk signals may be embedded within complex transactions, organizational relationships, and reporting environments that are difficult to identify through conventional monitoring mechanisms (Arum et al., 2023; Vanini et al., 2023). Recent research in financial fraud detection similarly emphasizes that the growth of digital financial environments has increased both the volume of available information and the complexity of financial risk, requiring more data-driven approaches to risk management (Anisykurlillah et al., 2022; Triyanto, 2019a, 2020). Accordingly, financial statement misstatement risk should be considered not only an accounting-control issue but also a financial governance problem involving information quality, monitoring, and organizational decision-making.

Big Data Analytics (BDA) has emerged as an important organizational capability for strengthening financial risk assessment and fraud-risk monitoring (Farida, 2021; Möller & Törrönen, 2003; Sari et al., 2020; Surjaatmaja, 2018). By integrating and processing large volumes of structured and unstructured information, BDA can facilitate anomaly identification, pattern recognition, continuous monitoring, and evidence-based risk assessment. The increasing application of artificial intelligence, machine learning, predictive analytics, cybersecurity, and big data has further expanded the role of analytical technologies in addressing financial risk and fraud-related challenges (Haryono & Paminto, 2015; Putra et al., 2022; Tan, 2016). Recent FinTech research indicates that fraud detection increasingly requires attention to robustness, reliability, interpretability, and operational usefulness under changing information environments (Chen et al., 2023; Roszkowska, 2021). These developments indicate that financial risk governance increasingly depends on the ability of organizations to transform heterogeneous information into actionable risk intelligence. Recent work on digital financial fraud and forensic accounting likewise emphasizes the role of data-driven tools and digital evidence in fraud prevention and risk management (C. Daraojimba et al., 2023; Dosumu et al., 2024).

However, technological capability alone may not be sufficient to address the behavioral and organizational conditions underlying financial risk. Fraud-related risks can emerge within organizational environments characterized by performance pressure, opportunities arising from weak controls, rationalization, managerial capability, excessive ego or arrogance, and collusive relationships (Dlalisa & Govender, 2020; Farida & Nuryakin, 2021; Yusrianti et al., 2020). The Fraud Hexagon provides a multidimensional framework for understanding these conditions by extending conventional fraud perspectives through the inclusion of collusion as an important organizational dimension (Cao et al., 2019). This perspective is particularly relevant to state-owned enterprises (SOEs), where complex governance structures, public accountability, managerial interests, and operational pressures may create heterogeneous financial risks (Lou & Wang, 2011; Nurcahyono et al., 2021; Vanini et al., 2023). Thus, digital transformation should be accompanied by governance mechanisms capable of identifying both behavioral warning signals and structural weaknesses in the financial information environment.

In addition to technological mechanisms, organizations require institutional channels through which potential misconduct can be reported, communicated, and investigated. Whistleblowing Systems (WBS) provide such a mechanism by enabling employees and other stakeholders to communicate information regarding suspected misconduct to organizational authorities (Achyarsyah, 2022). WBS can therefore be viewed not merely as an internal control mechanism but as an information-generation capability within the broader financial risk governance process (Achmad, Ghozali, et al., 2022a; Handayani et al., 2023). Information obtained through whistleblowing may provide signals concerning managerial behavior, control weaknesses, unusual transactions, or other forms of misconduct that are not readily observable from conventional financial data (Putra et al., 2022). Such information may contribute to earlier identification and assessment of potentially material financial risks. Recent evidence also points to the relevance of combining data-driven technology and whistleblowing mechanisms for fraud detection in Indonesian institutional settings (Achyarsyah, 2022; Uswati Dewi et al., 2015).

The effectiveness of WBS, however, may depend on an organization’s ability to process, integrate, and evaluate the information generated through reporting channels. This creates an important connection between WBS and BDA. WBS can generate potentially valuable risk signals, while BDA can enhance the organization’s ability to integrate these signals with financial, operational, governance, and other organizational data (Achyarsyah, 2022; Saluja, 2024; Triyanto, 2019b). In this sense, WBS represents an information-generation mechanism, whereas BDA represents an information-processing and analytical capability. Their interaction may therefore provide a stronger basis for financial risk monitoring than either mechanism operating independently. This perspective is consistent with research emphasizing that the usefulness of analytical systems depends not only on their ability to identify patterns but also on the reliability and operational relevance of the information used in organizational decisions (Anisykurlillah et al., 2023; Prasetiyo et al., 2024).

Prior research on Indonesian SOEs has examined the relationship between Fraud Hexagon dimensions and financial statement fraud and has also considered WBS as a moderating mechanism (Agung et al., 2021; Raihan Noval Akbar et al., 2022). Recent evidence has demonstrated that WBS can influence the relationships between Fraud Hexagon dimensions and financial statement fraud in Indonesian SOEs (Nikmah & Arjoen, 2023; Nugroho & Diyanty, 2022; Triyanto, 2019b). This literature provides important evidence that organizational reporting mechanisms can influence the relationship between fraud-risk conditions and financial reporting risk. However, the technological capability required to process, integrate, and exploit fraud-related information has received considerably less attention (Quraini & Rimawati, 2018; Sallal et al., 2021). In particular, limited evidence explains whether BDA can strengthen the effectiveness of WBS in mitigating the relationship between fraud-risk conditions and financial statement fraud. Evidence from Indonesian SOEs and other listed sectors also suggests that Fraud Hexagon proxies can produce heterogeneous results across institutional contexts (Achmad, Hapsari, et al., 2022; Muawanah & Sari, 2023) (Agung et al., 2021; Sari et al., 2022).

This gap is important because digital financial risk governance requires more than the existence of reporting channels or the availability of analytical technologies. The value of information generated through WBS depends partly on whether organizations possess the analytical capability to connect reported signals with financial and operational patterns. BDA may strengthen this process by enabling organizations to identify anomalies, cross-reference information, detect unusual relationships, and support continuous risk monitoring. The FinTech literature further suggests that analytical fraud detection should be assessed in terms of practical reliability and operational usefulness rather than relying solely on isolated measures of analytical performance (Roszkowska, 2021; Saluja, 2024). Models or analytical approaches that perform well under particular conditions may become less effective when information environments change, highlighting the importance of robustness and continuous evaluation (Chen et al., 2023).

Against this background, this study examines financial risk governance in Indonesian SOEs within the context of digital transformation. Specifically, it investigates whether the six dimensions of the Fraud Hexagon—pressure, opportunity, rationalization, capability, arrogance, and collusion—are associated with financial statement fraud risk, whether WBS weakens these relationships, and whether BDA strengthens the mitigating role of WBS. The study therefore links behavioral risk conditions with two organizational capabilities that are increasingly relevant to digital financial governance: the generation of internal risk information and the analytical processing of heterogeneous data (Achmad, Ghozali, et al., 2022a; Handayani et al., 2023) (Roszkowska, 2021).

The study contributes to the economics and finance literature in three ways. First, it extends financial risk governance research by connecting behavioral fraud-risk conditions with organizational information and analytical capabilities. Second, it conceptualizes WBS and BDA as complementary capabilities within a financial risk information architecture, with WBS generating human-based signals and BDA supporting their integration with financial, operational, and governance data. Third, it provides firm-level evidence from Indonesian SOEs, an institutional setting in which public accountability, governance complexity, and digital transformation intersect. The study therefore moves beyond a narrow accounting-control interpretation and examines how information capabilities shape the monitoring of financial risks in public enterprises.

This framing is particularly relevant to an economics and finance audience because weaknesses in financial reporting and financial risk governance can affect the reliability of information used by boards, investors, creditors, auditors, regulators, and other economic decision makers. The study does not estimate economy-wide or market-level effects; rather, it provides firm-level evidence on governance mechanisms that shape the information environment of public enterprises. By identifying how WBS and BDA interact with observable fraud-risk conditions, the analysis offers a micro-level perspective on digital transformation, financial risk management, and institutional accountability in an emerging-market economy.

2. Financial Risk Governance and Hypotheses Development

2.1. Fraud Risk as a Financial Governance Challenge

Fraud theories have evolved from the Fraud Triangle proposed by (Cressey, 1953), which identifies pressure, opportunity, and rationalization as the primary conditions underlying fraudulent behavior. Wolfe, David T., (2004) extended this perspective by introducing capability, emphasizing that fraudulent behavior may require individuals with the authority, skills, access, and ability to exploit weaknesses within organizational systems. The Fraud Pentagon subsequently incorporated ego or arrogance as an additional behavioral dimension, while extended the framework by introducing collusion (Ghaisani et al., 2022; Vousinas, 2019). The resulting Fraud Hexagon provides a broader perspective for understanding fraudulent behavior by incorporating behavioral, organizational, and relational conditions that may facilitate financial misconduct (Achmad, Ghozali, et al., 2022b; Khamainy et al., 2022). The framework is particularly relevant to contemporary organizations because fraud may arise not from a single condition but from the convergence of incentives, control weaknesses, individual characteristics, and coordinated behavior.

The Fraud Hexagon comprises six dimensions: pressure, opportunity, rationalization, capability, arrogance, and collusion (Wijayani & Ratmono, 2020). Pressure refers to financial, organizational, or performance-related demands that may create incentives to manipulate financial information (Sinarti & Nuraini, 2019). In SOEs, such pressure may arise from performance targets, managerial expectations, political demands, financial constraints, or organizational restructuring. Opportunity reflects weaknesses in internal control, governance, monitoring, or information systems that allow fraudulent activities to be committed and concealed (Awalluddin et al., 2022). Rationalization represents the cognitive process through which individuals justify unethical actions as acceptable, necessary, or consistent with organizational circumstances(Awalluddin et al., 2022). Capability refers to the knowledge, authority, access, and technical competence that enable individuals to exploit organizational systems and conceal fraudulent activities. Arrogance reflects excessive confidence, entitlement, or perceived superiority that may lead individuals to disregard organizational rules and ethical standards fraud(Triyanto, 2019b). Finally, collusion represents coordinated actions among two or more individuals to commit or conceal fraudulent activities.

These dimensions are particularly relevant to financial statement fraud because fraudulent reporting may involve both individual behavior and organizational conditions (Nikmah, N., & Arjoen, 2023; Nikmah & Arjoen, 2023). Pressure may motivate management to achieve financial or performance targets, while opportunity may arise when internal controls or monitoring mechanisms are ineffective. Capability may determine whether individuals possess sufficient authority and expertise to manipulate accounting information, whereas arrogance may reduce sensitivity to ethical constraints or governance mechanisms (Surjaatmaja, 2018). Collusion can further complicate detection because coordinated actions may distribute responsibility, conceal evidence, and reduce the effectiveness of controls based on individual-level monitoring (Cao et al., 2019). Rationalization, meanwhile, provides a behavioral mechanism through which individuals may justify actions that would otherwise be considered unacceptable. Thus, the Fraud Hexagon provides a multidimensional foundation for examining why financial statement fraud may emerge within complex organizations.

The relevance of the Fraud Hexagon becomes increasingly important in digitally transforming SOEs. Digital systems increase the volume, availability, and interconnectedness of financial and operational information, potentially improving organizational monitoring while simultaneously creating more complex information environments (Alkaabi et al., 2019; Al-Mohammedi, 2020). Recent FinTech fraud detection research similarly emphasizes that the increasing scale and complexity of financial information require analytical approaches capable of identifying fraud patterns that may not be readily captured by conventional methods (Vanini et al., 2023). However, the availability of more data does not eliminate the underlying behavioral and organizational conditions that generate fraud risk. Instead, digital transformation may change the environment in which those conditions operate. Control weaknesses may involve digital access and information systems, managerial capability may include the ability to exploit technological systems, and collusion may be facilitated through interconnected organizational relationships. Consequently, technological development should be considered alongside the behavioral and organizational mechanisms identified by the Fraud Hexagon.

Based on this reasoning, the first hypothesis is formulated as an integrated proposition:

H1: 

Fraud Hexagon dimensions pressure, opportunity, rationalization, capability, arrogance, and collusion are positively associated with financial statement fraud in Indonesian SOEs.

2.2. Whistleblowing Systems as an Information-Generation Capability

Whistleblowing Systems (WBS) represent an important organizational mechanism for identifying and reporting suspected misconduct (Hardika et al., 2018; Sari Pramono et al., 2020). WBS provide formal channels through which employees and other stakeholders can communicate information concerning unethical or fraudulent activities to responsible organizational authorities. Such channels may include confidential hotlines, online reporting platforms, email-based reporting, and established procedures for receiving, protecting, investigating, and responding to reports. In this respect, WBS constitute an important component of organizational fraud governance because they create an institutional pathway through which information concerning potential misconduct can enter the organization’s oversight processes.

From an information perspective, the function of WBS differs from conventional financial monitoring. Financial statements and accounting systems primarily provide structured information concerning recorded economic activities, whereas whistleblowing can generate information about suspected misconduct that may not yet be observable in financial records (Meitasir et al., 2022). WBS can therefore be conceptualized as an information-generation mechanism within an organization’s broader fraud detection architecture. By providing accessible and protected reporting channels, WBS increase the possibility that information concerning irregular behavior reaches internal audit, risk management, compliance, or other oversight functions.

This distinction is particularly important in complex fraud environments. Fraud-related information may exist in different forms and may originate from different organizational actors. Employees may observe unusual management decisions, control circumvention, inappropriate transactions, conflicts of interest, or coordinated activities before these conditions become clearly visible in financial statements (Serrasqueiro & Caetano, 2015). WBS provide a mechanism through which such human-generated information can complement formal accounting and governance information. Consequently, the effectiveness of WBS should not be understood solely in terms of whether a reporting channel exists, but also in terms of whether the information generated through that channel can enter organizational processes for assessment and response.

The potential effectiveness of WBS is particularly relevant to the six dimensions of the Fraud Hexagon. Reporting mechanisms may expose activities associated with performance pressure, identify weaknesses that create opportunities for misconduct, challenge rationalizations, reveal misuse of authority or technical capability, identify arrogant or unethical managerial behavior, and provide information concerning coordinated activities among multiple actors (Daun Tasik Allo & Rachmawati, 2022; Sinarti & Nuraini, 2019). In this way, WBS may weaken the process through which fraud-risk conditions are translated into financial statement fraud by increasing the probability that suspicious behavior is reported and investigated.

The role of WBS is also consistent with the broader development of contemporary fraud detection (Prasetiyo et al., 2024). Recent FinTech research emphasizes that fraud detection should not be evaluated solely according to whether an analytical system identifies fraud under controlled conditions, but also according to whether the resulting information remains useful and reliable for operational decision-making. In the organizational context examined here, WBS provides an important upstream source of fraud-related information. However, the existence of information alone does not guarantee effective detection. Reported information must be assessed, connected with other evidence, and translated into organizational action.

Previous research has examined WBS as a moderating mechanism between Fraud Hexagon dimensions and financial statement fraud in Indonesian SOEs (Prasetiyo et al., 2024). The present study extends this perspective by positioning WBS within a broader digital information architecture. Rather than treating WBS only as a conventional internal control mechanism, the study considers WBS as the organizational mechanism through which human-generated fraud signals can enter a subsequent analytical process (Achyarsyah, 2022). This conceptualization creates a direct theoretical connection between WBS and BDA.

Accordingly, the second hypothesis is formulated as an integrated proposition:

H2: 

Whistleblowing Systems weaken the positive relationships between Fraud Hexagon dimensions and financial statement fraud in Indonesian SOEs.

2.3. Big Data Analytics as an Information-Processing Capability

Big Data Analytics (BDA) provides organizations with the capability to collect, integrate, process, and analyze large volumes of structured and unstructured information. In fraud detection, BDA can support anomaly identification, pattern recognition, predictive analysis, relationship analysis, and continuous monitoring (Surjaatmaja, 2018). These capabilities are particularly relevant in information environments in which fraud-related signals may be distributed across financial transactions, operational activities, governance information, and other organizational records.

Recent research in the FinTech domain demonstrates the increasing importance of analytical capability in fraud detection (Roszkowska, 2021). Conventional rule-based and statistical approaches may be useful in structured environments but can face limitations when financial information becomes increasingly large, complex, dynamic, and interconnected. Contemporary fraud detection research therefore increasingly considers not only predictive performance but also the robustness, reliability, and operational usefulness of analytical outputs(Achmad, Ghozali, et al., 2022a). These developments support the view that analytical technology should be understood as an organizational capability rather than simply as the adoption of a particular software application or analytical technique.

The role of BDA in this study is conceptually different from that of WBS. WBS generates potentially relevant information about suspected misconduct, whereas BDA provides the analytical capability to process, integrate, and interpret such information. This distinction is important because the existence of a reporting channel does not necessarily guarantee effective fraud detection(Rustiarini et al., 2021a, 2021b). Reported information may be incomplete, dispersed across organizational units, difficult to verify, or disconnected from other sources of evidence. BDA can enhance the usefulness of WBS-generated information by connecting reported signals with financial transactions, operational information, governance data, and other observable organizational patterns.

This complementarity suggests that BDA may strengthen the mitigating effect of WBS on fraud risk (Izzalqurny et al., 2019; Rehman & Hashim, 2020; Yayangida & Mohammed, 2023). When WBS operates without adequate analytical support, reported information may remain isolated within individual cases or depend heavily on manual investigation. By contrast, an organization with stronger analytical capabilities may be able to cross-reference whistleblowing information with large datasets, identify unusual relationships, detect recurring patterns, assess risk concentrations, and prioritize cases requiring further investigation. The analytical processing of multiple information sources may therefore increase the organizational value of WBS-generated signals.

This argument is also consistent with recent FinTech research emphasizing that fraud detection should be evaluated under realistic operational conditions rather than relying exclusively on performance under a single or controlled information environment (Arum et al., 2023; Riskiyadi, 2023). The benchmark study shows that fraud detection performance can vary across changing data conditions and that analytical outputs need to be assessed in terms of operational reliability and robustness. Although the present study does not develop a machine-learning fraud detection model or test distribution shift, the underlying implication is relevant: the effectiveness of technology depends on the organizational capacity to use information appropriately in changing and complex fraud environments.

Within this framework, BDA does not replace human judgment, whistleblowing, internal audit, or governance mechanisms. Instead, BDA complements these mechanisms by increasing the organization’s capacity to process heterogeneous information and identify relationships that may otherwise remain difficult to detect. This is particularly relevant to collusion, where suspicious behavior may involve relationships among several individuals or organizational units rather than a single observable transaction. Analytical capabilities may also help connect information from whistleblowing reports with patterns in financial and operational data, thereby supporting more targeted investigation and risk assessment (Dammak et al., 2022).

The original manuscript proposes that integrating BDA with WBS can strengthen organizational oversight and facilitate proactive fraud detection through predictive analytics, anomaly detection, and pattern recognition. The present study refines this argument by conceptualizing BDA not simply as another fraud-control instrument, but as an information-processing capability that enhances the effectiveness of an organizational information-generation mechanism. This distinction represents the central technological contribution of the proposed framework.

Accordingly, the third hypothesis is formulated as an integrated proposition:

H3: 

Big Data Analytics strengthens the mitigating effect of Whistleblowing Systems on the relationships between Fraud Hexagon dimensions and financial statement fraud in Indonesian SOEs.

2.4. Integrated Framework for Digital Financial Risk Governance

The theoretical framework integrates behavioral, governance, and technological perspectives on financial statement fraud (Nikmah, N., & Arjoen, 2023). The Fraud Hexagon identifies the behavioral and organizational conditions that may increase fraud risk. WBS provides an organizational mechanism through which information concerning suspected misconduct can be generated and communicated to relevant oversight functions. BDA complements this mechanism by providing the analytical capability required to process, integrate, and interpret fraud-related information.

The proposed framework therefore distinguishes between fraud-risk formation, information generation, and information processing. Fraud Hexagon dimensions represent conditions associated with the formation of fraud risk. WBS provides an information generation and reporting mechanism that may weaken the translation of these conditions into financial statement fraud (Meitasir et al., 2022). BDA provides an information-processing capability that can enhance the effectiveness of WBS by connecting reported signals with broader financial, operational, and governance information.

This distinction is important because digital fraud detection increasingly involves more than identifying whether a particular observation appears abnormal (Westland, 2022). Contemporary FinTech research highlights the importance of assessing whether fraud-detection outputs remain reliable, robust, and useful for operational decisions. From this perspective, the effectiveness of organizational fraud detection depends partly on how information is generated, processed, interpreted, and translated into action. WBS and BDA therefore represent complementary stages in an organizational information process rather than two independent anti-fraud mechanisms.

The framework further proposes that the value of WBS-generated information may increase when the organization possesses stronger analytical capabilities. A whistleblowing report may identify a potential fraud signal, but its detection value can be enhanced when the signal can be connected to transaction patterns, governance information, operational records, or relationships among organizational actors (Ghozali et al., 2019). BDA consequently provides an analytical layer through which human-generated information can be integrated with other forms of organizational evidence. This integration may improve the organization’s capacity to identify suspicious patterns, prioritize investigations, and respond to fraud-related risks.

The resulting framework moves beyond treating Fraud Hexagon, WBS, and BDA as independent components of an anti-fraud model. Instead, it proposes that behavioral risks, organizational reporting mechanisms, and analytical technologies interact within a broader fraud detection process. In particular, the analytical capability provided by BDA may increase the value of information generated through WBS and strengthen the organization’s capacity to identify and respond to fraud-related signals.

3. Materials and Methods

3.1. Research Design and Data Sources

This study employs an explanatory sequential mixed-methods design consisting of a quantitative phase followed by a qualitative phase. The quantitative phase examines the relationships between Fraud Hexagon dimensions and financial statement fraud risk (FSF), as well as the moderating roles of Whistleblowing Systems (WBS) and Big Data Analytics (BDA). The qualitative phase subsequently provides contextual explanations of the documented quantitative findings by exploring how fraud-risk conditions, reporting mechanisms, analytical capabilities, and organizational responses operate within Indonesian state-owned enterprises (SOEs).

The quantitative phase uses archival firm-year data from Indonesian SOEs and the documented empirical analysis was conducted using variance-based structural equation modeling (PLS-SEM) in WarpPLS 8.0. The model estimates direct relationships between the six Fraud Hexagon dimensions and FSF risk, two-way interaction effects between WBS and each Fraud Hexagon dimension, and three-way interaction effects involving BDA, WBS, and each Fraud Hexagon dimension. The analysis therefore focuses on standardized path coefficients (β) and their documented significance levels rather than panel-logistic coefficients or marginal effects.

The qualitative phase consists of semi-structured interviews with professionals whose responsibilities are related to auditing, risk management, information technology, cybersecurity, and organizational oversight. The interviews are used to explain the organizational mechanisms underlying the documented quantitative relationships rather than to provide a separate statistical test. The overall research sequence therefore follows QUAN → QUAL → integration, allowing the empirical patterns documented in the WarpPLS analysis to be interpreted within the organizational and technological context of Indonesian SOEs.

This design is appropriate for the present study because financial statement fraud risk involves both observable organizational characteristics and less directly observable processes concerning reporting, investigation, information processing, and managerial response. The quantitative phase identifies systematic associations in archival data, while the qualitative phase provides contextual insight into how WBS-generated information may be processed and used within organizational fraud-detection processes.

3.2. Population and Sample

The population consists of Indonesian SOEs identified from the official government SOE portfolio and SOEs listed on the Indonesia Stock Exchange (IDX) during the 2020–2024 observation period. The selection of SOEs is appropriate because these organizations operate under substantial public accountability and complex governance arrangements while simultaneously undergoing increasing digital transformation.

An initial population of 68 SOEs was identified. Twenty entities were subsequently excluded because they were banking SOEs or did not provide sufficiently complete information for the measurement of the study variables. The final sample therefore comprises 48 non-financial SOEs, generating 107 firm-year observations over the observation period.

A purposive sampling approach was applied because the analysis requires relatively complete annual information concerning financial statements, corporate governance, management characteristics, whistleblowing mechanisms, and technology-related disclosures. The sampling process is summarized in Table 1.

3.3. Data Sources and Data Collection

The quantitative data are obtained from publicly available corporate disclosures, including annual reports, audited financial statements, corporate governance reports, sustainability reports, risk management disclosures, and other relevant corporate documents. Financial and market-related information is cross-checked with IDX filings and company websites where necessary.

The archival approach enables the study to construct comparable firm-year observations across the observation period. Particular attention is given to disclosures concerning corporate governance, management characteristics, related-party transactions, political connections, whistleblowing mechanisms, and the adoption of analytical technologies.

The qualitative data are collected through 25 semi-structured interviews with professionals working in auditing, risk management, information technology, cybersecurity, and organizational oversight. Interview participants are selected based on their professional relevance to financial reporting, fraud risk management, internal control, digital information systems, and organizational monitoring.

The qualitative interviews focus on five broad areas: (1) the organizational conditions associated with fraud risk, (2) the functioning of whistleblowing mechanisms, (3) the processing and verification of whistleblowing information, (4) the use of data analytics in fraud-risk monitoring, and (5) organizational responses to fraud-related signals.

3.4. Variable Measurement

3.4.1. Financial Statement Fraud

Financial statement fraud risk is measured using the F-Score developed by Dechow et al. (2011). The F-Score provides an archival screening measure of the likelihood of material financial misstatement based on financial statement characteristics.

For the documented quantitative analysis, the continuous F-Score is transformed into a binary indicator:

FSFᵢₜ = {1, if F-Scoreᵢₜ > 1.00; 0, if F-Scoreᵢₜ ≤ 1.00}

where denotes the financial statement fraud indicator for firm in year.

An FSF value of 1 indicates that the firm’s F-Score exceeds the unconditional benchmark and therefore represents a higher likelihood of financial statement misstatement. It should not be interpreted as legal evidence that fraud has been established. Accordingly, FSF is interpreted throughout this study as an indicator of elevated financial statement misstatement risk rather than confirmed fraud.

3.4.2. Fraud Hexagon Dimensions

The six Fraud Hexagon dimensions constitute the primary explanatory variables. Consistent with the conceptual framework, the study examines pressure, opportunity, rationalization, capability, arrogance, and collusion (Handayani et al., 2023).

Pressure

Pressure (PRS) represents financial or organizational demands that may create incentives to manipulate financial information. It is measured using the firm’s financial leverage: A higher ratio indicates greater financial obligations relative to the asset base and is interpreted as a stronger observable condition of financial pressure.

Opportunity

Opportunity (OPT) represents conditions that may permit fraudulent behavior because of weaknesses in monitoring or governance. It is proxied by the proportion of independent commissioners: The measure captures an observable governance condition associated with the monitoring environment. Because the proxy does not directly observe fraudulent opportunity, its interpretation is limited to the governance characteristic used to represent the construct.

Rationalization

Rationalization (RTZ) represents the extent to which organizational circumstances may provide a basis for justifying questionable behavior. It is proxied by related-party transactions relative to revenue: A higher ratio indicates greater exposure to related-party transactions, which may provide an observable context in which questionable transactions or managerial judgments can be rationalized.

Capability

Capability (CP) represents the managerial ability and characteristics that may facilitate the exploitation of organizational systems. It is operationalized using CEO educational attainment: This variable is treated as an observable proxy rather than a complete measure of managerial capability.

Arrogance

Arrogance (ARG) represents excessive managerial confidence, entitlement, or perceived superiority. It is measured by the frequency of CEO photographs appearing in the annual report. A greater frequency is interpreted as an observable indicator of managerial prominence and excessive executive self-focus.

Collusion

Collusion (COL) represents coordinated behavior among individuals who may jointly facilitate or conceal fraudulent activities. It is proxied by the proportion of politically connected commissioners: This measure captures an observable governance characteristic associated with political connectedness and is used as a proxy for the collusion dimension.

3.4.3. Whistleblowing System

WBS is operationalized as a binary indicator of the existence of a formally established whistleblowing system: A value of 1 is assigned when the company’s official disclosures explicitly identify an established whistleblowing mechanism, such as a dedicated hotline, online reporting platform, reporting email, or formal reporting procedure. A value of 0 is assigned when such a formal mechanism is not disclosed.

The binary specification is used because archival corporate disclosures do not provide sufficiently consistent information to construct a comparable continuous measure of WBS quality or intensity across all firms and years. Accordingly, WBS captures the organizational existence and formal disclosure of a reporting mechanism, rather than its effectiveness, utilization rate, response time, case-resolution quality, or employee trust. This measurement limitation should be considered when interpreting the moderation results.

Conceptually, WBS is treated as an information-generation mechanism. It provides an institutional channel through which employees and other stakeholders can generate and communicate fraud-related information to organizational oversight functions.

3.4.4. Big Data Analytics

BDA is operationalized as a binary indicator reflecting the disclosed adoption of data analytics capabilities relevant to organizational monitoring, risk management, auditing, or fraud detection: A value of 1 is assigned when official corporate disclosures explicitly identify the adoption or use of big data or data analytics in areas such as risk management, auditing, fraud detection, anomaly identification, predictive analysis, monitoring, or related analytical activities. A value of 0 is assigned when such adoption is not disclosed.

In this study, BDA represents an organizational analytical capability rather than a specific machine-learning algorithm or an analytical model developed by the researchers. Because the measure is disclosure-based and binary, it captures reported adoption or use rather than the maturity, sophistication, accuracy, or actual operational effectiveness of BDA. The study therefore examines whether the presence of a reported analytical capability is associated with a stronger WBS–FSF risk relationship rather than comparing particular analytical technologies.

3.4.5. Control Variables

Two control variables are included to account for observable firm-level characteristics that may influence the likelihood of financial statement fraud.

Firm Size

Firm size (SIZE) is measured as the natural logarithm of total assets: Return on Assets

Profitability is measured using return on assets: Only these two control variables are included to maintain a parsimonious empirical specification and avoid unnecessary expansion of the model.

3.5. Empirical Model

The documented quantitative analysis was conducted using Warp-PLS 8.0 and a variance-based structural equation modeling (PLS-SEM) framework. The dependent variable, FSF, is coded as a binary indicator based on the F-Score threshold described above. The structural model examines three sets of relationships. First, the direct paths from the six Fraud Hexagon dimensions to FSF are estimated. Second, WBS × Fraud Hexagon interaction terms are included to test whether WBS weakens the positive associations between fraud-risk conditions and FSF. Third, BDA × WBS × Fraud Hexagon three-way interaction terms are included to test whether BDA strengthens the mitigating role of WBS.

The direct-effects specification can be expressed conceptually as: FSF = β0 + β1PRS + β2OPT + β3RTZ + β4CP + β5ARG + β6COL + controls + ε.

The second specification adds the WBS moderation terms: FSF = β0 + ΣβkHEXk + β7WBS + Σβ8k(HEXk × WBS) + controls + ε, where HEXk denotes each of the six Fraud Hexagon dimensions. where represents each of the six Fraud Hexagon dimensions.

A negative and statistically significant documented interaction coefficient is interpreted as evidence that WBS weakens the corresponding positive association between the Fraud Hexagon dimension and FSF risk. Because the available evidence consists of documented Warp-PLS coefficients, the manuscript does not claim marginal-effect or predicted-probability estimates that are not available in the retained output.

The third specification introduces the three-way interaction between each Fraud Hexagon dimension, WBS, and BDA: FSF = β0 + ΣβkHEXk + β7WBS + β8BDA + Σβ9k(HEXk × WBS) + Σβ10k(HEXk × WBS × BDA) + controls + ε.

A negative and statistically significant three-way interaction coefficient is interpreted as evidence that the documented mitigating association of WBS becomes stronger when BDA is present. This interpretation concerns moderation/complementarity and should not be read as evidence of causal prevention by BDA.

Because the original estimation output available for this manuscript is the documented WarpPLS output, the analysis does not introduce new panel-logistic estimates, marginal effects, predicted probabilities, or alternative robustness statistics that cannot be verified from the retained evidence.

3.6. Estimation Basis and Analytical Limitations

The quantitative findings reported in this manuscript are reproduced from the documented WarpPLS 8.0 analysis. The reported β coefficients and p-values are retained as the empirical evidence available for the study. Accordingly, the manuscript does not describe these coefficients as panel-logistic estimates and does not infer odds ratios, fixed-effects parameters, marginal effects, or predicted probabilities.

The dataset contains 107 firm-year observations from 48 non-financial SOEs over 2020–2024. The relatively small and unbalanced archival sample, together with the number of direct and interaction paths, warrants cautious interpretation of statistical significance and generalizability. The study therefore emphasizes the documented direction and significance of the associations rather than presenting the results as definitive causal effects.

No additional robustness analysis is claimed unless it is directly supported by the retained empirical output. In particular, alternative panel-logistic specifications, fixed-effects logit models, clustered standard errors, or marginal-effect calculations are not reported because the underlying raw dataset and corresponding estimation files are not available for independent re-estimation.

This reporting approach is intended to preserve methodological transparency. Future research with the complete firm-year dataset should replicate the model using alternative estimators, assess robustness to specification choices, and provide effect plots or marginal effects for the interaction terms.

3.7. Qualitative Analysis

The qualitative phase follows the documented quantitative analysis and is designed to provide contextual explanation of the observed relationships. Twenty-five semi-structured interviews were conducted with professionals in auditing, risk management, information technology, cybersecurity, and organizational oversight. Participants were selected for professional relevance to financial reporting, fraud-risk management, internal control, digital information systems, and organizational monitoring.

The interview data were analyzed thematically to identify recurring patterns related to fraud-risk conditions, reporting mechanisms, information processing, analytical capabilities, organizational responses, and governance challenges. Particular attention was given to how organizations receive, assess, verify, and respond to information generated through WBS and how analytical technologies may support the processing of such information.

The qualitative analysis does not constitute an independent causal test. Instead, it provides contextual evidence that helps explain why particular Fraud Hexagon dimensions may be associated with FSF risk, why WBS may weaken several of these relationships, and how BDA may enhance the organizational value of WBS-generated information.

The integration follows a QUAN → QUAL → integration sequence. The documented quantitative results first identify statistically significant or theoretically relevant relationships. Qualitative findings are then used to explain how and why these patterns may occur within Indonesian SOEs. The integrated interpretation focuses particularly on the interaction between WBS and BDA because the central proposition of the study is that organizational reporting mechanisms may become more useful when supported by appropriate analytical capabilities.

3.8. Integration of Quantitative and Qualitative Findings

The mixed-method integration is organized around three interconnected stages of technology-enabled fraud detection:

(1) Fraud-risk formation → (2) Information generation → (3) Information processing.

Fraud Hexagon dimensions represent the conditions associated with fraud-risk formation. WBS provides the organizational mechanism through which human-generated fraud-related information can be reported and communicated. BDA provides the analytical capability required to process, integrate, and interpret such information together with financial, operational, and governance data.

The integration therefore focuses on convergence and explanation between the quantitative and qualitative evidence. Quantitative results identify the statistical relationships among fraud-risk conditions, WBS, BDA, and FSF, while qualitative findings provide organizational explanations for the observed relationships. Particular attention is given to whether interview evidence supports the proposed complementarity between WBS and BDA.

This integrated approach reflects the broader direction of contemporary fraud detection research, which increasingly emphasizes the importance of connecting analytical capability with practical decision-making and organizational governance rather than treating technology as an isolated technical solution.

Table 2. Summary of Variable Operationalization.

Variable	Role	Operationalization	Reference
FSF	Dependent	1 if F-Score > 1.00; 0 otherwise	(Dechow et al., 2011)
PRS	Independent	Total liabilities / total assets	(Achmad et al., 2023)
OPT	Independent	Independent commissioners / total commissioners	(Abriatika & Mutmainah, 2022)
RTZ	Independent	Related-party transactions / total revenue	(Nindito et al., 2024)
CP	Independent	1 if CEO holds master’s degree; 0 otherwise	(Rochmah & Budiwitjaksono, 2025)
ARG	Independent	Frequency of CEO photographs in annual report	(Achmad et al., 2023)
COL	Independent	Politically connected commissioners / total commissioners	(Oktaviany & Reskino, 2023) (Fatimah & Januarti, 2025)
WBS	Moderator	1 if formal WBS is disclosed; 0 otherwise	(Raihan et al., 2024)
BDA	Moderator	1 if relevant BDA adoption is disclosed; 0 otherwise	(R. E. Daraojimba et al., 2023) (Putra et al., 2022)
SIZE	Control	Ln(total assets)	(Dechow et al., 2011)
ROA	Control	Net income / total assets	(Abriatika & Mutmainah, 2022)

Source: Authors’ compilation based on the cited literature (2026).

Table 2. Summary of Variable Operationalization.

Variable	Role	Operationalization	Reference
FSF	Dependent	1 if F-Score > 1.00; 0 otherwise	(Dechow et al., 2011)
PRS	Independent	Total liabilities / total assets	(Achmad et al., 2023)
OPT	Independent	Independent commissioners / total commissioners	(Abriatika & Mutmainah, 2022)
RTZ	Independent	Related-party transactions / total revenue	(Nindito et al., 2024)
CP	Independent	1 if CEO holds master’s degree; 0 otherwise	(Rochmah & Budiwitjaksono, 2025)
ARG	Independent	Frequency of CEO photographs in annual report	(Achmad et al., 2023)
COL	Independent	Politically connected commissioners / total commissioners	(Oktaviany & Reskino, 2023) (Fatimah & Januarti, 2025)
WBS	Moderator	1 if formal WBS is disclosed; 0 otherwise	(Raihan et al., 2024)
BDA	Moderator	1 if relevant BDA adoption is disclosed; 0 otherwise	(R. E. Daraojimba et al., 2023) (Putra et al., 2022)
SIZE	Control	Ln(total assets)	(Dechow et al., 2011)
ROA	Control	Net income / total assets	(Abriatika & Mutmainah, 2022)

Source: Authors’ compilation based on the cited literature (2026).

4. Results

Table 3 presents the descriptive statistics of the variables examined in the study. The analysis covers 107 firm-year observations from 48 Indonesian non-financial SOEs during 2020–2024. The descriptive statistics indicate meaningful variation in the financial, governance, behavioral, and technological characteristics of the sampled firms.

The mean value of FSF is 0.18, indicating that approximately 18% of the observations are classified as having a higher-than-expected likelihood of financial statement misstatement based on the F-Score classification. Pressure has a mean of 0.55, suggesting substantial variation in financial leverage across the sampled observations. Opportunity has a mean of 0.46, while rationalization has a relatively low mean of 0.09.

Capability has a mean of 0.48, indicating that the CEO educational indicator takes a value of one for approximately 48% of the observations. Ego/arrogance has a mean of 6.95, with values ranging from 3 to 13, reflecting considerable variation in CEO visibility across annual reports. Collusion has a mean of 0.21, indicating variation in the proportion of politically connected commissioners.

The organizational mechanisms examined in this study are also relatively prevalent. WBS has a mean of 0.62, indicating that 62% of the firm-year observations disclose an established formal whistleblowing system. BDA has a mean of 0.57, indicating that 57% of the observations disclose the adoption or use of relevant big-data or data-analytics capabilities. These findings suggest that both reporting and analytical mechanisms have become relatively common among Indonesian SOEs, although their adoption remains heterogeneous.

The descriptive statistics provide an initial overview of financial statement misstatement risk, Fraud Hexagon conditions, whistleblowing mechanisms, analytical capabilities, and firm-level characteristics within the sampled SOEs. Because WBS, BDA, CP, and FSF are binary variables, their means can also be interpreted as the proportion of observations coded as one. The remaining variables provide information regarding the dispersion and distribution of the observable characteristics used in the documented Warp-PLS model.

The descriptive results should be interpreted as evidence of variation within the sample rather than as evidence of causal relationships. They provide the empirical context for the documented PLS-SEM associations reported in Table 4.

4.1. Documented Empirical Results

The documented empirical analysis examines three relationships. First, it evaluates whether Fraud Hexagon conditions are associated with FSF risk. Second, it examines whether WBS weakens these relationships. Third, it assesses whether BDA strengthens the mitigating role of WBS.

The coefficients reported below are reproduced from the documented Warp-PLS 8.0 analysis retained for this manuscript. They should not be described as newly estimated panel-logistic coefficients. Because the underlying raw dataset and original estimation project are not currently available for re-estimation, the results are presented as documented empirical evidence.

Table 4. Documented Empirical Results.

No.	Relationship	β	p-value	Interpretation
1	Pressure → FSF	0.265	0.010	Significant positive
2	Opportunity → FSF	0.188	0.035	Significant positive
3	Rationalization → FSF	0.092	0.112	Not significant
4	Capability → FSF	0.203	0.022	Significant positive
5	Ego/Arrogance → FSF	0.145	0.041	Significant positive
6	Collusion → FSF	0.157	0.028	Significant positive
7	WBS × Pressure → FSF	−0.132	0.039	Significant negative
8	WBS × Opportunity → FSF	−0.118	0.045	Significant negative
9	WBS × Rationalization → FSF	−0.074	0.098	Not significant
10	WBS × Capability → FSF	−0.121	0.042	Significant negative
11	WBS × Ego/Arrogance → FSF	−0.109	0.048	Significant negative
12	WBS × Collusion → FSF	−0.138	0.030	Significant negative
13	BDA × WBS × Pressure → FSF	−0.146	0.027	Significant negative
14	BDA × WBS × Opportunity → FSF	−0.133	0.034	Significant negative
15	BDA × WBS × Rationalization → FSF	−0.091	0.072	Not significant
16	BDA × WBS × Capability → FSF	−0.137	0.029	Significant negative
17	BDA × WBS × Ego/Arrogance → FSF	−0.124	0.041	Significant negative
18	BDA × WBS × Collusion → FSF	−0.149	0.021	Significant negative

Source: Authors’ calculations based on Warp-PLS 8.0 output (2026). Note: β and p-values are reproduced from the documented Warp-PLS 8.0 analysis. They are not panel-logistic estimates and should not be interpreted as odds ratios, marginal effects, or causal effects.

Table 4. Documented Empirical Results.

No.	Relationship	β	p-value	Interpretation
1	Pressure → FSF	0.265	0.010	Significant positive
2	Opportunity → FSF	0.188	0.035	Significant positive
3	Rationalization → FSF	0.092	0.112	Not significant
4	Capability → FSF	0.203	0.022	Significant positive
5	Ego/Arrogance → FSF	0.145	0.041	Significant positive
6	Collusion → FSF	0.157	0.028	Significant positive
7	WBS × Pressure → FSF	−0.132	0.039	Significant negative
8	WBS × Opportunity → FSF	−0.118	0.045	Significant negative
9	WBS × Rationalization → FSF	−0.074	0.098	Not significant
10	WBS × Capability → FSF	−0.121	0.042	Significant negative
11	WBS × Ego/Arrogance → FSF	−0.109	0.048	Significant negative
12	WBS × Collusion → FSF	−0.138	0.030	Significant negative
13	BDA × WBS × Pressure → FSF	−0.146	0.027	Significant negative
14	BDA × WBS × Opportunity → FSF	−0.133	0.034	Significant negative
15	BDA × WBS × Rationalization → FSF	−0.091	0.072	Not significant
16	BDA × WBS × Capability → FSF	−0.137	0.029	Significant negative
17	BDA × WBS × Ego/Arrogance → FSF	−0.124	0.041	Significant negative
18	BDA × WBS × Collusion → FSF	−0.149	0.021	Significant negative

Source: Authors’ calculations based on Warp-PLS 8.0 output (2026). Note: β and p-values are reproduced from the documented Warp-PLS 8.0 analysis. They are not panel-logistic estimates and should not be interpreted as odds ratios, marginal effects, or causal effects.

4.1.1. Fraud Hexagon and Financial Statement Fraud

The documented empirical results indicate that five of the six Fraud Hexagon dimensions are positively and significantly associated with financial statement fraud risk (FSF). Pressure exhibits the strongest documented association with FSF (β = 0.265, p = 0.010), followed by Capability (β = 0.203, p = 0.022), Opportunity (β = 0.188, p = 0.035), Collusion (β = 0.157, p = 0.028), and Ego/Arrogance (β = 0.145, p = 0.041). In contrast, Rationalization does not show a statistically significant relationship with FSF (β = 0.092, p = 0.112).

The pattern suggests that financial statement fraud in Indonesian SOEs is associated more strongly with observable organizational conditions and the ability to exploit existing weaknesses than with rationalization as an underlying psychological process. Pressure may arise from financial targets, operational expectations, and demands from multiple stakeholders (Sinarti & Nuraini, 2019). Opportunity reflects weaknesses in monitoring and control environments, while capability and ego relate to the authority and influence required to exploit such weaknesses (Yusrianti et al., 2020). Collusion further indicates that fraud risk may emerge through relationships among multiple actors rather than through isolated individual behavior.

From a digital financial environment perspective, these findings are relevant because digitalization does not necessarily eliminate the underlying conditions that create fraud risk (Rehman & Hashim, 2020). Instead, increasing reliance on interconnected information systems may alter how such risks emerge and become observable. Weaknesses in data governance, access controls, monitoring processes, and information integration may create opportunities that are difficult to identify through conventional financial review alone. Consequently, the technological transformation of financial processes needs to be accompanied by organizational mechanisms capable of identifying behavioral and structural warning signals.

The non-significant relationship involving Rationalization requires particular consideration. Rationalization concerns an individual’s internal justification for unethical behavior and may therefore be difficult to observe through archival corporate disclosures (Awalluddin et al., 2022). The absence of statistical significance should not be interpreted as evidence that rationalization is irrelevant to fraudulent behavior. Rather, it may indicate that this dimension is less readily captured by the observable proxy used in the study.

Overall, the findings support a multidimensional interpretation of financial statement fraud. Pressure, opportunity, capability, ego, and collusion appear to represent important organizational and behavioral conditions associated with FSF, whereas rationalization is less visible through archival information. This provides an important foundation for examining whether organizational information mechanisms, particularly WBS, can interrupt the relationship between these risk conditions and financial statement fraud.

4.1.2. The Moderating Role of Whistleblowing Systems

The documented interaction results indicate that WBS weakens the relationships between Pressure, Opportunity, Capability, Ego/Arrogance, and Collusion and FSF risk. This supports the interpretation of WBS as an information-generation mechanism within financial risk governance. The three-way interaction results further indicate that BDA strengthens the mitigating association of WBS for these dimensions. Together, the results support a complementary information architecture in which reporting channels generate human-based risk signals and analytical capabilities support their processing and integration. These findings should be interpreted as documented associations rather than evidence of causal fraud prevention.

These findings suggest that WBS can function as an organizational information-generation mechanism within the fraud detection process(Triyanto, 2020). Unlike financial statements and other structured accounting records, whistleblowing reports can provide human-generated information concerning suspicious behavior, control weaknesses, managerial conduct, and relationships among organizational actors. Such information may reveal risk signals that are not immediately observable through conventional financial monitoring.

The moderating effect is particularly relevant to Pressure and Opportunity. These conditions may develop within organizational processes without producing immediately identifiable accounting anomalies. Employees and other organizational stakeholders may, however, observe unusual instructions, excessive performance demands, control circumvention, or inappropriate access to organizational resources. A formal WBS provides an alternative channel through which these signals can enter the organization’s monitoring and investigative processes (Achyarsyah, 2022; Crumbley & Ariail, 2020).

The same logic applies to Capability and Ego/Arrogance. Individuals with substantial authority, expertise, or organizational influence may have greater ability to exploit weaknesses in financial and information systems (Surjaatmaja, 2018). A formal reporting mechanism provides an information pathway that does not depend exclusively on hierarchical supervision. This is particularly relevant in large organizations in which information is distributed across multiple organizational levels.

The negative interaction involving Collusion is especially relevant from a technology-enabled fraud governance perspective. Collusive arrangements can involve multiple actors and may therefore be difficult to identify through isolated transaction-level review. WBS can generate information about relationships, coordinated activities, or suspicious conduct that may otherwise remain outside formal monitoring processes. In this sense, WBS can complement conventional financial controls by introducing an additional source of organizational intelligence.

The non-significant interaction involving Rationalization again highlights the difficulty of observing internal psychological processes through organizational reporting mechanisms (Awalluddin et al., 2022). WBS can facilitate the reporting of observable misconduct, but it may have limited capacity to directly alter the cognitive justification underlying fraudulent behavior (Yendrawati, 2019).

From a FinTech perspective, the broader implication is that digital fraud governance should not rely exclusively on structured transaction data. WBS represents a source of human-generated information that can complement financial, operational, governance, and digital activity data (R. E. Daraojimba et al., 2023). Its value therefore depends not only on the existence of a reporting channel but also on the organization’s ability to assess, integrate, and act upon the information received.

Accordingly, WBS should be understood as an information-generation layer within a broader fraud detection architecture. Its effectiveness depends on accessibility, confidentiality, responsiveness, and the extent to which reported information is connected to organizational risk assessment, audit, and investigative processes.

4.1.3. Big Data Analytics and the Enhanced Effect of Wbs

The documented three-way interaction results indicate that BDA may strengthen the mitigating role of WBS in the relationship between Fraud Hexagon dimensions and FSF. Negative three-way interaction coefficients are documented for all six dimensions. Statistically significant effects are observed for Pressure (β = −0.146, p = 0.027), Opportunity (β = −0.133, p = 0.034), Capability (β = −0.137, p = 0.029), Ego/Arrogance (β = −0.124, p = 0.041), and Collusion (β = −0.149, p = 0.021). The three-way interaction involving Rationalization is not statistically significant (β = −0.091, p = 0.072). The strongest documented negative interaction is observed for Collusion.

The pattern provides support for the proposed relationship between organizational reporting and analytical capability (Surjaatmaja, 2018). WBS generates potentially valuable fraud-related information, whereas BDA provides an organizational capability for processing, integrating, and interpreting information from multiple sources. The combination of these capabilities may therefore strengthen the organization’s ability to identify fraud-related patterns that would be difficult to detect through either mechanism independently.

This distinction is important for positioning the study within the FinTech literature. BDA is not conceptualized here as a specific machine-learning algorithm or as an automated replacement for auditors and investigators (Chen et al., 2023). Rather, it represents an organizational analytical capability through which financial, operational, governance, and other relevant information can be integrated and examined systematically. This allows human-generated information from WBS to be considered alongside structured organizational data.

The effect involving Collusion is particularly noteworthy. Because collusion reflects interactions among multiple actors, its detection may require the examination of relationships and patterns across different sources of information. Analytical capabilities can potentially assist organizations in connecting whistleblower information with transaction patterns, organizational relationships, governance information, and other relevant indicators (Riskiyadi, 2023). Thus, BDA may increase the informational value of WBS by enabling broader analysis of otherwise fragmented signals.

At the same time, the non-significant interaction involving Rationalization indicates that analytical capability may not resolve all dimensions of fraud risk equally (Yayangida & Mohammed, 2023). Rationalization represents an internal cognitive process that may remain difficult to observe even when organizations possess extensive data and analytical capabilities. This reinforces the importance of combining technological capabilities with human judgment and organizational oversight.

Overall, the documented results are consistent with the central proposition of this study: WBS generates fraud-related information, while BDA enhances the organizational capacity to process and connect that information with other sources of evidence. The contribution of BDA therefore lies not simply in increasing the volume of analyzed data, but in improving the organization’s ability to transform heterogeneous information into actionable fraud-risk intelligence.

4.2. Integrated Interpretation

Taken together, the documented findings support an integrated view of technology-enabled fraud detection in which behavioral risk conditions, organizational information mechanisms, and analytical capabilities are interconnected (Rustiarini et al., 2021a). The direct relationships indicate that several Fraud Hexagon conditions are associated with greater FSF risk. The negative WBS interactions indicate that formal reporting mechanisms can weaken several of these relationships, while the additional negative three-way interactions suggest that this mitigating role may become stronger when BDA capabilities are present.

The findings reinforce the conceptual distinction developed in this study. Fraud Hexagon represents fraud-risk formation; WBS represents information generation; and BDA represents information processing and analytical capability. The resulting framework positions fraud detection as an information process rather than as a purely accounting or technological activity.

This perspective also provides the study’s main contribution to the FinTech context. Digital fraud detection requires organizations to combine different forms of information and translate them into decision-relevant intelligence. Human-generated signals from WBS may provide contextual information that cannot be obtained from structured financial data alone, while BDA can facilitate the integration and analysis of these heterogeneous information sources.

Importantly, technology is not positioned as a substitute for governance, auditors, investigators, or human judgment. Instead, BDA is viewed as a capability that can increase the organizational value of information generated through WBS. This complementarity provides the central basis for understanding how organizational reporting mechanisms and analytical technologies can jointly support financial statement fraud detection in digitally transforming SOEs.

5. Discussion

5.1. Financial Risk Governance and Digital Transformation

The findings indicate that Big Data Analytics (BDA) strengthens the mitigating role of the Whistleblowing System (WBS) in the relationship between fraud-risk conditions and financial statement fraud. This finding suggests that the effectiveness of an organizational reporting mechanism depends not only on the existence of a channel through which suspicious activities can be reported, but also on the organization’s capacity to process, integrate, and interpret the information generated through that channel.

In this study, BDA is conceptualized as an organizational analytical capability rather than simply as a collection of technological tools. The distinction is important because the availability of analytical technologies does not automatically improve fraud detection. Their organizational value depends on the ability to systematically use data to identify unusual patterns, connect information across sources, and support timely risk assessment and investigation (Achmad, Ghozali, et al., 2022a; Handayani et al., 2023). This interpretation is particularly relevant to digitally transforming financial environments, where financial and non-financial information is increasingly generated across interconnected systems.

The interaction between WBS and BDA provides an explanation for the enhanced mitigating effect observed in the results. WBS can generate human-based information that may not be visible in conventional financial datasets. Employees or other organizational actors may identify unusual transactions, inappropriate managerial instructions, control weaknesses, or suspicious relationships before such conditions become apparent through routine financial analysis. BDA can subsequently support the examination of these signals by connecting them with financial, operational, governance, and other organizational information. WBS and BDA therefore perform different but complementary functions: WBS generates fraud-related information, whereas BDA increases the organizational capacity to process and interpret that information.

This complementarity is particularly relevant to Pressure and Opportunity. Pressure-related risks may be reflected in unusual reporting patterns around performance targets, while opportunity may arise from weaknesses in processes, access controls, or monitoring arrangements (Daun Tasik Allo & Rachmawati, 2022). Such conditions may generate fragmented signals across different organizational systems. BDA can help connect these signals and examine them across broader datasets rather than evaluating each observation independently. In this respect, analytical capability can support a transition from predominantly reactive fraud investigation toward more systematic and continuous risk monitoring.

The strengthening effect of BDA is also relevant to Capability and Ego/Arrogance. Individuals with substantial authority, expertise, or organizational influence may possess greater ability to exploit weaknesses in organizational and information systems (Surjaatmaja, 2018). Analytical capability can provide an additional layer of scrutiny by allowing patterns to be examined across transactions, reporting periods, and organizational units. This does not eliminate sophisticated manipulation, but it can reduce reliance on isolated observations and increase the visibility of potentially inconsistent patterns.

The role of BDA becomes particularly important in relation to Collusion. When several actors coordinate fraudulent activities, individual transactions may appear legitimate when assessed separately. Analytical capability can provide a broader perspective by examining relationships, transaction patterns, timing, and deviations across multiple datasets. The integration of structured and unstructured information may therefore help organizations identify connections that would otherwise remain fragmented across different information sources.

The findings further demonstrate why WBS and BDA should not be treated as interchangeable anti-fraud mechanisms. WBS depends substantially on human observation, organizational trust, and willingness to report, whereas BDA depends on the organization’s ability to transform available data into meaningful analytical signals. WBS may therefore provide information that is absent from accounting databases, while BDA can assist in assessing whether reported concerns are supported by observable patterns in organizational data (Achyarsyah, 2022; Pamungkas et al., 2020; Wahyudi et al., 2019). Their combination represents a movement toward an integrated fraud-intelligence capability rather than a conventional control mechanism.

This interpretation strengthens the financial risk governance positioning of the study. The contribution of BDA is not conceptualized as the independent adoption of sophisticated technology, but as an organizational capability that improves the information architecture of financial risk management. From this perspective, digital fraud-risk monitoring involves not only collecting and analyzing increasingly large volumes of data but also connecting different forms of information so that organizational decision makers can identify, assess, and investigate emerging financial risks.(Izzalqurny et al., 2019)

Importantly, the findings should not be interpreted as evidence that BDA independently prevents or eliminates financial statement fraud (Oktaviany & Reskino, 2023). Analytical capabilities can identify patterns and generate risk signals, but organizational responses still depend on governance, investigation, professional judgment, and accountability. This distinction is particularly important in SOEs, where fraud risks may be embedded in institutional pressures, organizational relationships, and managerial structures that cannot be addressed through technology alone (Nurcahyono et al., 2021).

Overall, the findings support an information-processing perspective on technology-enabled fraud governance. WBS expands the type and amount of fraud-related information available to an organization, while BDA increases its capacity to process, integrate, and interpret that information. The combination provides a stronger basis for identifying fraud-related signals than either mechanism operating in isolation. This perspective positions BDA as an organizational capability that enhances the value of human-generated fraud information and provides a stronger conceptual connection between behavioral fraud risk, organizational reporting, and digital financial risk management.

5.2. Theoretical Contribution to Financial Risk Governance

The findings contribute to the literature by extending the understanding of financial statement fraud from a predominantly behavioral and control-oriented perspective toward an integrated technology-enabled information perspective. The theoretical contribution rests on the interaction among three elements: fraud-risk formation, information generation, and information processing.

First, the study extends the application of Fraud Hexagon Theory by placing fraud-risk conditions within a broader technology-enabled information environment. Previous applications of the Fraud Hexagon have primarily focused on identifying the behavioral and organizational conditions associated with fraudulent financial reporting. The present findings indicate that understanding these conditions also requires consideration of how information about potential fraud becomes visible and actionable within an organization. Pressure, opportunity, capability, ego, and collusion remain important sources of fraud risk, but their consequences may also be influenced by the organization’s ability to generate and process information concerning those risks.

Second, the study extends the conceptualization of Whistleblowing Systems beyond their conventional interpretation as internal control or compliance mechanisms. WBS is positioned as an information-generation capability because whistleblowing reports can provide information that is not necessarily available through financial statements, routine audit procedures, or formal monitoring systems. Employees and other organizational actors may possess direct knowledge of operational practices, managerial decisions, control weaknesses, and organizational relationships that remain difficult for external or hierarchical monitors to observe.

Third, the study develops a theoretical connection between organizational reporting mechanisms and analytical capability. BDA is conceptualized as an information-processing capability, rather than merely as technological investment. This distinction shifts attention from whether an organization adopts advanced analytical technology toward whether it possesses the organizational capability to integrate information and transform heterogeneous data into meaningful fraud-risk signals. The study therefore complements existing technology-oriented fraud research by connecting analytical capability to information generated through WBS rather than examining BDA as an isolated technological determinant.

Fourth, the study introduces the concept of WBS–BDA information complementarity as a central theoretical mechanism. WBS and BDA represent different but interconnected information channels. WBS captures observations, concerns, and contextual signals originating from organizational actors, whereas BDA processes broader volumes of structured and unstructured information to identify relationships and patterns. Their combination suggests that human-generated information and analytical technology can complement one another in fraud detection. This is theoretically important because neither mechanism alone provides a complete information environment: WBS may generate signals that are difficult to validate without additional data, while BDA may identify anomalies without necessarily explaining the organizational context behind them.

Fifth, the study contributes to research on financial economics, corporate governance, and digital transformation by connecting behavioral fraud-risk theory with analytical capability. Rather than viewing technology as a substitute for governance, the findings suggest a complementary relationship in which organizational mechanisms generate risk information and analytical capabilities increase the capacity to process that information. This shifts the discussion from technology adoption alone toward the information architecture through which financial risks are identified, interpreted, and addressed.

Taken together, the theoretical contribution can be conceptualized through three interconnected layers. The Fraud Hexagon captures the formation of fraud risk, the Whistleblowing System (WBS) facilitates information generation, and Big Data Analytics (BDA) provides information-processing and analytical capabilities. The integration of these three layers establishes a technology-enabled framework for financial risk governance in digitally transforming public enterprises.

This framework provides the central theoretical contribution of the study. It suggests that financial fraud-risk monitoring in digitally transforming organizations should not be understood solely as the identification of behavioral risk factors or the adoption of analytical technologies. Instead, it can be understood as an information process in which behavioral and organizational conditions generate risk, reporting mechanisms make otherwise hidden information visible, and analytical capabilities enable organizations to connect and interpret that information for governance and risk-management decisions.

5.3. Economic, Policy, and Managerial Implications

From an economic and policy perspective, the findings suggest that public-enterprise oversight should place greater emphasis on the quality of financial-risk information flows, the protection and credibility of reporting channels, accountability for follow-up investigations, and the organizational capacity to analyze heterogeneous risk signals. WBS and BDA should therefore be considered as complementary investments in the information infrastructure of financial risk governance rather than as isolated compliance or technology initiatives. Because both mechanisms are measured through disclosed adoption in this study, governance frameworks should distinguish formal availability from operational maturity, including accessibility, confidentiality, analytical integration, investigation responsiveness, and documented resolution. For managers, the findings imply that digital transformation investments may generate greater governance value when they are connected to internal reporting, risk management, internal audit, and decision-making processes.

6. Conclusions

6.1. Conclusion and Contribution

This study examined financial statement fraud risk as a financial-governance concern in Indonesian state-owned enterprises by integrating the Fraud Hexagon perspective with organizational whistleblowing and analytical capabilities. The documented Warp-PLS results indicate that Pressure, Opportunity, Capability, Ego/Arrogance, and Collusion are positively associated with FSF risk, whereas Rationalization does not show a statistically significant relationship. These findings suggest that observable organizational conditions and the ability of individuals or groups to exploit organizational weaknesses are important correlates of FSF risk, while rationalization is less readily captured by archival proxies.

The documented interaction results indicate that WBS weakens the relationships between Pressure, Opportunity, Capability, Ego/Arrogance, and Collusion and FSF risk. This supports the interpretation of WBS as an information-generation mechanism within financial risk governance. The three-way interaction results further indicate that BDA strengthens the mitigating association of WBS for these dimensions. Together, the results support a complementary information architecture in which reporting channels generate human-based risk signals and analytical capabilities support their processing and integration. These findings should be interpreted as documented associations rather than evidence of causal fraud prevention.

The three-way interaction results further indicate that BDA strengthens the mitigating association of WBS for Pressure, Opportunity, Capability, Ego/Arrogance, and Collusion. The finding is consistent with the proposed WBS–BDA information complementarity: WBS can generate human-based fraud signals, while BDA can provide an organizational capability for processing and connecting those signals with other sources of evidence. Because BDA is measured through disclosed adoption, the result should not be interpreted as evidence of the performance of a particular analytical algorithm.

The study therefore contributes an information-process perspective to technology-enabled fraud governance. Fraud Hexagon dimensions represent fraud-risk formation, WBS represents information generation, and BDA represents information processing and analytical capability. The contribution is particularly relevant to SOEs because financial, governance, operational, and organizational information may be distributed across multiple systems and organizational levels.

From a practical perspective, SOEs should avoid treating whistleblowing channels and analytical technologies as isolated initiatives. Greater organizational value may be achieved when WBS, internal audit, risk management, investigation, and analytical processes are connected so that human-generated information can be assessed alongside structured organizational data. Such integration can support more systematic identification and investigation of emerging fraud risks, while retaining human judgment and governance oversight.

Overall, the documented evidence supports a complementary relationship among fraud-risk identification, information generation, information processing, and organizational response. The study contributes to financial risk governance research by showing how behavioral risk conditions can be connected to organizational information mechanisms and digital analytical capabilities within a public-enterprise setting. The findings should not be interpreted as evidence that BDA or WBS independently causes fraud reduction; rather, they document associations consistent with an integrated information architecture for financial risk monitoring and governance. This distinction is important for economic decision makers because effective digital transformation depends not only on technology adoption but also on the institutional capacity to convert information into risk-management decisions.

6.2. Limitations

Several limitations should be considered when interpreting the findings. First, the quantitative analysis is based on 107 firm-year observations from 48 Indonesian non-financial SOEs. Although the dataset covers multiple firms and years, the relatively limited and unbalanced sample and its concentration in Indonesian SOEs constrain generalizability to private-sector organizations, financial institutions, or other institutional environments.

Second, WBS and BDA are measured using organizational disclosure regarding their implementation or adoption. These disclosure-based measures capture the reported presence of the mechanisms but do not fully capture their quality, intensity, sophistication, or actual effectiveness. Organizations classified similarly may therefore differ substantially in the maturity and operational effectiveness of their whistleblowing and analytical capabilities.

Third, several Fraud Hexagon dimensions are represented through observable archival proxies. Although these proxies provide a practical basis for examining fraud-risk conditions using secondary data, they cannot fully capture the psychological and organizational characteristics underlying constructs such as Rationalization, Capability, Ego/Arrogance, and Collusion. In particular, the non-significant finding for Rationalization should be interpreted cautiously because internal psychological justification is inherently difficult to observe through corporate disclosures.

Fourth, the dependent variable is based on the F-Score classification of financial statement misstatement risk rather than legally established cases of fraud. Accordingly, FSF in this study should be interpreted as an indicator of elevated financial statement misstatement likelihood rather than direct evidence that fraud actually occurred.

Fifth, the qualitative component provides contextual explanations of the quantitative findings but does not independently establish causal relationships. The interview evidence is used to explain organizational practices, experiences, and perceptions surrounding fraud risk, WBS, and BDA rather than to constitute a separate statistical test of the proposed relationships.

Finally, the study conceptualizes BDA as an organizational analytical capability and measures its reported adoption rather than testing the performance of a specific analytical algorithm. Consequently, the findings should not be interpreted as evidence regarding the comparative accuracy of particular machine-learning, artificial-intelligence, or anomaly-detection techniques.

6.3. Future Research

Future research can extend this study through comparative research involving SOEs and privately owned firms, including cross-country studies within ASEAN. Such designs could examine whether governance structures, regulatory environments, whistleblowing protection, political connections, and digital maturity alter the relationships identified here.

Future studies should develop more refined measures of BDA maturity. Rather than using a binary adoption indicator, researchers could examine predictive analytics, anomaly detection, machine learning, natural language processing, network analysis, and real-time monitoring to determine whether analytical maturity is associated with stronger fraud-risk detection capability.

Future research could also examine WBS quality and effectiveness rather than merely its existence. Accessibility, confidentiality, employee trust, protection against retaliation, response time, investigation quality, and management responsiveness may provide a more nuanced explanation of why some whistleblowing systems generate more useful fraud-related information than others.

At the operational level, future research could integrate transaction-level financial data, audit logs, procurement records, textual disclosures, whistleblowing reports, cybersecurity information, and other digital traces. Machine learning and explainable artificial intelligence could then be used to examine how different forms of evidence contribute to fraud-risk identification while maintaining appropriate human oversight.

Finally, longitudinal research could examine how organizations develop WBS and BDA capabilities over time and whether improvements in analytical capability are accompanied by sustained improvements in fraud-risk identification. Such research should also test alternative estimators and robustness specifications using complete raw firm-year data.