Therefore, this study aimed to examine the effect of Big Data, and forensic audit as mediating variables on fraud detection.. Forensic Audit connects the accounting, auditing and legal a
INTRODUCTION
Research objectives and research questions
Overall objectives: The study examined the determinants affecting to fraud detection Specific objectives:
This article examines the impact of big data on fraud detection within Vietnamese corporations, highlighting how the integration of forensic audit practices mediates this effect By leveraging big data analytics, companies can enhance their fraud detection capabilities, ultimately leading to more effective risk management and improved financial integrity.
(3) : Analyzed the effect of forensic audit on fraud detection in Vietnam Corporations.
In order to achieve those objectives, we conducted an empirical analysis in Vietnam Corporation to seek convincing answers for the following three research questions:
(1) Does Big data support auditors to detect fraud in Vietnam Corporations?
(2) Is Forensic audit an effective tool to delect fraud in Vietnam Corporations?
(3) Through Forensic audit as a mediator variable, does Big data support auditors to detect fraud belter in Vietnam Corporations?
Research Methodology
In this Chapter, we only present an overview of the research methods Details about the research methods will be presented in detail in Chapter 3.
Research subject: the effect of big data and forensic audit on fraud detection in Vietnam Corporations
Respondents: 300 internal, external and government auditors in Vietnam
Scope of the research: The research lasted from 1/1/2024 to 11/2/2024
This study employed a quantitative approach utilizing a survey method, distributing questionnaires via Google Forms It focused on primary data collection and identified respondents through the snowball sampling technique, which included Internal, External, and Government Auditors in Vietnam The data analysis was conducted using structural equation modeling (SEM) with SmartPLS tools.
Research Contribution
Forensic audit is an emerging field in Vietnam, with only two researchers currently exploring forensic accounting This study aims to lay the groundwork for further investigation into the impact and significance of forensic audits in detecting fraud.
This study explores the impact of Big Data and forensic audits on fraud detection, highlighting the mediating role of forensic audits in enhancing the effectiveness of Big Data in identifying fraudulent activities The findings provide valuable insights for internal, external, and government auditors, offering them a practical framework to improve fraud detection methods in the future.
Research Structure
Chapter 3: Research Methodology and research model Chapter 4: Data analysis and results
LITERATURE REVIEW AND RESEARCH MODEL
Some previous studies
The 2013 research titled "The Impact of Forensic Accounting on Fraud Detection," conducted by A O Enofe, P O Okpako, and E.N Atube in Benin City, Edo State, Nigeria, aims to explore the relationship between forensic accounting and fraud detection Utilizing a survey method, the study collected primary data through questionnaires, a common technique in such research The analysis of the data employed Chi-square statistics, which assess how well a model aligns with observed data To ensure validity, the data used for the Chi-square calculation was random, mutually exclusive, and derived from independent variables within a sufficiently large sample The study's significance level was computed using SPSS software.
- Square and Degree of freedom, and concluded the relationship between “Forensic
This article explores the use of accounting and fraud detection software alongside OLS regression analysis to test a specific hypothesis The results obtained will either support or refute this hypothesis, leading to a comprehensive discussion of the findings and the conclusions drawn from the analysis.
The following hypotheses were tested in this research:
- HOI: Forensic accounting does not affect fraud detection.
- H02: Forensic accounting cannot curb fraudulent activities.
The study demonstrated that forensic accounting plays a crucial role in fraud detection, proving to be an effective tool for identifying fraudulent activities within firms Additionally, it highlighted that forensic accounting can significantly help in reducing such fraudulent practices The findings were validated through Chi-square statistical software and OLS regression analysis, underscoring the importance of forensic accounting in enhancing organizational integrity.
The study titled "Effect of Forensic Audit on Bank Fraud in Nigeria," conducted by S J Inyada, PhD, Archie Nathanael Mulyawan, D.O Olopade, PhD, and John Ugbede in 2019, investigates the significant impact of forensic audits on bank fraud in Nigeria The primary objectives include assessing how forensic audits contribute to the detection of bank fraud and evaluating their effectiveness in preventing such fraudulent activities.
The research hypotheses were derived from the identified problems and objectives:
HOI: Forensic audit does not have positive impact on bank fraud detection
H02: Forensic audit does not have positive impact on bank fraud prevention.
The survey design was employed with a sample of 128 out of a population of
Data analysis employed the ordinary least square (OLS) regression model, revealing that forensic audits significantly improve the detection and prevention of bank fraud The findings confirm that forensic audits are an effective and efficient means of identifying, preventing, and mitigating bank fraud in Nigeria.
The research titled “Impact of Data Analytics on Reporting Quality of Forensic Audit: A Study Focus in Malaysian Auditors” was conducted by Kahyahthri Suppiah and Dhamayanthi Arumugam in 2019, aiming to examine how data analytics influences the quality of forensic audit reports Utilizing a positivism philosophy and a deductive approach, the study developed hypotheses and theories to guide the research A survey strategy was employed to collect relevant data from respondents, with a mono method focusing on quantitative analysis through self-administered questionnaires (SAQ) created via Google Forms The SAQ utilized Likert scales ranging from 1 (Completely disagree) to 5 (Completely agree) to gather responses from employees of commercial banks The study analyzed data from forensic audit and accounting service providers, employing various statistical techniques to assess the impact of four distinct variables on the use of data analytics in enhancing forensic audit reporting quality.
In Malaysia, a study utilizing the Statistical Package for the Social Sciences (SPSS) examined the relationship between data analytics and the quality of forensic audit reporting The results indicated a significant correlation between various factors and the effectiveness of data analytics in enhancing forensic audit reports among practitioners This research underscores the importance of data analytics in improving data privacy and user confidence for forensic auditors By promoting the use of data analytics, the study highlights its potential to streamline routine data collection and analysis, enabling auditors to concentrate on deeper analysis and findings relevant to their specific cases or assignments.
The 2019 research titled “Big Data in Auditing for the Future of Data Driven Fraud Detection,” conducted by Bambang Leo Handoko and colleagues in a CPA firm in Indonesia, explores how auditors must adapt to new technological approaches in auditing Utilizing a descriptive qualitative method and interviews with an auditor partner, the study reveals both advantages and disadvantages of big data in the auditing and fraud detection profession It emphasizes the necessity for audit firms to recognize internal and external obstacles that may impede the implementation of big data analytics As technology evolves, auditors are urged to rapidly adapt to these changes The findings underscore the significance of big data in auditing and provide valuable insights for auditors, accounting firms, and regulatory bodies in formulating effective fraud detection strategies.
The research titled "The Auditing Profession in Vietnam in the Context of the Information Age" by Nguyen Thi Phuong (2022) addresses two key questions regarding digital transformation in auditing Firstly, while there are currently no robots performing audits in Vietnam or globally, predictions suggest that within 15-20 years, robots could replace approximately 30% of auditors However, complete automation is unlikely due to factors such as the diversity of audit subjects, the necessity for professional judgment, and established standards Secondly, the study emphasizes the need for auditors to adapt to the digital age and the burgeoning Internet economy by thoroughly examining and evaluating transaction processes, ensuring the reliability of automation systems through checks on transaction existence, dates, and electronic signatures Utilizing qualitative methods, the research involved interviews with 20 auditors and audit assistants, with each session lasting 30 to 45 minutes, supplemented by data from internal documents and various sources.
Previous studies on fraud detection have often focused on individual factors rather than examining multiple influences, highlighting a significant research gap This study aims to explore the potential of big data as a powerful tool for auditors in analyzing fraud risks and uncovering their causes, particularly in the context of Vietnam, where research on big data in auditing is limited Additionally, forensic auditing is recognized as an effective method for detecting and preventing fraud, yet its application remains underexplored in Vietnamese corporations Despite the established importance of forensic audits in international research, this study will investigate the combined effects of big data and forensic auditing on fraud detection within Vietnam Corporations, contributing valuable insights to the existing literature.
Theoretical framework
Agency theory, proposed by notable authors such as Armen Alchian, Harold Demsetz, Michael Jensen, William Meckling, and S.A Ross, is grounded in the concept of the separation of ownership and management in joint-stock companies.
The article examines the contractual dynamics between shareholders and agents, such as attorneys and managers, who are tasked with acting in the owners' best interests While both parties aim to optimize their respective interests, there are instances where managers may not prioritize the owners' interests Eisenhardt (1989) identifies two critical issues in agency relationships: adverse selection and moral hazard.
Adverse selection occurs when principals question if an agent's skills match their salary, while moral hazard arises when shareholders doubt whether the agent is fully committed to their work or is instead seeking personal benefits, stemming from information asymmetry.
Agency theory highlights the conflict of self-interest in business, particularly between agents and shareholders Agents may face dilemmas when presented with projects that could increase company assets but also involve risks, leading them to prioritize short-term safety over potential long-term gains This conflict necessitates monitoring mechanisms, such as auditing, to bridge the information gap between agents and shareholders Auditing fosters trust by ensuring that agents act competently and ethically, thereby providing reliable assessments of financial statements Ultimately, this reinforces the integrity of financial reporting and protects shareholder interests.
Meanwhile, from the perspective of an agent (board of directors), auditing must guarantee service quality to meet the needs of clients.
The principal-agent problem and asymmetric information in Agency theory can lead management to manipulate financial reports, resulting in significant fraud cases that have harmed stakeholders and damaged Vietnamese corporations Recent scandals, such as Bibica Joint Stock Company's inflated operating expenses in 2002, Bach Tuyet Cotton Joint Stock Company's false revenue declaration in 2006, and Truong Thanh Furniture Corporation's unexpected loss of over 1,000 billion VND in 2016, highlight growing concerns among stakeholders Additionally, the manipulation of stock prices by former FLC Group chairman Trinh Van Quyet, which resulted in an estimated illegal gain of 975 billion VND, further erodes investor confidence in the Vietnamese market These incidents underscore the urgent need for experienced auditors to effectively detect and prevent fraud.
2.2.2 Computer-assisted audit techniques (CAATs)
Computer-Assisted Audit Techniques (CAATs), also referred to as Computer-Assisted Audit Tools and Techniques, are essential programs and data that auditors employ during the audit process These tools facilitate data processing and the execution of audit tests to achieve specific objectives Regardless of the environment—manual or technology-driven—the fundamental goals of auditing remain unchanged It is crucial to thoughtfully select appropriate technical tools based on the context to ensure the collection of relevant and adequate evidence.
CAATs significantly enhance the audit process by improving efficiency, allowing auditors to quickly and accurately analyze large data sets within tight deadlines These tools not only optimize resources by cutting costs and reducing time and effort but also play a vital role in identifying fraud through detailed examination of accounting software In an increasingly globalized environment, the implementation of CAATs paves the way for innovative audit methodologies Furthermore, CAATs improve the management of audit data and working papers, making them an essential addition to business workflows.
To effectively utilize information technology (IT) tools in auditing, auditors must possess a thorough understanding of these tools and their functionalities, highlighting the importance of training an IT-capable team within CPA firms Additionally, auditors and IT specialists should evaluate clients' information systems by examining general control activities and application controls to assess reliability General control activities cover the entire system, including setup, development, program modifications, physical security measures, backup procedures, and contingency planning, while application controls focus on data input, processing, and output control.
Basic issues about forensic audit
Forensic auditing, as defined by the Institute of Forensic Auditors (IFA), is a systematic process involving the collection, verification, analysis, and reporting of data, primarily aimed at gathering evidence for legal proceedings It entails a thorough examination of an individual’s or company’s financial information, serving as crucial evidence in court This discipline integrates accounting, investigative auditing, criminology, and litigation services, as highlighted by Dada, Owolabi, and Okwu (2011) Forensic audits involve collecting, verifying, and analyzing data to uncover facts and tangible evidence related to legal and financial disputes, irregularities, and fraud prevention, according to Enofe et al (2015) Such audits can be essential for prosecuting fraud, embezzlement, or other financial claims, and may also investigate negligence, misuse of powers, or undue benefits granted to individuals or companies.
A forensic audit is a thorough and objective examination of an individual’s or company’s financial statements, aimed at verifying their accuracy and identifying any financial gains achieved through misrepresentation or illegal activities.
2.3.2 The role of forensic audit
A forensic audit is initiated when there are suspicions of fraud, misconduct, or illegal activities within an organization Key operational red flags indicating potential fraud include a poor tone at the top, insufficient management oversight, absence of independent audits, weak internal controls, and Internal Audit recommendations for a fraud investigation If these warning signs are evident, it is crucial to consider conducting a forensic audit.
This study highlights the critical role of forensic auditing through four key stages: investigation, evidence gathering, reporting, and court proceedings During the investigation stage, the forensic auditor focuses on exploring the client's suspicions, such as potential fraud within the business related to provided materials The auditor then works to confirm these suspicions through thorough examination and analysis.
- What kind of fraud is being committed?
- When have the fraudulent activities been taking place?
- Who was involved in the fraudulent activities?
- How did they conceal the fraudulent activities?
- How do fraudulent activities affect the business?
The second stage of the audit process involves gathering evidence, which is crucial as it underpins the entire investigation During this phase, the forensic auditor must compile comprehensive evidence that details the nature of the fraud, its effects on the company, and identifies those responsible for the fraudulent activities.
In the final stage of the audit process, the auditor compiles a detailed report that outlines the identified fraud for the client This report includes essential components such as findings, a summary of evidence, the methods used in the fraudulent activities, and recommendations for preventing future occurrences Furthermore, it acts as vital evidence in court if the client decides to take legal action.
During court proceedings, the auditor's presence is crucial for clarifying the evidence collection process and identifying suspects It's essential for the auditor to simplify explanations, especially for those unfamiliar with accounting terms While a forensic auditor plays a vital role, they do not work alone; hiring a forensic auditor requires selecting a CPA firm to utilize their services effectively.
Forensic Audit involves assessing evidence linked to a claim to determine its alignment with court-acceptable standards An example of this process is conducting a Forensic Audit of sales records to calculate the rent owed under a lease agreement currently in litigation (ACFE, 2014).
Forensic audit aims to develop computerized applications that improve an entity's internal control system while facilitating the analysis and presentation of financial evidence.
Forensic auditors employ their financial-economical and audit backgrounds to uncover facts for solving commercial or legal disputes, often linked to fraud suspicion (Godwin Emmanuel Oyedokun, 2015)
According to Carla Tardi (2022), Paul Munler (2022), there are five techniques following:
Forensic auditors utilize advanced data analysis techniques and software to scrutinize extensive financial data and transactions By leveraging data analytics, they can identify trends, anomalies, and potential red flags that may indicate fraudulent activities.
- Interviews and interrogations: Forensic auditors conduct interviews with employees, management, and anyone involved in the inquiry Skillful interviewing tactics can aid in gathering information and identifying prospective leads.
- Document Examination: Forensic auditors extensively study and analyze financial papers, invoices, receipts, contracts, bank statements, and other related data for anomalies or inconsistencies.
- Reconciliation and Cross-Verification: Forensic auditors reconcile financial records and cross-check information from several sources to uncover inconsistencies and track the movement of monies.
- Surveillance and Observation: In some circumstances, forensic auditors may perform surveillance or observation in order to obtain evidence of suspected fraudulent activity.
According to Godwin Emmanual Oyedokun (2015), a proficient forensic auditor should possess the following essential characteristics:
Ethical integrity is a fundamental requirement for forensic auditors, paralleling the standards expected of law enforcement personnel These professionals must uphold a strict code of conduct as they collaborate with law enforcement agencies to uncover criminal activities, ensuring they maintain a strong moral compass in their vital mission.
Forensic auditors are essential for analyzing financial reports and source documents, as they carefully assess the validity of transactions to ensure accurate recording When discrepancies arise between documentation and reported figures, they investigate the origins of these numbers to verify the company's disclosures Their analytical skills enable them to break down complex problems into manageable parts, a crucial trait in the forensic auditing field This profession requires a high level of analytical thinking, akin to that of historical figures like Sherlock Holmes and Albert Einstein While practice can improve these skills, a natural aptitude often plays a significant role in their success Given the complexity of their work, forensic auditors frequently face intricate challenges, sifting through vast amounts of information to reveal hidden financial misconduct.
A forensic auditor must possess strong detail-oriented skills, demonstrating a keen eye for discrepancies between supporting documents and financial reports This role requires meticulous comparisons of figures, as the auditor diligently identifies inconsistencies and seeks detailed information for every transaction The process is time-intensive, demanding thorough attention to specifics to ensure accuracy in final totals In essence, a detail-oriented forensic auditor is efficient, highly organized, and committed to achieving flawless results.
Forensic auditors are inherently inquisitive, as their role requires them to gather essential information for investigating financial reports They actively ask questions while reviewing the figures, engaging with staff to clarify each reported number This curiosity deepens their understanding of the underlying reasons for the data presented If staff members cannot provide satisfactory explanations, forensic auditors continue their inquiry by consulting other colleagues or supervisors to ensure a thorough investigation.
The fifth essential characteristic for success in financial investigation is intuition, which sets experienced professionals apart from novices, irrespective of gender Intuition enables individuals to make logical connections and insights without relying on conscious reasoning, allowing them to weave together seemingly unrelated information While it is possible to work as a financial investigator without strong intuition, doing so may present additional challenges in the field.
Basic issues about fraud
Cheating, defined by the Vietnamese dictionary as "deceitful behavior, trickery" (Institute of Linguistics, 2003), aligns with the International Auditing Standard (ISA) 240, which describes fraud as an intentional act perpetrated by individuals within management, the board of directors, employees, or external parties to achieve illegal or unjust gains The Association of Certified Fraud Examiners (ACFE) further expands this definition, characterizing fraud as a crime committed through deceit to obtain an advantage or profit.
Fraud in business can originate from both internal and external sources, including employees, managers, owners, customers, and suppliers While often committed by individuals, external fraud encompasses various schemes such as bid-rigging by dishonest suppliers, fraudulent invoicing, and employee bribery Customers may also engage in deceitful practices, like falsifying payment information or attempting to embezzle funds Additionally, businesses face threats like intellectual property theft and violations of proprietary information Occupational fraud is defined as the misuse of one’s professional role for personal gain through the intentional abuse of organizational resources and assets.
Until now, researchers present several classifications of the frauds For example, ACFE classifies occupational frauds into three classes: Corruption, Asset misappropriation, and Financial statement fraud.
Corruption refers to the misuse of power for personal benefit, taking various forms such as embezzlement, bribery, and election manipulation It affects multiple sectors, including business, government, judiciary, civil society, media, education, health, sports, and infrastructure This pervasive issue involves a broad range of individuals, from public servants and politicians to business professionals and everyday citizens.
Corruption undermines democracy, diminishes trust, and stifles economic growth, while worsening social division, poverty, and inequality It directly harms the economy through tax evasion and money laundering, and indirectly distorts fair competition, raising business costs By reducing returns on productive activities, corruption fosters inequality and threatens justice Transparency is crucial in combating corruption, as it sheds light on both formal and informal processes and regulations The right to access information is a fundamental human right that protects against corruption and builds trust in decision-makers and public institutions.
Asset misappropriation is the unauthorized use of an organization’s resources for personal gain, making it the most prevalent form of workplace fraud It can manifest through methods such as skimming, larceny, embezzlement, data theft, and payroll fraud, leading to significant financial losses that are often hard to detect without effective internal controls Fraudsters frequently rationalize their behavior and cleverly hide their misconduct To combat asset misappropriation, organizations must establish robust internal controls, conduct regular audits, and maintain a strict zero-tolerance policy towards fraud.
Financial statement fraud is a white-collar crime where individuals intentionally misrepresent a company’s financial statements by omitting or exaggerating information to create a misleadingly favorable view of the company’s performance Typically perpetrated by senior management, the motives behind this fraud include personal gain, ensuring company survival, and maintaining leadership status Common tactics include overstating revenue, falsifying expenses, and misappropriation, often in organizations with weak internal controls and aggressive leadership To combat financial statement fraud, prevention is key, but if it fails, swift detection is essential Implementing strict controls can minimize opportunities for fraud, and recognizing fraud red flags can help in the timely identification and apprehension of perpetrators.
The "Fraud Triangle Model," developed by criminologist Donald R Cressey in the 1970s, is a key theory in understanding fraud, which refers to intentional fraudulent activities for personal gain This framework identifies three critical components that heighten the risk of fraudulent behavior, providing insight into the motivations behind an individual's decision to engage in fraud.
Opportunity in fraud refers to the circumstances that enable dishonest acts for personal gain or pressure to achieve financial targets Top management and business owners significantly influence this component There is an inverse relationship between a business's control structure and the incidence of fraud; a robust control framework decreases the likelihood of fraud, while a weak one heightens the risk Various factors contribute to the creation of opportunities for fraudulent activities.
Weak internal controls pose significant risks to the integrity of accounting and financial data, as they are designed to ensure accuracy and reliability Insufficient division of responsibilities, lack of oversight, and poor record-keeping are common weaknesses that can create opportunities for fraud Strengthening internal controls is essential to mitigate these risks and protect organizational assets.
A weak tone at the top significantly undermines a company's ethical culture and values, primarily influenced by senior management and the board of directors This lack of strong ethical leadership increases the organization's susceptibility to fraudulent activities.
- Inadequate accounting policies: Accounting policies refer to how financial statement items are reported Accounting policies that are poor (inadequate) may allow employees to distort numbers.
To mitigate the risk of fraud and safeguard financial integrity, businesses should focus on the three key factors mentioned above Maintaining vigilance and implementing proactive measures are crucial for fostering a fraud-resistant environment.
Pressure is a powerful motivator that can push individuals to engage in actions they would typically avoid, particularly in cases of fraud Desperation often drives people to extreme measures, with pressures arising from both financial and non-financial sources.
- Repayment of Debt: The burden of debt repayment can create immense pressure, pushing individuals toward fraudulent behavior
- Falling Stock Prices: Financial losses due to declining stock prices can trigger anxiety and motivate unethical actions.
- Maintaining Reputation: The fear of damaging one's reputation may drive someone to engage in fraudulent activities.
Pressure with financial contents: Financial pressures can be either short-term or long term They arise when individuals find themselves in need of cash Let's explore some specific categories:
- Itching Palm and Greediness: A desire for financial gain, coupled with greed, can lead to fraudulent actions.
- Desire for a Comfortable Life: Aspirations for a better lifestyle may push individuals to cross ethical boundaries.
- High Personal Debts and Health Expenses: The weight of financial obligations can create immense stress, prompting desperate measures.
- Unexpected Financial Needs: Sudden emergencies or unforeseen expenses can drive individuals to seek illegal solutions.
Bad habits create significant pressures that are recognized as key contributors to fraudulent behavior, often stemming from specific human traits Understanding these underlying pressures is essential to addressing the issue effectively.
- Gambling, Drug, or Alcohol Addiction: Individuals grappling with these addictions may experience intense financial strain, leading them to consider fraudulent actions.
- Nightlife Habits: Late-night lifestyle choices can contribute to financial stress, pushing individuals toward unethical behavior.
Pressures related with jobs: job-related pressures emerge from various workplace dynamics:
- Job Dissatisfaction: When employees feel displeasure with their roles or work environment, they may succumb to the temptation of fraud.
- The idea of an Unfair Attitude: The perception of unfairness—whether in promotions, recognition, or treatment—can create immense pressure.
- Not getting Expected Promotion: Employees who expected promotions but were overlooked may feel compelled to seek alternative gains.
- Lower Wages structures: Financial struggles resulting from inadequate compensation can drive individuals to fraudulent acts.
So, understanding these pressures is essential for effective fraud risk management.
Rationalization serves as the defense mechanism employed by fraudsters to justify their actions Let's explore some common examples:
- “I had borrowed the money, I would pay back'’: The fraudster convinces themselves that they will eventually repay the funds they’ve taken.
- “This is in return for my efforts for the business”: They argue that their actions compensate for their hard work or contributions to the organization.
- “Nobody has suffered as a result of this”: They believe that no one has been harmed by their actions, they rationalize their behavior.
- “I have taken the money for the good purpose”: The fraudster claims to have used the money for a noble cause.
- “I didn’t know that this was a crime”: They plead ignorance, stating they were unaware their actions constituted a crime.
- “Business has deserved this”: Some justify their actions by believing the business somehow deserved it.
To prevent such rationalizations, businesses should establish a strong moral code and provide comprehensive employee training.
Basic issues about Big Data
Big Data refers to vast and complex data sets generated daily from various sources, including satellite data, climate sensors, and web interactions While there is no single definition of Big Data, it encompasses both structured and unstructured information that presents challenges in storage, analysis, and visualization.
Big Data encompasses not only vast amounts of information but also analytical and predictive methods essential for informed decision-making According to Faravetto et al (2020), it is crucial for researchers in social sciences to define Big Data in practical terms, linking it to data collection and processing This approach enhances audit quality, aids in fraud detection, and enriches the auditing process by utilizing extensive data beyond traditional sampling methods However, the influx of data exceeding established capacities poses challenges to both internal and external auditing processes Traditional auditing methods become inefficient as they struggle to accurately assess large populations of relevant data, such as transactions, leading to unreliable outcomes To achieve precise results, it is vital to adopt reliable commercial perspectives, conduct thorough analyses, and maintain a high level of attention.
Auditors encounter distinct challenges in harnessing and managing vast amounts of big data By leveraging big data analytics, they can enhance the effectiveness and application of the data at their disposal.
This method assists auditors in identifying hidden patterns and relationships, as well as making informed decisions.
According to Insurance el al (2013), Big Data includes three attributes:
Structured data is organized in a defined format and adheres to a specific database schema, making it easy to query and extract relevant information Most organizations utilize this well-defined structure to manage their data efficiently.
Semi-structured data refers to information that, while organized in a structured format, is not stored in traditional databases but rather in flat files This type of data typically consists of a dynamic blend of data and metadata, which is formatted for user presentation Common examples include XML data, JSON files, and the HTML source code of websites.
Unstructured data represents the largest volume of data generated in today's world, with companies adopting various standards to manage it based on user needs This data can originate from internal sources, such as memos and meeting notes, or external sources like reports and journals Predominantly descriptive, unstructured data lacks a defined format, making it challenging to categorize consistently.
2.5.3 The main components constitute Big Data
According to ICAEW (2015), Big data is often characterized by the ‘3 Vs’ - variety, velocity, and volume.
Big Data is an extensive data source categorized into three main types: structured, semi-structured, and unstructured Structured data, stored in tagged data warehouses, is easily sortable, while unstructured data is random and challenging to analyze In contrast, semi-structured data consists of discrete chunks, making it unsuitable for fixed fields.
- Velocity: (or the size of data) is now larger than terabytes and petabytes The sheer volume and multiplication of data outstrips existing storage and processing methods.
- Volume: referring to the speed at which data is generated and processed in order to satisfy the demands and problems of growth and development.
2.5.4 Factors create the formation and development of Big Data
The Big Data movement is driven by three key factors: advancements in computing power, the emergence of new data sources, and the development of infrastructure for knowledge creation, all of which play a crucial role in its establishment and growth (ICAEW, 2015).
The remarkable growth in computing power and storage capacity is the cornerstone of big data, allowing for the efficient collection and processing of large and complex data sets This expansion in processing capabilities is often likened to an exponential increase, facilitating deeper insights and analysis from diverse data sources.
As computing power increases, it becomes easier to collect and manage data from diverse sources, including the vast information available on the internet from user interactions like search clicks and website visits aimed at making purchases Additionally, social media platforms contribute a rich array of data types, encompassing status updates, comments, likes, images, videos, and user networks.
The digital infrastructure has revolutionized collaboration and knowledge creation, fostering trends like crowdsourcing and open source software This exchange of information has united diverse communities and unveiled valuable insights from unexpected sources.
2.5.5 The importance of Big data to fraud detection
Fraud poses a significant challenge for auditors in the Big Data landscape, as an abundance of information does not guarantee its relevance The increased complexity of Big Data complicates the evaluation of audit evidence related to fraud (Srivastava et al, 2009; Srivastava, 2011; Fukukawa et al, 2014) Effective fraud detection requires a thorough understanding of the internal control system, regardless of the analytical tools available for processing all business transactions Importantly, even with a robust internal control framework, the sheer volume and intricacy of Big Data can hinder the detection of potential fraud.
RESEARCH METHODOLOGY
Research Process
The research team developed a path model comprising a structural and measurement model to illustrate the relationships between research variables and test the hypotheses The structural model outlines the connections between latent variables, while the measurement model links these latent variables to their observed scales (Hair et al., 2014) The research process begins with identifying the research problem, reviewing literature, and establishing a structural model This leads to the creation of a comprehensive measurement model, which includes both cause and result measurement models After defining specific scales for each research concept, the team tested the measurement model through data collection and scale evaluation, considering factors like unidimensionality, multidimensionality, reliability, and validity (Nguyen Dinh Tho, 2013) Following this, the team performed structural model analysis to test hypotheses by examining P-values and t-tests, utilizing the structural equation modeling (SEM) approach with SMART Partial Least Square (PLS) software.
The concluding phase of the research process involves analyzing the results from PLS-SEM, which includes testing the measurement and structural models, conducting a Bootstrapping test, and drawing conclusions regarding the research hypotheses in Chapter 4 Additionally, Chapter 5 will outline the implications of the research findings.
Research Methodology
This research employed a quantitative methodology utilizing a survey distributed via Google Forms, focusing on primary data collection through a snowball sampling technique The participants included Internal, External, and Government Auditors operating in Vietnam The study was conducted over a period from January 1, 2024, to November 2, 2024.
The Snowball method is an effective non-probability sampling technique ideal for recruiting participants who are otherwise hard to reach, making it particularly beneficial for studies focused on Vietnam corporations This approach begins with initial participants, known as "seeds," who refer additional participants, creating a chain of referrals until the desired sample size or saturation point is achieved For our research, we might initiate the process with a few qualified auditors who possess extensive experience and expertise in forensic auditing, subsequently asking them to recommend other auditors with similar qualifications.
Snowball sampling is an effective method for accessing sensitive or private populations while addressing ethical considerations In research areas such as fraud detection, participants often hesitate to share information due to concerns about privacy and potential risks This sampling technique prioritizes confidentiality, enabling researchers to collect valuable data without compromising participant trust.
Structural Equation Modeling (SEM) is utilized to examine the interplay between satisfaction and both financial and non-financial factors, necessitating a substantial sample size due to its reliance on sample distribution theory (Raykov and Widaman).
In research, a larger sample size enhances reliability by minimizing sampling errors, as supported by findings from 1995 This study utilizes a sample size of 247, which effectively meets the necessary criteria for the analysis method employed.
Structural Equation Modeling (SEM) is a linear analysis technique that effectively identifies measurement errors and integrates abstract concepts, as noted by Fornell (1982) and cited in Hair et al (2014) This method will be employed to test the research hypotheses, allowing for a comprehensive comparison between theoretical frameworks and empirical data By bridging the gap between theory and data, SEM enhances the understanding of complex constructs in research (Fornell, 1982; cited from Hoang Le Chi, 2014).
This study employs a complex research model, utilizing SmartPLS 3.0 software for PLS-SEM analysis, an effective data processing technique that integrates both the theoretical (structural) and measurement models PLS-SEM, based on ordinary least squares (OLS) regression, facilitates the evaluation of research hypotheses and assesses the reliability and validity of the research concepts' scales through SEM analysis.
In conclusion, SEM and SmartPLS 3.0 facilitate robust statistical analysis, while the Snowball method allows for effective engagement with specific target populations This approach is particularly valuable for addressing intricate research questions when conventional sampling techniques are inadequate.
Research model and Hypothesis
The integration of big data in auditing significantly enhances fraud detection by enabling auditors to analyze vast amounts of information from diverse sources, thus improving analytical processes and audit outcomes This aligns with agency theory, which posits that big data can address agency problems like fraud, particularly in government institutions (Syahputra & Afnan, 2020) Big data streamlines data collection, enhances visualization, and fosters better communication within teams (Hipgrave, 2013) Notably, companies like Alibaba have demonstrated the effectiveness of big data in fraud detection (Chen, Tao, & Wang, 2015) A study revealed that 72% of 466 surveyed companies recognize the vital role of big data technology in fraud prevention and detection, underscoring its efficiency and effectiveness (Ernst & Young, 2014).
Hl: Big Data positively and significantly affects Fraud Detection
Big data encompasses extensive and intricate data sets that necessitate advanced technology, significantly enhancing the audit field It empowers forensic auditors by enabling them to detect fraud more effectively, particularly through the analysis of non-financial and unstructured data, including management news, contract specifics, and meeting outcomes Furthermore, the vast volume and integrated nature of big data allow forensic auditors to conduct analytical procedures with greater speed and efficiency.
Big Data significantly enhances external audits by improving the quality of audit evidence and aiding in fraud detection (Kyunghee Yoon et al., 2015) Its potential for conducting population-based audits leads to more relevant evidence (Roshan Ramlukan, 2015) For instance, auditors can utilize GPS data to verify shipping information, resulting in more accurate shipment verification Thus, Big Data has a positive impact on audits (Syahputra & Athan, 2020), supporting the formulation of the following hypothesis.
H2: Big Data positively and significantly affects forensic audit
Forensic audits involve the meticulous collection, verification, processing, analysis, and reporting of data to establish legally valid facts and evidence aimed at preventing criminal activities and financial irregularities, including fraud (Enofe et al., 2015) To achieve this, forensic auditors must possess a robust strategy underpinned by extensive knowledge, skills, and experience Proficiency in diverse fields such as criminology, information technology, and accounting is essential for effective forensic auditing, making it one of the most reliable methods for detecting and uncovering fraud (Alao, 2016; Enofe et al., 2015; Zachariah et al., 2014).
H3: Forensic Audit positively and significantly affects Fraud Detection
Big Data Analytics is an efficient and intelligent tool, which helps to identify and avoid the risk of security and fraudulent behavior (T Sirisha Madhuri, E Ramesh Babu,
In 2023, forensic auditors with specialized expertise significantly enhance the effectiveness of Big Data in accelerating and refining fraud detection processes (Tang & Karim, 2019) Research by Tang and Karim (2017) indicates that integrating Big Data with forensic audits leads to more effective fraud detection Consequently, this forms the basis for the proposed hypothesis.
H4: Forensic audits mediate the effect of Big Data on Fraud Detection
Measuring variables in the model
Big Data encompasses vast and intricate datasets that necessitate advanced technology to enhance the auditing process As noted by Rezaee and Wang (2017), the measurement of Big Data is based on seven key components.
Encoding Variables Scale items Source
Big data can perform data mining and modeling in forensic accounting investigation.
BD2 Big data can extract, transform, and leverage syndicated data for use in forensic accounting practices.
BD3 Big data can use and interpret datasets that may not have standard data formats.
BD4 Big data has advanced analytical other data management skills.
BD5 Effectively present findings to diverse audiences using strong verbal, written, and visual communication skills?
BD6 Big data has a clear and coherent database digital strategy.
BD7 Big data aligns people, processes and culture.
Forensic audit involves the systematic collection, verification, analysis, and reporting of data to gather evidence for legal proceedings The evidence obtained from a forensic audit can be crucial in court cases and may also help rectify situations that could lead to fraudulent activities According to Enofe, A., Omagbon, P., & Ehigiator, F (2015), the effectiveness of forensic audits can be assessed through five key criteria.
Encoding Variables Scale items Source
Key objective of forensic audit is to detect fraud and report same.
FA2 Forensic audit can guarantee prompt detection of fraud in a firm.
FA3 Forensic audit can help in reviewing existing internal control.
FA4 Forensic audit can help guarantee the safeguard of assets from unauthorized use.
FA5 Forensic audit can guarantee strategic prevention of fraud.
Fraud involves deceptive practices designed to secure an unfair advantage or inflict financial damage, including activities such as embezzlement, identity theft, and fraudulent financial reporting It is quantified through five specific items as outlined by Enofe, A., Omagbon, P., and Ehigiator, F (2015).
Encoding Variables Scale items Source
FRl Fraud (FR) Income smoothing increases the risks of fraud in an organization.
FR2 None inclusion of fraud reporting in the duties of The conventional auditor increases risks of fraud.
FR3 Lack of segregation increases the risks of fraud.
FR4 CEO’s override increases the risks of fraud.
FR5 Security breach increases the risks of fraud.
DATA ANALYSIS AND RESULTS
Qualitative research results
A total of 300 auditors, including internal, external, and government auditors in Vietnam, participated in the survey After filtering out invalid responses through two screening questions, 247 valid answer sheets were retained for analysis This final sample size aligns with the initially proposed number and is suitable for running the model using SmartPLS.
Quantitative research results
In a survey comprising 247 valid responses, participation was split with 40% of respondents being men and 60% women This gender ratio aligns closely with the demographics of the audit industry, where women outnumber men in accounting roles.
The survey revealed that the majority of participants were aged 21-30, comprising 30.5% of respondents, followed by those aged 41-50 at 29.3%, and 31-40 at 20.6% Additionally, 19.6% of participants were aged 50 or older This demographic distribution aligns with the age profile of auditors in Vietnam.
In a survey of 247 participants, 47.3% held Associate positions, 27.2% were in Senior roles, and 25.4% occupied managerial positions All respondents possessed at least a bachelor's degree, with 3.6% working as government auditors, 55.6% as external auditors, and 40.8% as internal auditors.
The survey reveals that small businesses represent 31.5% of the respondents, while medium-sized businesses account for 40.8%, and large enterprises make up 27.7% The distribution among these business sizes is closely aligned with actual industry proportions, enhancing the reliability and consistency of the survey findings.
The following is a table summarizing the descriptive statistics results related to the survey subjects from the collected data:
Category Sample size n$7 Category Sample size n$7
Education Bachelor 49,60% Company size work
(Source: compiled by the author)
To assess reliability and validity in research, we utilize specific metrics such as Cronbach's alpha and Composite Reliability (CR) for measuring reliability For validity, including both Convergent and Discriminant Validity, we employ the HTMT ratio, Average Variance Extracted (AVE), and a correlation matrix among the research variables.
+ Evaluate the value of the scale
The value of the scale is evaluated through the Outer Loading index of observed variables.
Outer Loading results show that all factors have observed variables greater than 0.7 According to Hair et al., 2016 (PLS-SEM book, page 51), Outer Loading is best when the result is >=0.7.
(Source: compiled by the author from Smart PLS)
+ Evaluate the reliability of the scale
All scales achieve reliability, as indicated by Cronbach’s Alpha (DcVellis, 2012) and composite reliability (CR) values above 0.7 (Bagozzi & Yi, 1988) for each scale.
Table 4.3: Results construct Reliability and Validity
(Source: compiled by the author from Smart PLS)
+ Evaluate the convergence of the scale
To assess convergent validity using SMART PLS, we will utilize the average variance extracted (AVE) index Hock & Ringle (2010) state that a scale exhibits convergent validity when the AVE is 0.5 or higher The lowest AVE recorded is 0.765 for the scale measuring the impact of Big Data, indicating that the results satisfy the required convergence value.
Table 4.4: The average variance extracted results
(Source: compiled by the author from Smart PLS)
+ Evaluate the discrimination of the scale
Table 4.5: Fornell-Larcker Criterion results
(Source: compiled by the author from Smart PLS)
The results presented in Table 4.5 indicate that all Fornell-Larcker indices are located on the diagonal, showcasing higher values for the same measurement scale in comparison to other indices from different scales Notably, the BD scale exhibits a Fornell-Larcker index of 0.875, which signifies a considerably higher discriminant value than other scales such as FA and FR This evidence confirms that the research sample maintains clear discriminant validity among the measured constructs.
Collinearity can significantly impact model outcomes, prompting the study to conduct a collinearity test According to Hair et al (2019), a Variance Inflation Factor (VIF) greater than 5 indicates a high likelihood of collinearity, whereas a VIF below 3 suggests minimal collinearity The findings reveal that all VIF coefficients for the examined factors are below 3, indicating that multicollinearity does not influence the results of the model analysis.
(Source: compiled by the author from Smart PLS)
+ Evaluate the relationship between factors
(Source: compiled by the author from Smart PLS)
Using SmartPLS software and running the Bootstrapping calculation 500 times, the results show the relationship and level of impact between factors.
Big Data (BD) has the same impact on Forensic Audit (FA) (beta coefficient = 0.683 and p-value less than 0.05) Shows that as Big Data (BD) increases, Forensic Audit (FA) will increase.
Big Data (BD) has the same influence on Fraud Detection (FR) (beta coefficient = 0.353 and p-value less than 0.05) Therefore, increasing Big Data (BD) will increase Fraud Detection (FR).
Forensic Audit (FA) has the same impact on Fraud Detection (FR) (beta coefficient 0.429 and p-value less than 0.05) Therefore, increasing Forensic Audit (FA) will increase Fraud Detection (FR).
Table 4.8: Relationships between groups of factors
(Source: compiled by the author from Smart PLS)
From the results table, it shows that the factor group Big Data (BD) -> Forensic Audit (FA) -> Fraud Detection (FR) has an impact and influence (beta coefficient= 0.293 and p-value < 0.05).
+ For the Forensic Audit mediating variable
4- From the Results table, it shows that Big Data (BD) -> Fraud (FR) (P-Value