The role of data analytics for detecting indications of fraud in the public sector

Technological developments play an important role in the audit process, one of which is the use of data analytics that are useful to assist auditors in analyzing data, collecting audit evidence, predicting risks that occur and will occur, and other things. The use of data analytics is also applied by public sector auditors to maintain accountability and responsibility for state finances. This study aims to examine the effect of using data analytics on indications of fraud for public sector examiners in Indonesia. Testing and data analysis techniques used STATA version 14 which processed answers from 33 auditors from two representative offices of public sector auditors in Java Province and Sumatra Province. The results of the study state that the use of data analytics has a positive and significant effect on indications of fraud for public sector examiners in the examination process. This means that public sector auditors can detect fraud using data analytics. © 2022 by the authors. Licensee SSBFNET, Istanbul, Turkey. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (


Introduction
The high number of fraud cases in Indonesia resulted in huge losses faced by the state, thus harming many parties.The important role of government agencies in achieving government goals that are free from fraud is highly prioritized, but there are still cases of fraud that occur even among government agencies, both central and regional.Factors that can cause fraud in government agencies are the competencies of the perpetrators, the code of ethics that is violated, having a leadership spirit that is not trustworthy in carrying out their responsibilities, and the use of technology that is misused so that the consequences of fraudulent actions that occur will have a significant influence on economic growth (Watriningsih, 2020).The importance of a strong internal control system must be applied to the public sector to be able to maintain the quality of financial reports (Mulyadi & Haryoso, 2019).Based on data from the corruption eradication commission, corruption cases that occurred from 2004-2020 increased to 409 cases among districts and cities, 382 cases in ministry institutions, 152 cases in provincial governments, 86 cases in state or regional-owned enterprises (BUMN/BUMD), 74 cases in parliament as public representatives, and 20 cases in the parliament's commission (Databoks, 2021).This data strengthened with data released by ICW in 2020, losses faced by the state amounted to Rp56.7 trillion with 1,218 corruption cases (Kompas, 2021).
Fraud is an act that is carried out intentionally by individuals or groups of people to take the results of resources that do not belong to them and that are used for their interests (ACFE, 2018).Acts of fraud and corruption in the public sector are carried out by members of the organization as a result of organizational management that lacks a professional attitude at work, thus triggering opportunities for fraud to occur (Agyei, 2017) Therefore, everyone who works with an organization needs to implement a code of ethics that limits the rights and obligations at work (Watriningsih, 2020).Fraud can occur due to factors contained in the fraud triangle (Cressey, 1953) (Tuana Kotta, 2016) namely pressure, opportunity, and rationalization.Factors that are often used by perpetrators to commit fraud are opportunities and situations that allow them to carry out fraudulent actions smoothly, such as taking advantage of situations and positions.Good leadership will benefit the organization by achieving superior performance as expected and avoiding fraudulent actions (Ferry, L., & Ahrens, 2017) but if the leader has unprofessional ethics, the organization will face failure in its performance (Rahim et al., 2017).The case of corruption in social assistance, which was carried out by the minister of social affairs, illustrates that positions can be misused for personal interests.The fraudulent act was exploited by the perpetrators due to the COVID-19 pandemic conditions by asking the public for funds of IDR 10,000 per family card and taking advantage of the perpetrator's position as a social minister who distributed social assistance to the community so that he dared to take the opportunity to commit the fraud of up to IDR 20.8 billion (Kompas, 2020).This is in line with research (Dewi et al., 2019) that fraud that occurs in the public sector can be carried out due to pressure, opportunity, and rationalization factors.The fraudulent act that is carried out begins with a small number of numbers, this can occur due to a lack of supervision system.
Based on Law no. 15 of 2004 the Public Sector Examiner in this case the Supreme Audit Agency (in Indonesia called BPK) has a role in eradicating corruption because it is tasked with calculating, assessing, and determining state losses in the use of the budget by entities.The findings of public sector examiners that contain criminal indications will be reported to law enforcement officers such as the prosecutor's office, the corruption eradication commission (KPK), and the police.In addition to leadership, the role of technology and information can also assist organizations in identifying fraud, if the organization utilizes technology properly (Watriningsih, 2020).Therefore, public sector auditors can apply data analytics to adapt to technological developments and maximize their responsibility for audits.The application of data analytics by public sector examiners is expected to help detect fraud that has occurred since the beginning (IAI, 2021).According to (Hartono, 2019) big data can be used in detecting fraud and used as tools to increase the effectiveness of fraud detection by maximizing the data contained in big data by using data analytics tools, to make it easier for auditors to predict the risk of fraud that will occur.Data analytics is useful to make it easier for auditors to detect fraud (Hipgrave, 2013).The use of data analytics implemented by public sector auditors can improve organizational governance that is accountable, and transparent, and becomes an example for other institutions (BPK, 2021).This is in line with several previous researchers who have analyzed technology's effect, especially in data analytics to detect fraud.The use of data analytics can assist auditors in maintaining the quality of financial reports that are clean from fraud by increasing anti-fraud controls because financial reports containing fraud will interfere with the reliability, accuracy, and efficiency of financial statements in terms of their existence and continuity (Kılıç, 2020).In addition, using data analytics could strengthen risk management based on big data processing and technology (Chen et al., 2015).The use of data analytics can also help in making decisions in more real-time and analyzing more relevant in the business environment (Appelbaum et al., 2017).Furthermore, data analytics can also be used as a tool for analyzing user behavior and transaction history can also be seen when using data analytics to detect fraud more effectively (Jennifer et al., 2013).Computer-assisted audit techniques can assist auditors in detecting fraud because auditors can more easily process data in the form of computer files.So that the auditor can plan fraud detection more quickly and precisely compared to detecting it manually (Atmaja, 2016).The use of artificial intelligence (AI) can assist in the process of monitoring and controlling personnel, one of which is in the field of code of ethics and discipline (Fauzan, 2020).In addition, the use of AI can also prevent and detect indications of fraud, making it easier for auditors to analyze and carry out investigative activities such as identifying, collecting evidence, examining, and protecting audit evidence obtained (Indrika et al., 2021).
The difference between this study and previous research is that it includes the concept of data analytics used in public sector inspections to detect indications of fraud.Another thing that underlies the difference between this study and previous research is that public sector auditors have the task of examining the management and accountability of state financial accountability that is free from misstatement and fraud.The purpose of this study is to determine the effect of using data analytics to detect indications of fraud in the examination process carried out by public sector examiners.

Fraud Triangle Theory
Fraud is an act that is carried out intentionally by an individual or a group of people to carry out these actions by using resources that do not belong to them for their interests (ACFE, 2018).Fraud acts are divided into three groups, namely (ACFE, 2018): Misappropriation of assets, fraudulent acts that are easily detected for stealing assets so that they can be known by calculating the losses due to the fraud.Fraud in financial statements is fraudulent acts that are carried out intentionally to misstate financial statements or misrepresent amounts to disclose financial reporting to deceive users of financial statements.Corruption and fraudulent acts are difficult to detect because the perpetrators cooperate with many other parties in a network and have a systematic way of working.Fraud triangle theory developed by (Cressey, 1953) (Tuana Kotta, 2016) fraud can occur due to three factors, namely: Pressure, which means that someone commits the fraudulent activity due to the pressure he faces such as financial pressure, lifestyle, work, and other pressures.Opportunity is a factor that causes a person to be compelled to commit fraud, such as the lack of strong internal control that makes someone think there is an opportunity and take advantage of the opportunity.Rationalization is a factor that comes from within a person in the form of a mind that always considers what he does as natural and morally acceptable.

Data Analytics
Big data can help increase its effectiveness in the audit process to issue audit opinions (Purda, L, dan Skillicorn, 2015).With the help of a system of data analytics tools, Big data is done physically and conceptually separately from audited accounting data, it will be difficult to cover up fraud and manipulation due to the use of big data for the audit process in 100% sampling or analyzing the entire population (Chang, 2008).So finding red flags and suspicious findings will increase the potential to detect fraud (Soeprajitno, 2019).The areas examined for research using big data techniques in audit practice will assist the auditor in predicting the risks that will occur and are occurring in more depth (Soeprajitno, 2019).In increasing the use of data analytics, an analytical strategy component is needed that aims to maintain a more structured use of data analytics, and analytical governance built on the values of openness, responsibility, accountability, transparency, and fairness.The analytical framework and Community analytics aim to be able to assist in the use of targeted data analytics and can assist in decision making.

Data Analytics and Detecting Indications of Fraud in the Public Sector
Big data can assist auditors in expanding their search for sources and measures of information needed to detect fraud.With the support of analytical processes that have an impact on improving the quality of examination results.Data analytics assistance helps the auditor in predicting the risks that occur and are happening in more depth (Soeprajitno, 2019).According to (Appelbaum et al., 2017) the use of data analytics can help in making decisions in more real time and help in analyzing more relevant in the business environment.Reinforced by research (Jennifer et al., 2013) that the use of data analytics to detect fraud more effectively can be seen by analyzing user behavior and transaction history.Computer-assisted audit techniques can assist auditors in collecting and evaluating electronic data as accurate audit evidence, besides that auditors must also have knowledge related to accessing and analyzing electronic data (Atmaja, 2016).The use of data analytics can increase anti-fraud control and the opportunity for fraud to occur because financial reports containing fraud will interfere with reliability, accuracy, and financial efficiency in terms of their existence and continuity (Kılıç, 2020).In addition, using data analytics can also be used to strengthen risk management based on big data processes and technology (Chen et al., 2015).In using data analytics tools or audit software, the auditor must have confidence in the validity of the data obtained.In addition, auditors must also prioritize security by identifying and examining data sources that aim to prevent fraud and misuse of information (Oktavia, 2015).Based on these arguments, we concluded that data analytics has a positive influence on the detection of indications of fraud in the public sector.It means the better use of data analytics by public sector auditors, the more accurate and quicker the detection of fraud will be.The hypothesis is thus formally stated as follows:

Sample Selection
This type of research is quantitative using primary data obtained from distributing questionnaires to examiners who work as Public Sector Auditors of the Republic of Indonesia, namely BPK as the population in this study.The reason for choosing BPK is that it is an institution authorized to calculate, assess, and determine state losses in the use of budgets by entities.With this responsibility, public sector auditors will maximize their duties by using software that helps in improving the quality of their work, namely using data analytics.The sampling technique used non-probability sampling with purposive sampling where sampling is based on special criteria that have been set by the researcher to answer research problems (Sugiyono, 2013).The criteria are BPK RI auditors, understanding or using data analytics in the audit process, becoming a member of a professional organization, having a professional certification, and having a minimum of one year of work experience.Respondents who filled out the questionnaire in this study were 33 examiners in the State Finance Auditor I to VII and the Main Investigation Auditor consisting of 22 examiners working from BPK Java Island and 11 auditors working at BPK Sumatra Island.The data collection technique uses primary data obtained from distributing questionnaires to respondents whose criteria have been determined.The questionnaire contains respondents' perceptions of the variables consisting of questions regarding respondent data and questions related to research variables.This study uses a Likert scale which aims to measure the indicators of each variable in the questionnaire with a score of 1 to 5. The questionnaire was distributed through a google form, the process of distributing the questionnaires starts from April to May 2021.

Conceptual Model
This study examines the effect of data analytics as an independent variable on indications of fraud in triangle theory.The framework for this research is as follows: Based on 33 public sector auditors who met the criteria for distributing questionnaires using the google form link. Then it is classified based on the respondent's criteria as follows: Source: Questionnaire Distribution, 2021 Based on Table 1 regarding respondent data, all 33 examiners have met the research criteria and understand data analytics.Several auditors have more than one professional certification so the results in the table are 41 and there is one examiner who is a member of the profession more than one so the results in the table are 34.This indicates that the public sector examiner is competent in accounting and auditing.

Validity and Reliability Test
Based on the validity test table on data analytics variables with analytical framework indicators that produce very high correlations (very high) which is greater than 0.90, while indicators of analytical strategy, analytic governance, and analytical communities produce high correlation values (high) which is greater than 0.70.Overall the indicators in the data analytics variable are valid because they have a greater than 0.50.
The fraud detection variable with opportunity and rationalization indicators produces a very high correlation which is greater than 0.90, while the pressure indicator produces a moderate correlation because it has greater than 0.70.Overall, the indicators in the fraud detection variable are valid because they have a greater than 0.50.The conclusion is that all questions in this research are valid and can be continued for further testing.Based on the validity test table on data analytics variables with analytical framework indicators that produce very high correlations (very high) which is greater than 0.90, while indicators of analytical strategy, analytic governance, and analytical communities produce high correlation values (high) which is greater than 0.70.Overall, the indicators in the data analytics variable are valid because they have a greater than 0.50.The fraud detection variable with opportunity and rationalization indicators produces a very high correlation which is greater than 0.90, while the pressure indicator produces a moderate correlation because it has greater than 0.70.Overall, the indicators in the fraud detection variable are valid because they have a greater than 0.50.The conclusion is that all questions in this research are valid and can be continued for further testing.Based on the results of X²-chi-square, showing good results with a value of 0.000<0.05can be accepted because the smaller X², the better the model.The results of the Tucker-Lewis Index / Non-Normed Fit Index (TLI/NNFI) are 1.00≥ 0.90 and the comparative Fit Index (CFI) is 1.00≥0.90with good fit results.It is said there is the goodness of fit criteria that meet the cut-off value and the evaluation of the model is very good.

Findings
The results of testing the hypothesis contained in Table 6 are the relationship between data analytics variables on fraud detection which shows an estimated value of 0.9207927 and the magnitude of the Z value of the influence of data analytics on fraud detection is 22.68 with a probability value of 0.000, significant at the level alpha 0.05 (0.000<0.05).It can be said that data analytics has a positive and significant effect in detecting fraud committed by public sector auditors more effectively The data analytics in question is software or tools used by auditors in the audit process.The use of data analytics can help in making decisions in more real time and help in analyzing more relevant information in the business environment (Appelbaum et al., 2017).In addition, using data analytics strengthens risk management based on big data processing and technology (Chen et al., 2015).The use of data analytics for auditors in the public sector becomes more effective because data analytics makes it easier for auditors to collect evidence, establish a broad population, predict risks that occur, and increase responsibility as a public sector examiner to the state and society.Data analytics can also assist auditors in analyzing data more quickly and making financial reports that are free from material misstatement and fraud.The Supreme Audit Agency (BPK) as an examiner in the public sector has been using big data since 2009.The first system introduced is the e-audit program which is used to send data from the auditee's information system to the BPK data center, where it will be processed by the master agent.consolidator.This is used to examine the management and responsibility of state finances.Furthermore, auditors in the public sector also apply the system information in public sector accounting (SiAP) which is used by the auditors in managing inspection activities in the field.The data inputted in the SiAP application is only structured data such as financial reports, corrections, and examiner procedure data related to documents, images, and audio as examiner evidence (Dezar, 2021).In the use of data analytics, the public sector examiner puts forward a strong analytical strategy, especially in the purpose of using it to manage state finances better and more responsibly, one of the uses of data analytics is an examination related to the citizenship identification number at population and civil registration service of the ministry of home affairs.and examinations related to the e-catalog system at the government goods and service procurement policy institution with the similarity index method used to group citizenship identification numbers based on the level of similarity of elements and text analysis to obtain retail prices for products found on e-commerce websites.It is also used to make it easier to check when reconciling data through the BPK analytics portal.Knowledge of public sector auditors regarding data analytics is also prioritized by conducting training, brevet, and holding seminars for both internal parties and the general public.
Analytics governance implemented by BPK as a public sector examiner in using data analytics puts forward the values of transparency and responsibility (responsibility) because as a manager of state money, they must consistently report transparently to the public, to produce accurate information.Accountability (accountability) because if the public sector shows good accountability performance to the community, it will be a driving force for reducing fraudulent actions that occur in the public sector.Independence using data analytics can assist public sector auditors in determining the relevant evidence easily and making the right decisions.
Fairness is needed because it relates to the code of ethics of an auditor who prepares a report that must comply with auditing standards.The application of analytical framework in designing the framework remains good by utilizing technology infrastructure to keep data more secure.Utilization of technology infrastructure used by public sector auditors, namely by using the cloud to back up data and issuing SiAP applications to collect audit evidence in more real-time.The data analytics used are also monitored by the IT audit party who understands data analytics so that they are maintained if at any time the system goes wrong.Components of the analytics community need to be prepared for the use of data analytics because it can help in analyzing short-term and long-term planning by looking at the trends that exist in a graph from the processed data, and can also assist in the planning and investigation process.
Computer-assisted audit techniques can assist auditors in detecting fraud because auditors can more easily process data in the form of computer files.So that the auditor can plan fraud detection more quickly and precisely compared to detecting it manually (Atmaja, 2016).
Public sector accounting has a scope under the auspices of high state institutions, public sector accounting practices are the same as accounting practices in general but have limitations on stakeholders, namely parties from the government and the wider community (Hamidah, 2020).Based on the fraud triangle theory, fraud can occur due to pressure, opportunity, and rationalization factors.Pressure factors often occur in the scope of the public sector such as pressure from job demands, and community demands.Opportunity factors are more often used by perpetrators to commit fraud, such as lack of internal control, lack of a culture of honesty applied by those in top positions, taking advantage of positions to commit fraud, and having a special relationship between internal and external parties to commit fraud.Analyzing user behavior and transaction history can also be seen when using data analytics to detect fraud more effectively (Jennifer et al., 2013).The weaknesses of data control and documentation systems can also provide opportunities for perpetrators to commit fraudulent acts, therefore system security is very important to protect important data.In the rationalization factor, the perpetrator feels that what he has done is considered correct.Public sector examiners have whistleblowers by making applications that are provided to the public who have information relating to indications of violations that occur in the environment of public sector examiners.This app helps examiners in the public sector in finding evidence in investigations.The use of data analytics can assist auditors in maintaining the quality of financial reports that are clean from fraud by increasing anti-fraud controls because financial reports containing fraud will interfere with the reliability, accuracy, and efficiency of financial statements in terms of their existence and continuity (Kılıç, 2020).Other ways to prevent and detect fraud, forensic accounting, and investigative auditing are options that must be taken to overcome this problem, assisted by computer forensics to facilitate auditors in analyzing and carrying out investigative activities such as identifying, collecting evidence, examining, and protecting audit evidence obtained (Indrika et al.,2021).

Conclusions
Based on the results of this research, the use of data analytics has a positive and significant effect in detecting indications of fraud in the public sector.The use of data analytics helps public sector examiners in the audit process because it makes it easier to collect audit evidence, analyze data, and determine risks that have occurred or will occur.The main thing in using data analytics is to maintain data security because data is currently an asset for every organization.Public sector auditors use data analytics to increase their responsibilities to maintain report results that must be transparent under existing conditions.Assist in making the right decisions in real-time.
Suggestions put forward for all government agencies, with increasingly significant technological developments can help in the process of operational activities but the main thing that must be instilled in all parties involved in the public sector is honesty.With honesty applied from parties who have high positions, it will affect the attitude of other employees not to commit fraud.Suggestions for examiners in the public sector to continue to increase vigilance in data security which at any time can be attacked by irresponsible parties and increase skepticism in the audit process because with the development of data technology it is easy to be manipulated by actors.To support the ability to use data analytics, examiners can participate in continuing education related to data analytics and information technology, and other supporting tools.
There are limitations in the study, namely the number of respondents who were only 33 public sector examiners due to pandemic conditions and the period of distributing questionnaires during the peak season so that the BPK head office limited access to research for distributing questionnaires directly to auditors.For further research, other variables like auditor ethics can be included that can strengthen the influence of data analytics in detecting indications of fraud in the public sector.

Figure 1 :
Figure 1: Conceptual Model of the Study; Source: Data Processing, 2021

Table 1 :
Description of Respondents

Table 2 :
Data Analytics Variable Validity Test

Table 3 :
Fraud Detection Variable Validity Test

Table 4 :
Reliability Test Based on the reliability table, the results show that Cronbach's alpha data analytics variable is 0.9318>0.70 or 0.9318>r table 0.344 with a very high correlation value (very high) and Cronbach's alpha fraud detection variable 0.8634>0.70 or 0.8634> r table 0.344 with a high correlation value (high).It means data analytics variables and fraud detection variables have reliable and reliable results.

Table 5 :
Test of Goodness of Fit