The occurrences and potential mitigation of occupational fraud in the international non-governmental organisation (INGO) sector in Vietnam: an empirical holistic approach

The occurrences and potential mitigation of occupational fraud in the international non-governmental organisation (INGO) sector in Vietnam: an empirical holistic approach

Purpose – This research investigated the overall research question as follows: How do factors (individual-level, organisational-level, industry-level, and country-level) influence occupational fraud in the INGO sector in Vietnam? This thesis addresses occupational fraud or internal fraud, rather than external fraud, because the former accounted for the majority of losses due to fraud. Additionally, this thesis has two foci: what may contribute to fraud – the second focus, and what may prevent fraud – the first focus. Specifically, the first focus is on how effective organisational-level factors are at preventing or reducing fraud (undetected, suspected, actual/detected, and future). Whilst, the second focus is on factors (individual-level, organisational-level, industry-level, and country-level) that may contribute to fraud (future only).

Design/methodology/approach – This study develops a conceptual model, built on the fraud triangle by Cressey in 1950. The model combines: fraud (undetected, suspected, actual/detected, and future); individual-level determinants (e.g. pressures); organisational-level determinants [e.g. organisational ethical culture (OEC), organisational ethics program (OEP), and organisational internal control (OIC)]; the industry-level determinant (e.g. the non-profit nature); and country-level determinants (e.g. cultural, political, legal, economic, and technological). The methodological choice of research design is quantitative, using the survey strategy with a self-administered questionnaire to collect data. The questionnaires were administered to the entire target population of 408 INGOs (a census) in Vietnam. In each INGO, only one potential survey respondent was approached. Potential respondents were high-level staff members such as Chief Executive Officers (CEOs), Chief Financial Officers (CFOs), Chief Accountants (CAs), accounting/auditing/financial employees, and other managers (e.g. human resources managers). In order to approach potential respondents, the current research used two survey modes: the paper-based (drop-off) mode (administered through Indochina, a research company) and the web-based (online) mode (administered through the Internet). Furthermore, descriptive and inferential data analysis was performed. In inferential analysis, testing the goodness of the data, including the scales used, was undertaken, prior to hypotheses testing. Chi-square test for independence, independent-samples t-test, exploratory factor analysis (EFA), confirmatory factor analysis (CFA), multivariate analysis of variance (MANOVA), logistic regression, and multiple regression were statistical techniques employed in inferential analysis. CFA was executed by AMOS, whilst the rest were performed by SPSS.

Results – This thesis attained 177 usable returned questionnaires, representing the overall response rate to be 47.6%. The results of statistical analyses are as follows.

What may prevent fraud? – The first focus.

Actual fraud
At the model level, the combination of OEC, the effectiveness of OEP, and the effectiveness of OIC, to some extent, significantly predicted actual fraud. Likewise, the combination of the degree of OEP and the degree of OIC, to some extent, significantly predicted actual fraud. Moreover, at the independent variable level, the dimension of supportability under OEC as well as both the degree and effectiveness of OIC make unique statistically significant contributions to predicting/explaining actual fraud. Specifically, the latter is related to the effectiveness of the three dimensions under OIC. They are: management-performed monitoring activities and process-level control activities; commitment-to-competence control environment; and audits-performed monitoring activities and gifts-hospitality control activities. In which, commitment-to-competence control environment made the strongest unique contribution to predicting actual fraud. Conversely, the degree and effectiveness of OEP play no role in mitigating actual fraud.

Undetected fraud
At the model level, the combination of OEC, the effectiveness of OEP, and the effectiveness of OIC, to some extent, significantly predicted undetected fraud, whilst the combination of the degree of OEP, and the degree of OIC did not. In addition, at the independent variable level, only the dimension of supportability under OEC made a unique statistically significant contribution to predicting undetected fraud, but not the effectiveness and degree of OEP and OIC.

Suspected fraud
At the model level, the combination of OEC, the effectiveness of OEP, and the effectiveness of OIC, to some extent, significantly predicted suspected fraud, whilst the combination of the degree of OEP, and the degree of OIC did not. Additionally, at the independent variable level, there were three dimensions making unique statistically significant contributions to predicting suspected fraud. They were: the dimension of discussability under OEC, the effectiveness of the dimension of ethics-incorporated performance and employment under OEP, and the effectiveness of the dimension of monitoring activities and process-level control activities under OIC. In which, ethics-incorporated performance and employment made the strongest unique contribution to predicting suspected fraud. Conversely, the degree of OEP and OIC played no role in mitigating suspected fraud.

Future fraud
At the model level, the combination of OEC, the effectiveness of OEP, and the effectiveness of OIC, to some extent, significantly predicted future fraud, whilst the combination of the degree of OEP, and the degree of OIC did not. Moreover, at the independent variable level, only the dimension of supportability under OEC made a unique statistically significant contribution to predicting future fraud, but not the effectiveness and degree of OEP and OIC.

What may contribute to fraud (future only)? – The second focus.

For the purposes of this thesis, individual-level determinants (e.g. pressures), organisational-level determinants (e.g. poor OEC, poor OIC), and the industry-level determinant (e.g. the non-profit nature) are micro factors, whilst country-level determinants (e.g. cultural, political, legal, economic, technological) are macro factors. At the model level, the combination of the determinants, to some extent, significantly contributed to engendering future fraud. In addition, at the independent variable level, there were five factors making unique statistically significant contributions to engendering future fraud. They are: two individual-level factors: non-financial personal pressure (e.g. lack of personal discipline, greed) and financial corporate/employment pressure (e.g. management financial interest, low salary); one organisational-level factor: poor OIC; one industry-level factor: the non-profit nature in the operation of INGOs; and one country-level factor: the legal factor in Vietnam. In which, the non-profit nature was the strongest source of future fraud, whilst poor OIC is considered to be a silent source of future fraud.

Originality/value – This thesis has made three distinct contributions to the literature. Firstly, it is the first comprehensive study worldwide to incorporate individual-level, organisational-level, industry-level, and country-level factors into predicting future fraud in the INGO/NPO (non-profit organisation) sector in a specific country (e.g. Vietnam), with statistically empirical evidence. Secondly, it is the first comprehensive study worldwide to examine different stages of fraud occurrence (undetected, suspected, actual/detected, and future), with statistically empirical evidence. Thirdly, it is the first comprehensive study in Vietnam as well as in developing countries to investigate fraud in the INGO/NPO sector (industry-level), with statistically empirical evidence. Besides, this thesis also provided other remarkable theoretical contributions.

Practical implications – The results of this thesis are useful for different stakeholders in the INGO sector in Vietnam. The results may also be beneficial to stakeholders in the INGO/NPO sectors in other developing countries, which have similar or close national context factors as in Vietnam. The groups of stakeholders include: (1) CEOs, CFOs, CAs and other directors/managers of INGOs; (2) donors of INGOs; (3) auditors related to INGOs; and (4) the Vietnamese government. Furthermore, fruitful areas of further research were identified, in which five research areas, being equivalent to five research projects, could use the data collected by this thesis.