data mining as a tool in fraud investigation 29 june 2015
TRANSCRIPT
Data mining as a tool in fraud investigation
29 June 2015
Reputation promise/mission
The Auditor-General of South Africa has a constitutional mandate and, as the Supreme Audit Institution (SAI) of South Africa, exists to strengthen ourcountry’s democracy by enabling oversight, accountability and governance in the public sector through auditing, thereby building public confidence.
Outline
• Project timeline
• Development of analytical model
• Guide status
• Comments
• 31 October 2013
Chapter Layout
• 31 January 2014
Design and specifications
document
• 31 March 2014
Development of analytic model
• 31 March 2015
Fraud risk assessment
• Brazil 2016
Finalisation and approval
Project timeline
Development of analytical model
Country assignments: To design and submit a fraud scoring model for their SAI using and considering their audit methodology.
• Is the outcome to determine an overall risk rating of the auditee to address ISA240 requirements on planning (high/medium/low risk determination)?
• Are you identifying riskier areas within the auditee to focus on?
What to achieve?
• What factors to use when determining an outcome: Use of analytical and/or non-analytical factors?Decision model
• Will you apply weighted scoring? What basis for the weights?• How to aggregate the score to fit your outcome?
How will scoring work if applied to decision model?
• Rating of the likelihood of fraud?• Response process to determined riskOutcome
Guide status
Purpose of the guide
Definition of fraud
Building of the platform
Fraud scoring model
Fraud detection
Completed
Completed
In progress
Evaluate: Country assignments
Building of the platform
Guide status: Fraud detection
Known schemes
- Normal practice
Fraud scheme
Characteristics Fraud indicator Analytical procedure
Unknown schemes
- ISA 240
- Audit risk analysis
Guide status: Fraud scoring model
Country assignments
- Two (2) members responded
Methodology
- What brings everything together?
- Elements to form a picture; instructions to draw the picture
- SAI SA : practical implementation
Guide status: Fraud methodology
Fraud scheme 1
Characteristics
Fraud indicator
Analytical procedure
Fraud scheme 2
Characteristics
Fraud indicator
Analytical procedure
Proactive analytics
ISA240 Fraud indicator
Analytical procedure
Conclusion
Guide status: Building the platform
Software independent guide
Methodology and fraud scoring model
- Decision tree
- Flowcharting
Comments
Tobie Bruyns (CISA)email: [email protected]:+27 12 422 9831SAI – South Africa