State-of-the-art: data analytics for public sector fraud detection and mitigation
Scott Mongeau, Senior Manager Business Solutions, Fraud and Financial Crime (SAS Institute)
While it is not always apparent, everyone is affected by public sector fraud. It imposes costs in the form of higher taxes, disrupted services, and breakdowns in trust. Crucial services become more costly and difficult to administrate. In extreme forms, prolific fraud is a cyclical affliction of failed municipalities and states. Even in stable societies, it has been estimated that fraud encompasses upwards of 5% of all transactions.
The public sector encompasses all services that a government provides for the administration of public and commercial life in a society. Although each country bundles and designates public services in unique ways, common service categories include defence, intelligence, policing, transport, finance, education, and healthcare. Additional core services encompass agriculture, utilities, housing, telecommunication, infrastructure, and public financing. The degree to which a service is privatised, semi-privatised, or nationalised varies by country. Typically a government exerts at least a regulatory and oversight role in those areas where functions are deemed of central importance to national stability.
In terms of the types of fraud which afflict public sector entities, the areas of risk are broadly delineated by the scope of the particular agency, although all common forms of well-known internal and external fraud are generally in scope: theft, improper claims, ghost employees, bribery, bid-rigging, kick-backs, money laundering, internal threats, collusion, and organised crime.
Professionals dedicated to fraud detection and mitigation agree that the scope of public sector fraud is expanding in complexity and scale. Fraud is subject to the power of economic incentives; when there are sufficient motivations and means, fraud will occur. Public sector agencies, due to the importance of their functions and services, in particular face unprecedented and growing challenges due to emerging sophisticated forms of fraud. Globalisation and the advancement of communication and computer technology has enabled increasingly sophisticated forms of fraud.
Public sector agencies increasingly face a technology forged double-edged sword. Technological progress and ever growing data volumes present both opportunities and threats to public sector agencies. While agencies face growing demands for improved and streamlined services to serve the public interest, the increasingly interconnected nature of these technologies also creates new vulnerabilities for abuse and attack. The very power of rapidly developing computer and communication technologies are creating new vehicles and methods for the perpetration of fraud. A particularly worrying trend is cyber fraud, the use of computer and network means by criminals and criminal organisations to facilitate ‘traditional’ forms of fraud.
Whereas the internet and prolific computing has made possible increasingly sophisticated forms of fraud, the same tools also can be deployed as a powerful weapon to detect and mitigate fraud. The advent of powerful computing technology has provided a promising tool in the struggle to detect and prevent public sector fraud. The latest innovation has been data analytics, also referred to as data science - a method to automate the detection of fraud patterns in large sets of data.
Data analytics-driven big data solutions offer powerful approaches to improve fraud detection accuracy, even potentially to detect and block fraud in real-time as it is being perpetrated. In addition to mitigating fraud, analytics also offers a platform for streamlining services, reducing costs, and improving reporting, accountability, transparency, and trust.
The benefits of applying data analytics for public sector fraud detection and mitigation are broad and powerful. It is not just about ‘stopping the bad guys’ – a fraud analytics program establishes a platform from which a diverse range of benefits grows, including:
- Preventing the direct cost of fraud
- Reducing fraud monitoring costs
- Mitigating persistent risks
- Improved metrics and oversight
- Improved forecasting capabilities
- Streamlining processes and efficiency
- Improving data management and governance
- Addressing compliance and reporting requirements
- Building collaboration and sharing across public functions
- Improving transparency, governance, and accountability
- Strengthening citizen engagement through improved reporting
- Establishing a platform for real time monitoring and transaction control
Whereas fraud examiners traditionally conduct spot-checks and hope to detect and stop a small percentage of fraud, many organisations now are moving to automate the monitoring of fraud risks by analysing large sets of data, or big data. Big data is processed via data analytics to search for complex fraud patterns, to potentially pinpoint ALL violations of a certain type, and to discover new types of fraud patterns through exploratory techniques. Data analytics brings the power of advanced computing and big data together as a powerful tool to combat public sector fraud.