The fight against fraud has to be at all levels, and use all possible means available to the organization.
However, it is important to distinguish between political, organizational and technical means. Persuading states to organize themselves better to facilitate exchange of information between administrations can be decisive, even with the need to respect individual data confidentiality. Many sources of information are still isolated, incomplete or inadequate in terms of data quality, sometimes because of lack of resources or political will. Some states have therefore decided to improve joint working across state agencies by creating an anti-fraud secretariat or national directorate to organize resource pooling (e.g. police and judicial). It is, however, important that enough resources be made available for this type of organization, to give it the strength to act, and the technical credibility and authority required for a collective approach to the fight against fraud.
Some states have also turned to the market to obtain reliable technical solutions to reduce fraud detection cycles, and also increase the efficiency of audit and investigation teams. Anti-fraud technologies are mature and well-proven, especially in the field of data analysis and statistical interpretation. They also provide full transparency on the deployed models and variables that are used within them. This improves their speed of uptake across the business user community, including by audit and inspection teams.
These unique tools allow us to explore and identify hidden links between individuals and organizations, even as particular social and criminal networks become more active, more essential and more ubiquitous. This is partly because their user and operational interfaces have been greatly improved and made more user-friendly in recent years by research and development companies such as SAS Institute.
It is increasingly possible to monitor individuals who have raised concerns because of their actions, even within data protection limits. And this becomes even more relevant when it is possible to identify and interpret the frame in which the individual is connected, because it affects the level of risk to which the organization is exposed. Combining information from several agencies or from multiple sources can often identify an individual who would not have raised any alarms from any one of them in isolation. Agencies can use powerful algorithms to combine information and improve the analysis of any suspicious links established within a frame, as the UK’s HM Revenue and Customs is doing to reduce tax evasion. This consistent and traceable business analysis can then be passed to the police or other authorities if required.
A good analysis of networks should cover companies and individuals working together, whether employed or not by the organization. Fraud, after all, can emerge from both within and beyond the company or organization: witness recent interest in procurement fraud. The analysis can be mainly based on internal organizational data, rather than public or protected external data, but good, solid analysis offers potential for faster, earlier and more efficient detection of cases of complex fraud, as well as abuses, anomalies and, more benignly, consumer marketing trends and preferences. Reliable knowledge about these areas, gained through analysis for the purpose of fraud prevention, can be shared intelligently across the organization, benefiting other departments with similar missions. Subject to good governance, this kind of information-sharing can help organizations to manage multiple challenges.
But there is another issue that emerges from this. Information-sharing also requires holistic and integrated end-to-end management of the information processing system. This is not just to ensure that the information is suitable for use, but is also essential in the context of regulatory requirements, including ensuring the traceability of the data. Good data governance serves multiple masters.Get your free copy of an eBook: Fraud analytics with SAS