De-cluttering data to make the fight against fraud more effective
The digital revolution offers new possibilities, but also new challenges for fraud fighters. European Anti-Fraud Office (OLAF) analysts are working on finding the best ways to exploit the huge amounts of data available to them, not only to detect but also contribute to preventing fraud. Lara Dobinson and Corinna Ullrich are heading units in (OLAF) that deal with large amounts of information. They explain how OLAF aims to enhance its analytical capability, and improve both the connectivity of databases and the quality of data.
By Lara Dobinson and Corinna Ullrich, European Anti-Fraud Office (OLAF)
The issue of handling large amounts of data has been on everybody’s mind lately. Getting data, mining it, sharing and protecting personal data have topped the agenda of EU leaders and regular citizens alike.
The truth is that complex organisations tend to hold huge volumes of data. However, these are often scattered across several departments and therefore underused. What institutions need are not only analytical and data management tools, but also clear plans — how can we exploit the data we have to further the goals we want to achieve? In OLAF’s case, how can we be more effective in not only fighting fraud, but also in preventing it?
Efficiency gains in operational analysis
Today’s world of investigation has reached an unprecedented level of complexity due to the multitude of data sources, the variety of formats of the information and, equally importantly, the size of the data that is gathered in a case. This means that identifying entities and the links between them has become more and more problematic. In this respect, OLAF is currently working on solutions that tackle these issues in an efficient and innovative manner, both at the level of its software and its processes.
The part of the analytical process that has been subject to the most important changes in the last decade is that of collation, or assembly. The massive production and easier availability of data have turned the investigative phase of data collation into a very time consuming and resource intensive stage. The traditional methods of sorting data in a common format that can then easily be processed and extracted is, due to the increased flow of information, no longer a viable practice. This can actually paralyse investigations. Finding sufficient evidence that a fraud has been committed within the exponential increase of data has become more challenging than ever.
OLAF has therefore extensively developed its internal data analysis capacity to support the investigative function of the office. This relates to both analysing incoming data of possible investigative interest that may result in the opening of an investigation, as well as the operational analytical support assisting both investigators and, where needed, the national authorities in situations where an investigation has been referred to them.
In all these tasks, searching and dynamically presenting clear and comprehensive data is key to a better understanding of the information gathered. At the same time, data should be ready for enrichment without affecting the raw information already gathered.
The tools that OLAF is developing aim to improve operational analysis through more efficient data aggregation, better searching and retrieving of key information, extraction of all possible meanings from a single piece of information and, of course, providing the ability to pivot on a specific entity in order to obtain as much information as possible.
All these features aim to eliminate the risk of missing crucial information, while offering the possibility to cross-check existing entities against newly extracted ones and alert the analysts and investigators on incoming potentially interesting data. For OLAF, investigative success depends on the ability to get key fraud indicators out of the clutter of big data into a structured and searchable environment through improved data analysis tools.
A focus on fraud prevention
In addition to its investigative work, OLAF also works closely with Member States, the other European Commission services and other EU Institutions to prevent or discourage fraud. For example, under the previous Commission’s Anti-Fraud Strategy, OLAF experts had developed over the last years a system of ‘red flags’ which can indicate whether a particular procurement project has a higher probability of suffering from fraud or corruption. Red flags can relate to bidders (who could for example, have multiple undeclared connections between them), to evaluation teams, or to the procedures themselves (very large tenders, too short timespan for the application process, changes in the project description after the award, etc.).
Figure -collection and analysis of fraud related data
The recently adopted new Commission Anti-Fraud Strategy (CAFS)1 will take fraud prevention in the Commission to the next level, and OLAF will be the one steering its implementation. The 2019 CAFS rests on two pillars: further improved cooperation between the Commission services and an enhanced analytical capability through innovative analytical tools, higher connectivity of databases and improved data quality. The Commission commits to increasing the amount of data at its disposal and enhancing its exploitation for analysis purposes, in particular by bringing together information collected in different databases and concerning different areas of the EU budget.
These more in-depth analyses based on much broader, yet tailored data collection and intense cooperation with the relevant stakeholders will provide more meaningful information in general and in relation to specific sectors and/or Member States. In the future, through the more efficient use of data, OLAF could become not only more effective as an investigative service, but also a European center for excellence in fraud prevention and analysis.