IT Forensics

Every year, white-collar crime causes billions in losses in Germany. According to figures from the Federal Criminal Police Office (BKA), the losses recorded by the police in 2024 amounted to around 2.76 billion euros. Although economic offences accounted for only around one per cent of all recorded criminal offences, they were responsible for more than a third of the total financial losses recorded. Experts also assume that the number of unreported cases is very high, as many companies do not report incidents for reputational reasons.

A valuable source of information for detecting and investigating acts of white-collar crime is a company’s accounting data. This can reveal, for example, that funds have been misappropriated or that entries have been manipulated or fabricated to conceal illegal activities.

A well-known example of fraud detection in financial data is Benford’s law, which examines the frequency of digits in numbers. Since this method fails under certain realistic conditions that directly affect digit frequencies—such as posting limits—Fraunhofer SIT has developed a more flexible and far more effective alternative: model-based digit analysis. Unlike Benford’s law, this approach allows specific characteristics of a dataset to be modeled, thereby significantly reducing the false positive rate.