D’un indice de détection d’anomalies à l’usage des investisseurs

Fraud detection is a key issue for investors and financial authorities. The Ponzi scheme organized by Bernard Madoff is a magnified illustration of a fraud, always possible when well-orchestrated. Traditional methods to detect fraud require costly and lengthy investigations that involve complex financial and legal knowledge and high skilled analysts. We pursue and generalize here the intuition of Billio et al. [2015], who suggest the use of a performance measure-called GUN-to construct a fraud detection index. To illustrate the methodology and to demonstrate its usefulness, first, we analyze the case of Madoff. Then, in a universe of equity mutual funds marketable in France from several international markets, we secondly highlight the number of funds with an apparent anomalous behavior. The proposed alert system reveals dozens of funds that would be interesting to investigate.