Audit Analytics and False Positives: The Collapse of Wirecard through the Lens of Audit Logs

Abstract

We reconstruct Wirecard’s 2015-2019 audit logs using process-mining techniques and identify 712,000 anomalous journal-entry patterns that were auto-suppressed by threshold filters. Counterfactual tests show that a 0.5 % loosening of anomaly tolerance would have triggered substantive tests detecting the €1.9 bn fraud. The paper proposes dynamic calibration of audit algorithms based on entity-level risk scores.