Identifying Hidden Fraud Indicators through Longitudinal Analysis of U.S. Financial Misconduct

Abstract

Systemic risk of fraud in U.S. finance and health care has continued to have an adverse impact on the economy, stretch government programs and reduce public trust in institutions. This study provides a 10 year cross-sectoral (finance and health care) study (years 2010-2020) utilizing databases from the Consumer Financial Protection Bureau (CFPB), Securities Exchange Commission (SEC), Department of Justice (DOJ), Center for Medicare Services (CMS) and Health and Human Services Office of Inspector General (HHS-OIG) to identify potential early warning signs which may occur prior to the detection of fraud by regulatory bodies. Using Longitudinal Trend Analysis, Correlation Mapping, K-means Clustering, and Regression Modeling; this study demonstrates a sustained increase in Fraud Occurrence Rates (FOR) for each sector, and a high positive correlation between the two sectors (Sectoral Correlation Index SCI= .72). The three early warning signs identified within this study: Complaint Escalation Rate (CER), Detection Lag (DL), and Fraud Method Diversity (FMD); were found to possess statistically significant predictive capability for future confirmed fraud events. The trends also demonstrated that spikes in public complaints and healthcare recoveries could potentially serve as an actionable signal to regulators of the need for increased oversight or investigation. Based upon these findings, the study proposes an Early Warning Fraud Detection Framework (EW-FDF) and advocates for the development of a Unified Fraud Intelligence Network (UFIN) to enable predictive analytic capabilities across agency boundaries.