Persistent Oversight Failures in State-Administered Federal Grant Programs:A Longitudinal Analysis of Single Audit Findings and the Case for Risk-Based Compliance Reform

Abstract

State-administered federal grant programs distribute over $700 billion annually, yet improper payments reached $161.5 billion in fiscal year 2024 and GAO estimates annual fraud losses between $233 billion and $521 billion. Despite three decades of Single Audit oversight, there is limited empirical evidence on whether current monitoring structures interrupt persistent noncompliance. This paper presents a longitudinal analysis of Single Audit finding patterns using Federal Audit Clearinghouse data from 2016 to 2023 (N = 7,147 state-program-year observations across 955 unique entities in 56 states and territories). We find that compliance failures are highly concentrated: the top 12.40% of uniquely identified entities account for 90% of all finding-years, with a Gini coefficient of 0.8902. Fixed-effects regression using a two-year lagged predictor, which is definitionally independent of the outcome, shows that prior finding history is associated with a 9.48 percentage point increase in finding persistence probability (SE = 0.026, p = 0.0002), confirming genuine organizational persistence beyond the mechanical structure of the immediate lag. Finding rates remained stable at approximately 47% across the full analytical period with no descriptive evidence of improvement. Descriptive post-finding trajectories show no sustained improvement below baseline, consistent with symbolic compliance dynamics in which formal oversight artifacts are produced without reliable compliance improvement. Drawing on these findings, we propose a risk-stratified compliance reform framework grounded in COSO Internal Control principles. Policy implications for Congress, OMB, and state practitioners are discussed.