Audits of a federal loan program : a classification model of reported irregularities

Abstract

The primary objectives of this research are to: (1) determine factors significant in differentiating between projects with and without reported irregularities, and (2) develop an empirical classification model for reported irregularities. The study examines audit reports of Farmers Home Administration rural housing projects. This federal loan program has a high incidence of reported irregularities. The dependent variable of interest is the presence or absence of reported irregularities. The opportunity to allow, the motivation to commit, and the attitude to permit the occurrence of irregularities have been identified in prior literature as factors affecting the risk of fraud. Measures of these factors are used as independent variables in developing an irregularity classification model. Probit analysis indicates measures of opportunity, motivation, and attitude are significant in differentiating between projects with and without reported irregularities. These measures of internal control, cash flow, and the number of developer projects can successfully classify projects as to the presence or absence of reported irregularities with 77% accuracy. Compared to projects with no reported irregularities, projects with reported irregularities are found to be significantly associated with reported internal control weaknesses, a larger number of related parties, lower net income, smaller cash flow, and fewer projects per developer. The results of this research as well as extensions could have monitoring applications for users of the audit report and implications for federal regulation.