Optimal auditing of social benefit fraud: a case study

Abstract

The aim of this paper is to study the effect of different audit strategies on fraud in one particular social benefit system in Sweden. The efficiency of different audit strategies is compared using a computer-based optimization algorithm. Two types of audit strategies are used. One is to adapt the audit intensity to the propensity for errors and fraud in different segments of the group studied. This type of strategy is denoted segmentation strategy. The second type of audit strategy is based on adaptation of behaviour through information. The model developed by Erard and Feinstein for tax auditing is adapted for benefit fraud. In this model, the audit intensity is controlled by a variable, and the auditees are informed of the relationship between control variable and audit intensity. The control variable used in this paper is the benefit amount claimed during a certain period. As the audit intensity increases with the claim amount, the rational fraudster understands that reducing the amount of fraud decreases the risk of being audited. This type of strategy is denoted information strategy. One main result is that the Erard and Feinstein model can be successfully adapted to benefit fraud. Using coarse estimates of audit unit costs, the result of the study is that all persons should be audited. For higher audit costs, it is shown that the information strategy is much more effective when compared to the segmentation strategy.