Information economics approach to the performance evaluation of error-prone internal auditors

Suppose that an internal auditor inspects financial records of a company or government agency, such as social security payments. Human errors are inevitable in any inspection process, and such an auditing procedure is error-prone as well. This implies that the internal auditor may classify correct payments as incorrect (type I error) or may fail to detect some incorrect payments (type II error). In this article, we consider various methods of estimating (i) the internal auditor's two types of errors and (ii) the total number of incorrect records. If we treat each internal auditor as an information structure with imperfect information, we can calculate the value of information via an information economics approach. We use the same approach to measure and compare the performances of multiple internal auditors with different type I and II errors. In a simulation study, we first show that our proposed estimation method performs very well and then demonstrate how to evaluate the performance of several internal auditors from 0 to 100% on the same scale. Our estimation and evaluation method can be applied to any practical situations in which testing devices, inspection methods, or screening procedures are error-prone.