A Directed Graph Based Model for Fault Diagnosis in Internal Control Systems

Auditors rely on two important types of knowledge when performing audit tasks. The first type is the knowledge of financial transactions, internal controls, their interrelationships, and the associated processes, i.e., the knowledge of the structure of internal control systems and financial reporting in accounting cycles. The second type is the more heuristic knowledge that comes from experience. Most of the existing research on modeling auditor knowledge and behavior has focused on the second type of knowledge, i.e., probabilistic models and expert systems that rely on heuristics and expert auditor experience. While knowledge based on heuristics and experience is important, evidence suggests that knowledge about the structure of internal control systems is equally important in helping auditors perform their job of assessing and testing internal controls. This paper proposes a new approach to modeling the structure of internal control systems. The model uses a directed graph to represent the various elements in an internal control system, such as financial statement assertions, control activities, financial processes, and the causal relationships that exist among these elements. The paper demonstrates the usefulness of the model by presenting and discussing an algorithm based on this model to help external auditors in fault diagnosis during substantive testing and internal auditors in assessment of internal control systems