Allocating Internal Audit Resources to Minimize Detection Risk Due to Theft.
研究内部审计资源分配策略与员工盗窃策略的互动，发现审计师应将更多资源分配给审计成本更高的地点，而员工则倾向于从更大或审计成本更高的地点盗窃。
Abstract We analyze internal audit resource allocation strategies and employee theft strategies in a setting in which there are many asset types or "locations" controlled by a single employee where theft can occur. The term "location" refers to an auditible group of assets of equal value and identical audit sampling cost for each unit in the group. Theft detection likelihoods are based on a discovery sampling formulation (the Poisson approximation to the binomial distribution) commonly used in theft and fraud scenarios; detection risk depends on the extent and allocation of auditing, the extent and allocation of theft, and the size of locations. In our analysis, the auditor, by choice of a sampling strategy, minimizes detection risk, or equivalently maximizes the probability that theft is detected. The agent maximizes the expected net gain from undetected theft, where an increasing penalty is imposed on the agent as the amount of detected theft increases. Using these behavioral assumptions, we find that the auditor's strategy involves allocating to each location a share of the audit budget proportionate to the ratio of the cost of auditing the entire location to the cost of a complete (100 percent) audit of all locations. One implication of this result is that if the sampling cost at a location increases, then, contrary to an optimal allocation in a non-interactive setting, the auditor allocates more resources to the more costly location. We also find that the agent's theft strategy involves more theft from larger locations and from locations with higher audit costs.