Practical Benefits of Discounting Historical Audit Samples Using Normalized Power Priors

Abstract

When establishing an overall audit strategy, auditors must assess the risk of material misstatement and determine the nature, timing and extent of further audit procedures. Among other things, the nature, timing and extent of further audit procedures is affected by the nature and extent of misstatements identified in previous audits and thereby the auditor’s expectations in relation to misstatements in the current audit. Deciding and justifying the extent to which these historical data should be considered is challenging. Consequently, auditors often incorporate these data in a nonstatistical manner. However, there are practical benefits to doing this statistically, such as increased transparency and justifiability. In this article, we introduce a statistical approach to incorporate historical data into the current audit based on the normalized power prior. This approach eliminates the need for auditors to decide how much to discount the historical data and enables them to learn this using the current data. We demonstrate that the normalized power prior improves audit efficiency over time.