Evaluating the Sufficiency of Causes in Audit Analytical Procedures

SUMMARY When auditors perform analytical procedures, one concern is that they may render an incorrect conclusion about an observed fluctuation. In this paper, we argue that one way that auditors might make this kind of judgment mistake is to conclude that only one cause is responsible for an unexpected fluctuation when, in fact, multiple causes are responsible. We present a model of the evaluation of causes in analytical procedures that decomposes this task into two components -- plausibility checking and sufficiency checking. We argue that auditors are more likely to make such judgment mistakes if they only perform the plausibility check. To address this prediction, we conducted an experiment in the context of analytical procedures using 72 experienced auditors as subjects. Our experimental results show that, as predicted, auditors who merely perform a plausibility check are more likely to erroneously conclude that one cause is responsible for a fluctuation when, in fact, other causes are operating. In contrast, auditors who perform both plausibility and sufficiency checks are much better at discerning when there is one vs. multiple causes responsible for a fluctuation. Key Words: Analytical procedures, Plausibility, Explanation, Sufficiency, Quantification. Data Availability: Available from the authors upon request. INTRODUCTION The purpose of this paper is to explore how auditors performing analytical procedures reach incorrect conclusions about the causes of unexpected fluctuations and how such tendencies can be mitigated. Specifically, we investigate how auditors might conclude that one cause is responsible for a fluctuation when, in fact, there are multiple causes. Archival data from Coglitore and Berryman (1988) show that financial-statement fluctuations are often the result of multiple causes. Of the 19 companies they analyzed, nine had unexpected fluctuations that appeared to be the result of at least two material causes. Despite this, field research by Hirst and Koonce (1996, 471) reveals that auditors routinely assume that only one cause is responsible for a fluctuation.(1) Moreover, those who review analytical-procedures work report that auditors frequently stop their investigation too soon and thus tail to completely account for unexpected fluctuations (Hirst and Koonce 1996, 475). Taken together, these two studies suggest that a potentially serious problem facing audit practice involves auditors erroneously concluding that only one cause is responsible for a fluctuation. Most of the prior research on analytical procedures has focused on how the auditor accepts the wrong cause of a fluctuation (e.g., Bedard and Biggs 1991). That is, this research assumes that there is one cause responsible for a fluctuation and investigates how the auditor might conclude that cause X is responsible when, in fact, it is cause Y. In this study, we relax the one-cause assumption and examine why the auditor might conclude that cause X is solely responsible for a fluctuation when, in fact, both causes X and Y are explanatory. Investigating the latter type of judgment error is important because it may be more likely to remain undetected and thus compromise the effectiveness of the audit. That is, if the auditor accepts the wrong cause, then the supervisory review process is likely to detect that mistake because the erroneous cause will either not fit the pattern of fluctuations or will not have information supporting its occurrence. If, instead, the auditor concludes that one cause is responsible when in fact multiple causes are responsible, then the reviewer would have to realize that there is a missing cause. We suspect that an error of omission is harder to detect than an error of commission. To guide our investigation, we develop a two-part process model of the evaluation of causes during analytical procedures. This model is useful because it allows us to understand how the judgment error under investigation can occur and how it can be mitigated. …