Audit reviewers who are aware of the personal biases of audit preparers are still influenced by them, according to a new study.
The study, which appears in the current issue of the American Accounting Association journal The Accounting Review, found the result of bias awareness to be the reverse of what one might expect. Instead of causing reviewers to discount audit preparers’ judgments, it led them to rely on those conclusions more than they would without that awareness.
“Reviewers who are not informed about the preparer’s affect are not significantly influenced by the preparer’s judgment and arrive at judgments more consistent with the audit evidence,” said the study, written by Vicky B. Hoffman, a professor of business administration at the University of Pittsburgh, who carried out the study with doctoral candidate Michele L. Frank. “In contrast, reviewers who receive the same preparer judgment and workpaper evidence, but who are also informed of the preparer’s affect, are significantly influenced by the preparer’s judgment."
The new study corresponds with earlier research on the difficulty people experience when confronted with an adviser's conflict of interest, or that jurors encounter when they are instructed to ignore information improperly introduced in the course of a trial.
The new study focuses on whether auditors fall prey to what is known as the “ironic rebound effect,” allowing them to be swayed by judgments of colleagues who acknowledge personal bias.
“Failure to mitigate doesn't mean that reviewers simply adopt a preparer's recommendation,” Hoffman said in a statement. “It does mean letting a preparer's bias significantly impede a reviewer's own good judgment.”
She and Frank were both auditors at Big Four accounting firms before pursuing careers in academia. They believe the issue to be important, since, in the words of the study, “partners often visit audit staff on client engagements, taking them to lunch or interacting with them in other informal social settings. In these settings it is not unusual for preparers with relatively strong personal feelings toward the clients’ personnel to discuss them with the rest of the audit team."
The paper's findings came from an online experiment involving 119 audit managers and senior managers from two Big Four accounting firms who had an average of more than nine years of experience. All the participants received the same background information, of approximately 500 words, on a manufacturer of electronic components whose primary competitor has developed a prototype of a product that may be technically superior to the focal company's version and is expected to sell for about 20 percent less. The participants in the experiment were asked to assume the role of an audit manager assigned to review inventory judgments made by a hypothetical senior auditor, the preparer.
In half the cases the preparer reached a conclusion, very favorable to the client, that no writedown of inventory is required, while in the remaining cases a highly unfavorable writedown of $1.2 million was recommended. Half the reviewers of the favorable recommendation were told that the preparer had mentioned previously working with the client’s controller at a different company and really liking him.
Meanwhile, half the reviewers of the unfavorable conclusion were informed that the preparer had an unhappy prior experience at another firm with the controller, whom he found arrogant, condescending and extremely unpleasant. The remaining subjects, serving as control groups, received no information at all about personal likes and dislikes.
The researchers noted that the background information presented to participants "indicated that client personnel were competent and that the audit firm had issued unqualified opinions in each of the last five years." In addition, "by design the preparer's relationship with the controller was based on their interactions at a different company to avoid unintentionally providing participants in the affect conditions with additional information about the company's current audit situation." Questioning of the research participants after the experiment revealed that they "perceived client personnel to be relatively competent and audit risk to be relatively low."
Despite assurances of client competence, the audit reviewers made widely different judgments that were apparently influenced by whether the preparer liked or disliked the controller. Among the subjects who reviewed judgments favorable to the client (no writedown required), those who were informed that the controller was very pleasant urged an average writedown of $42,000, which was only about one-third of the $119,200 suggested by those who were told nothing of personal relationships.
As for participants who reviewed judgments unfavorable to the client ($1.2 million writedown proposed), the audit reviewers with no knowledge of personal likes or dislikes recommended an average of $183,103, while those whose preparer said the controller was extremely unpleasant called for almost double that, a whopping $357,333.
What accounts for these wide disparities? It is highly unlikely, the study surmised, that reviewers were naive about preparers' potential to be unduly influenced by their feelings toward clients. In a supplementary experiment involving 40 seasoned auditors, 38 said that there was at least some chance that the audit preparers' personal feelings toward their clients would lead them to make more favorable or unfavorable judgments than warranted, and 27 estimated that the chance of this was 50 percent or better.
Instead the authors sought a cause in the "ironic rebound effect." Psychological research, they explained, has found that when people attempt to ignore or minimize facts (in this study, the preparer’s biased judgment), a subconscious monitoring process occurs to alert them when they are relying on this information. If there is uncertainty about how much to discount the facts, the monitoring process intensifies, keeping the awkward info prominent in working memory and thereby increasing its effect on judgments. Among settings where the effect has been studied are law courts, where it was found that jurors told to disregard a piece of evidence ironically attend to that evidence more than those who are not trying to ignore it.
The authors also surmised that in actual audits this effect on reviewers may be increased by “secondhand affective reactions, which are subconsciously positive toward client personnel that the preparer likes and negative toward those that the preparer dislikes."
As to how the ironic rebound effect can be countered, the authors were cautious, given the complexity of the psychological mechanisms involved. "Hopefully, future research can test how to improve the review process," said Hoffman. "Meanwhile, it may help to simply recognize that the process is not as effective as generally believed and that reviewers should recognize their susceptibility to being unduly swayed by preparer bias. A little alarm bell should go off when personal feelings about client personnel come to light, prompting reviewers to consider what judgment the preparer would likely have reached based on the evidence in the workpapers."
The study, "How Audit Reviewers Respond to an Audit Preparer's Affective Bias: The Ironic Rebound Effect," appears in the March/April issue of The Accounting Review, published every other month by the American Accounting Association.
