Why Expertise Is Important for the Detection of Abnormal Performance: The Hot Hand Strikes Back. (October 2016)

• Record Type: Journal Article
• Authors: Frame, David; Hughson, Eric; Leach, J. Chris
• Is Part Of: Journal of accounting, auditing & finance. Volume 31:Number 4(2016:Oct.)
• Page Start: 428
• Page End: 440
• Publication Date: 2016-10
• Subjects: auditors -- Sarbanes–Oxley -- hot hand -- regulation -- PCAOB; Accounting -- Periodicals; Auditing -- Periodicals; Finance -- Periodicals; 657
• DOI: 10.1177/0148558X16628933
• Languages: English
• ISSNs: 0148-558X

Abstract

The Sarbanes–Oxley Act emphasizes auditor and analyst independence as defining characteristics of quality financial reporting. We argue that this focus on independence may have adverse effects on auditor expertise. In particular, an independent auditor will lack firsthand knowledge regarding the construction of a client's financial statements, knowledge that is obtained through the provision of a variety of non-audit services. This lack of firm-specific expertise introduces the potential for a bias analogous to that alleged in sports fans' belief in the "hot hand" in athletic performance. In studies of professional sports, little evidence of serially abnormal performance has been documented by empirical researchers. As a result, many assert the absence of a hot hand in sports and conclude that athletic spectators' beliefs and assertions thereof are not rational. We demonstrate that commonly used tests lack power specifically in the detection of abnormally elevated streaks in performance, even for large sample sizes. Importantly, in contrast to the usual non-structural (and non-parametric) approach, incorporating a structural-data-generating process dramatically enhances an observer's ability to make the correct inference concerning the existence of a hot hand even if only a few observations are available. Of relevance to the debate regarding the (dis)advantages of independent auditors, we argue that auditor expertise regarding the underlying structure of the data-generating processes is paramount in increasing statistical power in (a) an auditor's detection of anomalous corporate behavior and (b) inquiries by the Public Company Accounting Oversight Board (PCAOB) regarding auditor inference and potential corruption.