Hun-Tong Tan

Modeling the determinants of audit expertise

Abstract Libby & Luft [ Accounting, Organizations and Society (1993) pp. 425–450] presented a model of the relations between experience, ability, knowledge, and performance in audit judgment. This paper extends the model by developing a framework for predicting the structure of these relations in different judgment in different judgment settings and provides an initial test of the predictions by analyzing data from Bonner & Lewiss [ Journal of Accounting Research (Supplement 1990) pp. 1–20] four tasks using LISREL. Key predictions were that problem-solving ability would directly affect performance only in unstructured tasks and would indirectly affect performance through its effect on knowledge acquisition where the learning environment was impoverished. The predictions were supported in most cases. In addition, the paper provides and tests a basis for predicting the associations between performance on different audit tasks. Construct measurement problems that need to be addressed in future research are also indicated.

Analysts' Reactions to Earnings Preannouncement Strategies

Preannouncements of earnings tend to overstate negative or understate positive news, which decreases the chance of a negative surprise when actual earnings is announced. We conduct an experiment to investigate how experienced sell-side analysts’ earnings forecasts are affected by preannouncements that either understate, accurately state, or overstate the magnitude of positive or negative total earnings news, holding total earnings news constant. We find that firms with negative (positive) total news receive the highest post-earnings announcement forecasts of future earnings when the earlier preannouncement overstates (understates) the magnitude of the news. These forecasts are consistent with the analysts’ perceptions about the firms’ future prospects, but not their perceptions of management. While analysts expect preannouncements to be lower than actual earnings, they do not adjust their forecasts for these beliefs. These insights into analysts’ responses have implications both for managers and analysts.

When the Use of Positive Language Backfires: The Joint Effect of Tone, Readability, and Investor Sophistication on Earnings Judgments

Recent studies document that market participants react positively to the positive language sentiment or tone embedded in financial disclosures, and that investors’ reactions to negative news are more muted with poor disclosure readability. However, while language sentiment and readability co-occur in practice, their joint effects remain largely unexplored. In an experiment with MBA students as participants, we investigate how the effect of language sentiment varies with readability and investor sophistication level. We find that language sentiment influences investors’ judgments when readability is low, but not when readability is high. Specifically, when readability is low, disclosures couched in positive language lead to higher earnings judgments for less sophisticated investors, but lower earnings judgments for more sophisticated investors. These findings show that the main effects of readability and language sentiment documented in prior studies have boundary effects, and may reverse when both variables are jointly considered along with investor sophistication.

Effects of the preparer's justification on the reviewer's hypothesis generation and judgment in analytical procedures

Abstract The process gains arising from the review process have been the subject of much research interest, but few studies have examined how features in the audit environment can influence the review process. In this paper, we investigate how one such feature — the preparers justification for a conclusion made — affects the nature of hypotheses generated by reviewers and their likelihood assessments. Using the context of an analytical procedures task, our results showed that exposure to a preparers justification for a non-error cause reduced the number of alternative error hypotheses generated by the reviewer, and increased the reviewers belief in the likelihood of a non-error cause. We also showed that three mechanisms moderated the influence of the preparers justification on the reviewers judgment — generation of alternative hypotheses, explaining an alternative hypothesis, and making an initial independent assessment. Implications of the results are discussed.

Are Reviewers' Judgements Influenced by Memo Structure and Conclusions Documented in Audit Workpapers?*

This paper investigates how the initial conclusion reached by the preparer of audit workpapers and the manner with which the preparer structures the associated evidence can influence the reviewers judgment in an audit setting. We conducted an experiment in which auditors reviewed the work of a preparer who had either concluded that the account of the clients major customer was collectible or not collectible. The preparers memo was structured either in a neutral manner, or stylized to emphasize (de-emphasize) evidence consistent (inconsistent) with the preparers conclusion. Results showed that reviewers placed less reliance on the conclusions reached by the preparer when the preparers memo was stylized than when the memo was structured in a neutral manner. These results suggest that reviewers are sensitive to stylization attempts by preparers. Implications of the paper are discussed.

When Do Analysts Adjust for Biases in Management Guidance? Effects of Guidance Track Record and Analysts’ Incentives

Research in accounting indicates that management has strong preferences for actual earnings to exceed market expectations or analysts’ consensus forecast (e.g. Brown 2001; Matsumoto 2002). Research also indicates that management tends to issue downwardly-biased guidance (Soffer, Thiagarajan, and Walther 2000; Choi and Ziebart 2002; Baik and Jiang 2006). To the extent that analysts adjust for the downward bias in management’s guidance by issuing forecasts that are above management’s guidance, they would increase the likelihood that the firm’s actual earnings will miss analysts’ forecasts, thereby resulting in adverse stock price impact for the firm. However, prior archival and experimental research indicates that analysts do not fully adjust their forecasts for the downward bias in management’s earnings guidance issued between quarterly earnings announcements (e.g., Cotter, Tuna, and Wysocki 2006; Tan, Libby, and Hunton 2002). This adjustment failure accounts in part for the predominance of firms with actual earnings that meet or beat the consensus analysts’ forecast (e.g., Bartov, Givoly, and Hayn 2002; Richardson, Teoh, and Wysocki 2004). Prior research has examined potential explanations for the bias in management guidance (e.g., Skinner and Sloan 2002; Richardson et al. 2004), but not the causes of analysts’ adjustment failure. Our study examines how analysts’ incentives might interact with the consistency and magnitude of bias in management’s guidance in determining

Effect of Concession-Timing Strategies in Auditor-Client Negotiations: It Matters Who is Using Them

In this study, we examine how norms about the use of negotiation strategies by different parties in an auditor–client negotiation influence the relative efficacies of these negotiation strategies. We conduct an experiment with experienced auditors/financial managers as participants, who enter into a negotiation on an income-decreasing audit adjustment with a hypothetical client/auditor who uses a strategy where the same concessions are given either at the start, gradually, or the end of the negotiation. We find that the concession-end strategy is more effective than the concession-start strategy when used by auditors; however, the reverse is true when these same strategies are used by financial managers. The concession-gradual strategy leads to superior outcomes when used by either auditors or clients. We also provide evidence that auditors’ and financial managers’ perceptions of the norms relating to the use of these strategies correspond to what we propose in our theory.