Research Article
The Effect of Audit Hours on Financial Analysts' Earnings Forecast Bias
Yeungnam University
Published: January 2014 · Vol. 43 No. 4 · pp. 1079-1111
Full Text
Abstract
This study examines the relation between abnormal audit hours and bias in analystsʹ earningsforecasts. We expect a negative relation between abnormal audit hours and optimistic forecasterror. This study is motivated by prior studies which report positive association between abnormalaudit hours and overall reliability of accounting information. The increase in the reliability ofpubic information would likely induce analysts to depend less on private information. The relativeindependence fromprivate information may allow decrease in analystsʹ issue of optimistic forecasts. The proxy for forecast error is measured as signed error(SE, actual-forecast) following Jeong(2003) and Jeong et al.(2012). Forecast greater than actual EPS is interpreted as optimisticforecast. As SE decreases, the forecast becomes more optimistic. Abnormal audit hours areestimated following Ma et al.(2012) and Park and Park(2007). The hypothesis is tested by using sample firms listed on the Korean Stock Exchange(KOSPI)from the year of 2006 to the year of 2011 inclusively. After controlling for variables relatedwith bias in analystsʹ earnings forecasts as reported in previous studies, the results of multipleregression show a significantly positive coefficient for audit hour variable. The decrease in theoptimistic bias indicates decrease in the demand for private information. In addition we performed additional tests by dividing the sample based on several firmcharacteristics(abnormal audit hours above zero vs. abnormal audit hours below zero; abnormal audit feeabove zero vs. abnormal audit fee below zero; big4 vs. non-big4). The results of the additionaltests provide similar results. We also performed 2-Stage Least Squares(2SLS) test to control for potential endogeneity. The additional result is consistent with our main findings. Our resultsindicate decrease in optimistic bias increases analystsʹ independence from managementʹs privateinformation. Overall, the results of this study suggest that the quality of accounting information isincreasing in audit hours. And the improvement of the quality of accounting informationreduces analystsʹ dependence on managementʹs private information.
