발생액의 질과 재무분석가의 정보 환경이 이익예측 정확도에 미치는 영향

This study examines how accruals quality and individual analysts’ information environments affect on the accuracy of analysts’ earnings forecasts. Barron et al.(1998) reveal about the properties of analysts’ information environments, developing BKLS consensus, which is the measure of information commonality individual analysts possess at the time they announce earnings forecasts. BKLS consensus assumes analysts’ information can be represented as either common to all analysts or idiosyncratic to some individual analysts. Thus, by using BKLS consensus, we can measure the degree to which analysts base their forecasts on common information relative to idiosyncratic information. Francis et al.(2005) suggest that accruals quality is a proxy for the information risk associated with earnings because this measure tells investors about the mapping of accrual-based accounting earnings into cash flows. Relatively poor accruals quality weakens this mapping and, therefore, increases information risk. First, we investigate whether accruals quality is associated with BKLS consensus, the ratio of common-to-total information in the average analysts’ forecasts. Second, we examine whether analysts’ behavior to change the degree of information commonality depending on accruals quality contributes to improve the accuracy of earnings forecasts. We find that higher accruals quality decreases BKLS consensus, which shows lower information risk decreases investors’ demands on common information which is publicly available to all investors. Accordingly, such investors’ demands enable analysts to generate more idiosyncratic information rather than common information. Our results are consistent with Indjeijikian(1991), which find that higher level of disclosures afford analysts more opportunity to develop and convey more uniquely private information in their individual forecasts. Further, we provide evidence that the accuracy of analysts’ earnings forecasts decreases as BKLS consensus increases, which suggests that high ratio of common-to-total information in the average analysts’ forecasts decreases analysts’ earnings forecast accuracy. However, the increased reliance on common information as for firms with high accruals quality contributes to increase analysts’ earnings forecast accuracy. Therefore, these findings suggest that, as for firms with high accruals quality, analysts do not have to spend their time and economic resources to acquire or develop idiosyncratic information because using common information is helpful to increase earings forecast accuracy. In sum, while analysts tend to place more reliance on idiosyncratic information rather than common information as for firms with high accruals quality, we find that their efforts on generating idiosyncratic information do not contribute to increase the accuracy of earnings forecasts. Our study suggests that it is cost effective for analysts to evaluate firms’ accruals quality before they work harder to exploit more idiosyncratic information to increase the accuracy of earnings forecasts. Our results are consistent with Lobo et al.(2006) which find that low accruals quality are followed by a great number of analysts in response to greater demand on analyst services and more effort by analysts when firms have high information risk. Prior studies find that corporate disclosures are important source for analysts in forming their earnings forecasts. Byard and Shaw(2002) find that analysts rely more heavily on publicly available financial data rather than privileged communication with management. Evidence on how analysts use accounting information is still one of the main issues among those interested in understanding analysts’ role as information intermediaries. Our study contributes to better understand how analysts react to market participants by analyzing that earnings forecast accuracy differs according to not only information environments but also its interaction effect with the quality of accounting information.