수익비용대응 수준이 재무분석가의 허딩에 미치는 영향

This study examines whether matching principle affects analysts' herding behavior. An accounting practice that poorly matches expenses to revenues impairs earnings quality and increases information uncertainty. Then analysts tend to make forecasts close to consensus of other analysts because they are not certain about their earnings forecasts and try to hide their incompetent forecasting ability, so－called ‘herding.’ On the other hand, when poor matching increases information uncertainty, analysts are also likely to decrease their herding behavior and release earnings forecasts that are differentiated from the consensus of other analysts by using their private information. Based on these two conflicting opinions, we examine how matching principle affects analysts’ herding behavior. Further, this study investigates whether the impact of matching principle on herding differs according to the level of selling and administrative costs. The results show that the degree of matching has significantly negative relationship with analysts' herding behavior. This suggests that when poor matching increases information uncertainty, analysts have difficulty in analyzing the firm and are concerned with releasing inaccurate earnings forecast, resulting in their herding behavior to increase. However, when a firm has good matching practice despite its high selling and administrative costs, which is relatively poorly matched to the revenue, or when a firm has poor matching practice despite its has low selling and administrative costs, we find different results. In this case, analysts have difficulty in understanding the firms' matching practice, resulting in stronger information uncertainty. Thus, analysts decrease herding behavior because they would try to carefully analyze the firm by using their private information. This paper makes a direct contribution to analyzing whether the degree of matching, which enhances earnings quality affects, the decision of analysts' forecasting activity. This study provides an empirical evidence that the difference in information uncertainty according to the level of matching principle affects analysts' herding behavior.