재무분석가의 이익예측 허딩 및 허딩 결정요인 (Analyst Herding and the Determinants of Herding)

서지정보

· 발행기관 : 한국경영학회
· 수록지 정보 : 경영학연구 / 35권 / 4호 / 1241 ~ 1260페이지
· 저자명 : 안윤영, 장진호, 신현한, 유영태

초록

본 연구는 2001년부터 2003년까지 3년 동안 FN-Guide 데이터베이스에 포함되어 있는 국내 재무분석가의 이익예측자료를 이용하여, 재무분석가의 이익예측 허딩행태가 우리나라 증권시장에서 존재하는지 여부 및 이익예측 허딩 결정요인을살펴보았다. 분석결과 첫째, 우리나라 증권시장에서 재무분석가의 이익예측 허딩이 존재하고 있음을 발견하였다. 둘째, 재무분석가의 개인특성 요인은 허딩확률에 유의적인 관련이 없음을 발견하였다. 셋째, 재무분석가의 환경요인과 관련된변수로서, 규모가 작은 증권회사에 소속되었거나 규모가 작은 기업 및 기관투자자지분율이 높은 기업에 대한 분석을 담당하는 재무분석가일수록 허딩확률이 높음을 발견하였다. 이는 재무분석가의 환경요인이 허딩확률에 유의한 영향을 미치는요인임을 의미한다.

영어초록

Herding behaviour occurs when individuals ignore their private opinions and alter their private beliefs to correspond more closely with the publicly expressed opinion of others (consensus). Most recent researches have been identified that the degree of herding behaviour varies from person to person dependent on individual and environmental factors.
So various factors have been hypothesized to influence herding. But perfect explanation for the determinants of analyst herding behaviour is so difficult work. Because the analyst’s output(forecast) is observable, but the inputs(the process leading to the forecast) are unobservable.
Based on the prior herding theories and literatures, this paper investigates an empirical test for the presence of herding phenomenon and its determinants in the context where herding is most frequently mentioned by security analysts. Using a sample of domestic financial analyst’s earnings forecast data included in the FN-Guide database, this paper examines whether analysts’ herding behavior exists in the Korean security market and the determinants of herding behaviour for a three year period from 2001 to 2003.
First, We find that analysts exhibit propensity of prior consensus herding behavior in the Korean security market. This result indicates that the prevailing consensus has influence on analysts’ choices. Second, we find that there are no significant relations between financial analyst characteristics and the likelihood of analyst herding. Third, we find that herding forecasts are more likely to be issued by analysts employed by smaller brokerages, covered smaller companies and covered firms with higher institutional investor ownership. This result indicates that analyst’s environmental factors do influence analysts to herd in an earnings forecast.
We end with the cry for future researches. First, as shown here, we defined and measured ‘herding’ as the notion of the ‘consensus forecast’ which is the simple mean of analysts’ forecasts. But it is widely accepted that there are sheep and shepherds, analysts who are ‘followers’ and analysts who are ‘opinion leaders’. This implies that naively calculating a consensus analyst forecast by averaging individual analyst forecasts is inappropriate. So further research might focus on the new herding measure as mentioned above. Second, if herding behaviour serves as proxy for analyst ineffectiveness, it can be detrimental to firm valuation. So examining the relationship between analyst herding and market valuation is one of the attractive research subjects. These much more demanding tasks are left to future research and we plan to investigate related matters in subsequent work.

참고자료

· 없음